Fixed vm bugs, refactored the entire broject to use regexis024 namespace, removed some junk, added tests

This commit is contained in:
Андреев Григорий 2024-07-30 01:02:07 +03:00
parent b11afa72ea
commit 76f3742521
79 changed files with 4988 additions and 4875 deletions

View File

@ -72,7 +72,6 @@ struct Libregexis024BuildSystem {
"libregexis024fa/graph_to_bytecode/core.cpp",
"libregexis024sol/common_codesets.cpp",
"libregexis024sol/part_of_expr_that_tracks.cpp",
"libregexis024sol/expr_compiler.cpp",
"libregexis024sol/square_bracket_expression.cpp",
"libregexis024sol/sol_misc_base.cpp",

View File

@ -10,19 +10,20 @@
#include <libregexis024vm/vm_opcodes.h>
#include <libregexis024fa/tracking_fa_nodes.h>
const char* one_char_read_color = "black";
const char* forking_color = "darkorchid1";
const char* look_one_behind_color = "darkslateblue";
const char* look_one_ahead_color = "coral1";
const char* track_array_mov_imm_color = "lightblue2";
const char* track_array_mov_halfinvariant_color = "lightseagreen";
const char* match_pending_lob_color = "darkgoldenrod2";
const char* match_color = "gold";
const char* det_char_crossroads_color = "navy";
const char* error_color = "crimson";
const char* STAR = "";
namespace regexis024 {
const char* one_char_read_color = "black";
const char* forking_color = "darkorchid1";
const char* look_one_behind_color = "darkslateblue";
const char* look_one_ahead_color = "coral1";
const char* track_array_mov_imm_color = "lightblue2";
const char* track_array_mov_halfinvariant_color = "lightseagreen";
const char* match_pending_lob_color = "darkgoldenrod2";
const char* match_color = "gold";
const char* det_char_crossroads_color = "navy";
const char* error_color = "crimson";
const char* STAR = "";
const char* get_associated_color(FA_Node* node){
const char* get_associated_color(FA_Node* node){
switch (node->type) {
#define ccase(tn) case tn: return tn##_color;
ccase(one_char_read)
@ -38,20 +39,20 @@ const char* get_associated_color(FA_Node* node){
return "black";
#undef ccase
}
}
}
struct NodesProblems{
struct NodesProblems{
size_t actual_refcount = 0;
bool refcount_problem = false;
size_t edges_point_to_null = 0;
};
};
struct EdgesProblems {
struct EdgesProblems {
bool points_to_null = false;
explicit EdgesProblems(bool points_to_null): points_to_null(points_to_null) {}
};
};
std::string get_applied_edge_attributes(FA_Node* node, const NodesProblems& np, const EdgesProblems& ep){
std::string get_applied_edge_attributes(FA_Node* node, const NodesProblems& np, const EdgesProblems& ep){
std::string res = "color=";
if (ep.points_to_null) {
res += error_color;
@ -61,9 +62,9 @@ std::string get_applied_edge_attributes(FA_Node* node, const NodesProblems& np,
res += " style=bold";
}
return res;
}
}
std::string get_applied_node_attributes(FA_Node* node, const NodesProblems& bd){
std::string get_applied_node_attributes(FA_Node* node, const NodesProblems& bd){
std::string res = "color=";
res += get_associated_color(node);
if (bd.refcount_problem)
@ -72,9 +73,9 @@ std::string get_applied_node_attributes(FA_Node* node, const NodesProblems& bd){
(node->type == det_char_crossroads && dynamic_cast<FA_NodeOfDetCharCrossroads*>(node)->matching))
res += " shape=doublecircle";
return res;
}
}
void append_reverse_hex(std::string& res, uint32_t num){
void append_reverse_hex(std::string& res, uint32_t num){
if (num == 0){
res += "0";
} else {
@ -84,9 +85,9 @@ void append_reverse_hex(std::string& res, uint32_t num){
num >>= 4;
}
}
}
}
std::string stringify_codeset(const codeset_t& cs){
std::string stringify_codeset(const codeset_t& cs){
std::string res;
for (long i = static_cast<long>(cs.size()) - 1; i >= 0; i--) {
uint64_t start = cs[i].first, end = cs[i].second;
@ -102,9 +103,9 @@ std::string stringify_codeset(const codeset_t& cs){
}
std::reverse(res.begin(), res.end()); /* ascii works wonders */
return res;
}
}
std::string get_extended_node_lable(FA_Node* node){
std::string get_extended_node_lable(FA_Node* node){
if ((node->type == one_char_read && dynamic_cast<FA_NodeOfOneCharRead*>(node)->second_ns) ||
(node->type == det_char_crossroads && dynamic_cast<FA_NodeOfDetCharCrossroads*>(node)->second_ns)) {
return std::string(" ") + STAR;
@ -115,9 +116,9 @@ std::string get_extended_node_lable(FA_Node* node){
return std::string(" pending loa ") + stringify_codeset(mn->pending_filter);
}
return "";
}
}
std::string get_node_lable(FA_Node* node, const NodesProblems& bd){
std::string get_node_lable(FA_Node* node, const NodesProblems& bd){
std::string res;
switch (node->type) {
#define tcase(tn, str) case tn: res = str; break;
@ -135,9 +136,9 @@ std::string get_node_lable(FA_Node* node, const NodesProblems& bd){
if (bd.refcount_problem)
res += ("!refcount: " + std::to_string(node->refs) + "!");
return res;
}
}
void print_edge(FA_Node* start, const FA_Node* dest, const std::string& label, FILE* fd, NodesProblems& np){
void print_edge(FA_Node* start, const FA_Node* dest, const std::string& label, FILE* fd, NodesProblems& np){
if (!dest){
fprintf(stderr, "NULL transition going from node %lu\n", start->nodeId);
fprintf(fd, "%lu->NULL_%lu_%lu [label=\"%s\" color=crimson]", start->nodeId,
@ -146,9 +147,9 @@ void print_edge(FA_Node* start, const FA_Node* dest, const std::string& label, F
}
fprintf(fd, "%lu->%lu [label=\"%s\" %s]\n", start->nodeId, dest->nodeId, label.c_str(),
get_applied_edge_attributes(start, np, EdgesProblems(false)).c_str());
}
}
void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
const RegexPriorityTable& priority_table){
assert(fa.start);
assert(fd);
@ -184,7 +185,7 @@ void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
/* Two Infoboxes */
auto stringifyTrackingVarType = [](tracking_var_type type) -> std::string {
auto stringifyTrackingVarType = [](tracking_var_type_t type) -> std::string {
switch (type) {
case tracking_var_types::range:
return "range";
@ -199,7 +200,7 @@ void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
for (auto& p: ktr.track_names){
const SubtrackingNameInfo& tu = ktr.retrieval_info[p.second];
auto getRole = [](bool presence, tracking_var_type type, int first, int second,
auto getRole = [](bool presence, tracking_var_type_t type, int first, int second,
const std::string& ARR_NAME) -> std::string {
if (!presence) {
assert(first == -1 && second == -1);
@ -264,7 +265,7 @@ void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
if (!isImmMovOpcode(cn->operation))
fprintf(stderr, "bad operation in node %lu\n", node->nodeId);
snprintf(buf, 1024, "%s %hu %lu",
regex024_opcode_tostr(cn->operation), cn->key, cn->imm_value);
opcode_to_str(cn->operation), cn->key, cn->imm_value);
print_edge(node, cn->nxt_node,std::string(buf), fd, bd);
} else if (node->type == track_array_mov_halfinvariant){
FA_NodeOfTrackArrayMovHalfinvariant* cn = dynamic_cast<FA_NodeOfTrackArrayMovHalfinvariant *>(node);
@ -272,7 +273,7 @@ void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
if (!isCurPosMovOpcode(cn->operation))
fprintf(stderr, "bad operation in node %lu\n", node->nodeId);
snprintf(buf, 1024, "%s %hu",
regex024_opcode_tostr(cn->operation), cn->key);
opcode_to_str(cn->operation), cn->key);
print_edge(node, cn->nxt_node,std::string(buf), fd, bd);
} else if (node->type == det_char_crossroads){
FA_NodeOfDetCharCrossroads* cn = dynamic_cast<FA_NodeOfDetCharCrossroads *>(node);
@ -284,9 +285,9 @@ void print_fa(const FA_Container& fa, FILE* fd, const KnownTrackingTools& ktr,
}
}
fprintf(fd, "}\n");
}
}
FILE* get_fd(const char* apath){
FILE* get_fd(const char* apath){
errno = 0;
FILE *fd = fopen(apath, "w");
if (!fd)
@ -297,9 +298,9 @@ FILE* get_fd(const char* apath){
if (!fd)
perror("fopen a");
return fd;
}
}
void show_fa_with_sxiv_after_dot(const FA_Container& fa, const KnownTrackingTools& ktr,
void show_fa_with_sxiv_after_dot(const FA_Container& fa, const KnownTrackingTools& ktr,
const RegexPriorityTable& priority_table) {
const char* temp_gv = "FAGraph.gv";
const char* temp_png = "FAGraph.png";
@ -322,4 +323,5 @@ void show_fa_with_sxiv_after_dot(const FA_Container& fa, const KnownTrackingTool
assert(chw >= 0);
unlink(temp_gv);
unlink(temp_png);
}
}

View File

@ -5,8 +5,10 @@
#include <libregexis024sol/part_of_expr_that_tracks.h>
#include <libregexis024fa/selarr_priority_table.h>
/* Uses temporary file FAGraph.gv,png, dot command and sxiv */
void show_fa_with_sxiv_after_dot(const FA_Container& fa, const KnownTrackingTools& ktr,
namespace regexis024 {
/* Uses temporary file FAGraph.gv,png, dot command and sxiv */
void show_fa_with_sxiv_after_dot(const FA_Container& fa, const KnownTrackingTools& ktr,
const RegexPriorityTable& priority_table);
}
#endif

View File

@ -2,40 +2,38 @@
#include <functional>
#include <libregexis024vm/utils.h>
TreeWithStringsNode::TreeWithStringsNode(const std::string &val): val(val) {
}
namespace regexis024 {
static const char* ch_empty = " ";
static const char* ch_passing_by = "\u2502 ";
static const char* ch_connect_right_and_forward = "\u251c\u2500\u2500\u2500";
static const char* ch_connect_right_last = "\u2514\u2500\u2500\u2500";
static const char* ch_empty = " ";
static const char* ch_passing_by = "\u2502 ";
static const char* ch_connect_right_and_forward = "\u251c\u2500\u2500\u2500";
static const char* ch_connect_right_last = "\u2514\u2500\u2500\u2500";
static const char* ch_box_left_side = "\u2551";
static const char* ch_box_right_side = "\u2551";
static const char* ch_box_top_side = "\u2550";
static const char* ch_box_bottom_side = "\u2550";
static const char* ch_box_crn_top_left = "\u2554";
static const char* ch_box_crn_top_right = "\u2557";
static const char* ch_box_crn_bottom_left = "\u255A";
static const char* ch_box_crn_bottom_right = "\u255D";
static const char* ch_box_left_side = "\u2551";
static const char* ch_box_right_side = "\u2551";
static const char* ch_box_top_side = "\u2550";
static const char* ch_box_bottom_side = "\u2550";
static const char* ch_box_crn_top_left = "\u2554";
static const char* ch_box_crn_top_right = "\u2557";
static const char* ch_box_crn_bottom_left = "\u255A";
static const char* ch_box_crn_bottom_right = "\u255D";
size_t length_of_line(const std::string& str) {
size_t length_of_line(const std::string& str) {
size_t ch = 0;
size_t pos = 0;
while (pos < str.size()) {
int32_t code;
size_t adj;
utf8_string_iterat(code, adj, pos, reinterpret_cast<const uint8_t*>(str.data()), str.size());
utf8_string_iterat(code, adj, pos, str.data(), str.size());
if (code < 0)
return ch;
ch++;
pos += adj;
}
return ch;
}
}
/* Warning: recursion used */
void toLines_dfs(const TreeWithStringsNode& node, lines& out, std::vector<bool>& prefix) {
/* Warning: recursion used */
void toLines_dfs(const TreeWithStringsNode& node, lines& out, std::vector<bool>& prefix) {
out.push_back("");
size_t n = prefix.size();
for (size_t i = 0; i < n; i++) {
@ -54,21 +52,21 @@ void toLines_dfs(const TreeWithStringsNode& node, lines& out, std::vector<bool>&
toLines_dfs(node.childeren[i], out, prefix);
}
prefix.pop_back();
}
}
void TreeWithStringsNode::toLines(lines &out) const {
void TreeWithStringsNode::toLines(lines &out) const {
std::vector<bool> prefix;
toLines_dfs(*this, out, prefix);
}
}
std::string strMul(size_t n, const char* str) {
std::string strMul(size_t n, const char* str) {
std::string res;
for (size_t i = 0; i < n; i++)
res += str;
return res;
}
}
lines wrapWithBox(const lines &in) {
lines wrapWithBox(const lines &in) {
lines out;
size_t max_width = 0;
for (auto& l: in)
@ -80,10 +78,10 @@ lines wrapWithBox(const lines &in) {
}
out.push_back(ch_box_crn_bottom_left + strMul(max_width, ch_box_bottom_side) + ch_box_crn_bottom_right);
return out;
}
}
void printLines(const lines &in) {
void printLines(const lines &in) {
for (auto& l: in)
printf("%s\n", l.c_str());
}
}

View File

@ -6,20 +6,19 @@
#include <vector>
#include <string>
typedef std::vector<std::string> lines;
namespace regexis024 {
typedef std::vector<std::string> lines;
struct TreeWithStringsNode {
struct TreeWithStringsNode {
std::string val;
std::vector<TreeWithStringsNode> childeren;
explicit TreeWithStringsNode(const std::string &val);
TreeWithStringsNode() = default;
void toLines(lines& out) const;
};
};
lines wrapWithBox(const lines& in);
lines wrapWithBox(const lines& in);
void printLines(const lines& in);
void printLines(const lines& in);
}
#endif

View File

@ -2,15 +2,16 @@
#include <stdio.h>
#include <string>
std::string thread_to_str(const REGEX_IS024_Thread& thread){
namespace regexis024 {
std::string thread_to_str(const Thread& thread){
if (!(thread.slot_occupation_status & SLOT_OCCUPIED))
return "{ unoccupied }";
char buf[1024];
snprintf(buf, 1024, "{ IP = %lu }", thread.IP);
return buf;
}
}
std::string stack_to_str(const REGEX_IS024_Stack& stack){
std::string stack_to_str(const SSID_Stack& stack){
std::string res = "{ ";
for (uint32_t i = 0; i < stack.sz; i++){
if (i != 0)
@ -19,9 +20,9 @@ std::string stack_to_str(const REGEX_IS024_Stack& stack){
}
res += " }";
return res;
}
}
std::string slots_to_str(const REGEX_IS024_CONTEXT& ctx){
std::string slots_to_str(const VMContext& ctx){
if (!ctx.initialized)
return "uninitialized";
std::string READ_slots;
@ -41,18 +42,19 @@ std::string slots_to_str(const REGEX_IS024_CONTEXT& ctx){
stack_to_str(ctx.READ_halted_stack_new_second).c_str(),
stack_to_str(ctx.READ_halted_stack_old).c_str(), stack_to_str(ctx.FORK_halted_stack).c_str());
return buf;
}
}
void debug_print_context(const REGEX_IS024_CONTEXT& ctx, const char* place) {
void debug_print_context(const VMContext& ctx, const char* place) {
printf("== DEBUG `%s` ==\n", place);
printf("Active thread: %s, sifting_with: %s, match: %s\n%s\n",
thread_to_str(ctx.active_thread).c_str(),
ctx.sifting_with ? thread_to_str(*ctx.sifting_with).c_str() : "NO", thread_to_str(ctx.matched_thread).c_str(),
slots_to_str(ctx).c_str());
}
}
void debug_print_thread(const REGEX_IS024_Thread& thr, const char *place) {
void debug_print_thread(const Thread& thr, const char *place) {
printf("== DEBUG `%s` ==\n", place);
printf("This thread: %s\n", thread_to_str(thr).c_str());
}
}

View File

@ -4,8 +4,10 @@
#include <libregexis024vm/libregexis024vm.h>
#include <libregexis024vm/instruction_implementation.h>
void debug_print_context(const REGEX_IS024_CONTEXT& ctx, const char* place);
namespace regexis024 {
void debug_print_context(const VMContext& ctx, const char* place);
void debug_print_thread(const REGEX_IS024_Thread& thr, const char *place);
void debug_print_thread(const Thread& thr, const char *place);
}
#endif

View File

@ -1,7 +1,8 @@
#include <libregexis024fa/codeset.h>
#include <assert.h>
codeset_t invert_set(const codeset_t &X) {
namespace regexis024 {
codeset_t invert_set(const codeset_t &X) {
if (X.empty())
return {{0, UINT32_MAX}};
codeset_t res;
@ -13,7 +14,7 @@ codeset_t invert_set(const codeset_t &X) {
if (X.back().second != UINT32_MAX)
res.emplace_back(X.back().second + 1, UINT32_MAX);
return res;
}
}
#define elA (A[i])
#define elB (B[j])
@ -23,7 +24,7 @@ codeset_t invert_set(const codeset_t &X) {
#define Aended (i == An)
#define Bended (j == Bn)
codeset_t merge_sets(const codeset_t &A, const codeset_t &B) {
codeset_t merge_sets(const codeset_t &A, const codeset_t &B) {
codeset_t res;
prepare
std::pair<uint32_t, uint32_t> cur;
@ -63,9 +64,9 @@ codeset_t merge_sets(const codeset_t &A, const codeset_t &B) {
}
}
return res;
}
}
codeset_t intersect_sets(const codeset_t &A, const codeset_t &B) {
codeset_t intersect_sets(const codeset_t &A, const codeset_t &B) {
codeset_t res;
prepare
while (true){
@ -82,34 +83,34 @@ codeset_t intersect_sets(const codeset_t &A, const codeset_t &B) {
Binc;
}
return res;
}
}
codeset_t subtract_sets(const codeset_t &A, const codeset_t &B) {
codeset_t subtract_sets(const codeset_t &A, const codeset_t &B) {
return intersect_sets(A, invert_set(B));
}
}
bool is_inside(uint32_t start, uint32_t end, codeset_t &X) {
bool is_inside(uint32_t start, uint32_t end, codeset_t &X) {
for (auto& p: X){
if (p.first <= start && end <= p.second)
return true;
assert(end < p.first || p.second < start);
}
return false;
}
}
codeset_t set_add_char(const codeset_t& X, uint32_t cp) {
codeset_t set_add_char(const codeset_t& X, uint32_t cp) {
return merge_sets(X, {{cp, cp}});
}
}
codeset_t set_add_range(const codeset_t& X, uint32_t start, uint32_t end) {
codeset_t set_add_range(const codeset_t& X, uint32_t start, uint32_t end) {
return merge_sets(X, {{start, end}});
}
}
codeset_t codeset_of_one_char(uint32_t ch) {
codeset_t codeset_of_one_char(uint32_t ch) {
return codeset_t({{ch, ch}});
}
}
std::string stringifyCodesetBase10(const codeset_t& CS) {
std::string stringifyCodesetBase10(const codeset_t& CS) {
std::string cs;
for (auto p: CS) {
if (!cs.empty())
@ -117,4 +118,5 @@ std::string stringifyCodesetBase10(const codeset_t& CS) {
cs += std::to_string(p.first) + "-" + std::to_string(p.second);
}
return cs;
}
}

View File

@ -6,22 +6,24 @@
#include <stdint.h>
#include <string>
typedef std::vector<std::pair<uint32_t, uint32_t>> codeset_t;
namespace regexis024 {
typedef std::vector<std::pair<uint32_t, uint32_t>> codeset_t;
codeset_t invert_set(const codeset_t& X);
codeset_t merge_sets(const codeset_t& A, const codeset_t& B);
codeset_t intersect_sets(const codeset_t& A, const codeset_t& B);
codeset_t subtract_sets(const codeset_t& A, const codeset_t& B);
codeset_t invert_set(const codeset_t& X);
codeset_t merge_sets(const codeset_t& A, const codeset_t& B);
codeset_t intersect_sets(const codeset_t& A, const codeset_t& B);
codeset_t subtract_sets(const codeset_t& A, const codeset_t& B);
/* Aborts if segment in question hit the edge (unsafe function) */
bool is_inside(uint32_t start, uint32_t end, codeset_t& X);
/* Aborts if segment in question hit the edge (unsafe function) */
bool is_inside(uint32_t start, uint32_t end, codeset_t& X);
codeset_t set_add_char(const codeset_t& X, uint32_t cp);
codeset_t set_add_range(const codeset_t& X, uint32_t start, uint32_t end);
codeset_t set_add_char(const codeset_t& X, uint32_t cp);
codeset_t set_add_range(const codeset_t& X, uint32_t start, uint32_t end);
codeset_t codeset_of_one_char(uint32_t ch);
codeset_t codeset_of_one_char(uint32_t ch);
#define codeset_of_all codeset_t({{0, UINT32_MAX}})
std::string stringifyCodesetBase10(const codeset_t& CS);
std::string stringifyCodesetBase10(const codeset_t& CS);
}
#endif //LIBREGEXIS024_CODESET_H

View File

@ -2,13 +2,14 @@
#include <assert.h>
ColoredCodesetSegment::ColoredCodesetSegment(uint32_t color, uint32_t right_code): color(color), right_code(right_code) {}
namespace regexis024 {
ColoredCodesetSegment::ColoredCodesetSegment(uint32_t color, uint32_t right_code): color(color), right_code(right_code) {}
ColoredCodesetSegmentList::ColoredCodesetSegmentList() {
ColoredCodesetSegmentList::ColoredCodesetSegmentList() {
first = new ColoredCodesetSegment(0, UINT32_MAX);
}
}
void ColoredCodesetSegmentList::replace_myself(const ColoredCodesetSegmentList &other) {
void ColoredCodesetSegmentList::replace_myself(const ColoredCodesetSegmentList &other) {
assert(other.first);
ColoredCodesetSegment** in_cur = &first;
ColoredCodesetSegment* in_other = other.first;
@ -17,36 +18,36 @@ void ColoredCodesetSegmentList::replace_myself(const ColoredCodesetSegmentList &
in_cur = &((**in_cur).next);
in_other = in_other->next;
}
}
}
ColoredCodesetSegmentList::ColoredCodesetSegmentList(const ColoredCodesetSegmentList &other) {
ColoredCodesetSegmentList::ColoredCodesetSegmentList(const ColoredCodesetSegmentList &other) {
replace_myself(other);
}
}
void ColoredCodesetSegmentList::free_myself() {
void ColoredCodesetSegmentList::free_myself() {
ColoredCodesetSegment* cur = first;
while (cur) {
ColoredCodesetSegment* nxt = cur->next;
delete cur;
cur = nxt;
}
}
}
ColoredCodesetSegmentList::~ColoredCodesetSegmentList() {
ColoredCodesetSegmentList::~ColoredCodesetSegmentList() {
free_myself();
}
}
ColoredCodesetSegmentList& ColoredCodesetSegmentList::operator=(const ColoredCodesetSegmentList &other) {
ColoredCodesetSegmentList& ColoredCodesetSegmentList::operator=(const ColoredCodesetSegmentList &other) {
free_myself();
replace_myself(other);
return *this;
}
}
ColoredCodeset::ColoredCodeset(uint64_t dummy_n): DummyN(dummy_n) {
ColoredCodeset::ColoredCodeset(uint64_t dummy_n): DummyN(dummy_n) {
requests = {{}};
}
}
void ColoredCodeset::split_phase(const codeset_t &X) {
void ColoredCodeset::split_phase(const codeset_t &X) {
uint32_t cA = 0;
ColoredCodesetSegment* cur_seg = list.first;
@ -106,9 +107,9 @@ void ColoredCodeset::split_phase(const codeset_t &X) {
advance_old();
}
}
}
}
void ColoredCodeset::apply_divisor(const codeset_t &X) {
void ColoredCodeset::apply_divisor(const codeset_t &X) {
split_phase(X);
size_t X_id = nxt_request_id++;
size_t m = requests.size();
@ -154,9 +155,9 @@ void ColoredCodeset::apply_divisor(const codeset_t &X) {
cur = cur->next;
}
}
}
}
void ColoredCodeset::get_splits_of_non_dummy(std::vector<codeset_t> &res_input,
void ColoredCodeset::get_splits_of_non_dummy(std::vector<codeset_t> &res_input,
std::vector<std::vector<size_t>> &res_color_to_requests) {
size_t n = requests.size();
std::vector<ssize_t> nonclean_to_clean(n, -1);
@ -180,4 +181,5 @@ void ColoredCodeset::get_splits_of_non_dummy(std::vector<codeset_t> &res_input,
L = cur->right_code + 1;
cur = cur->next;
}
}
}

View File

@ -7,9 +7,9 @@
#include <libregexis024fa/codeset.h>
/* Used for determinizer. Nowhere else */
struct ColoredCodesetSegment {
namespace regexis024 {
/* Used for determinizer. Nowhere else */
struct ColoredCodesetSegment {
uint32_t color;
uint32_t right_code;
ColoredCodesetSegment* next = NULL;
@ -18,11 +18,11 @@ struct ColoredCodesetSegment {
bool divisor_on_left = false;
ColoredCodesetSegment(uint32_t color, uint32_t right_code);
};
};
/* Warning!!! This stupid class is OOM-unsafe!!!
/* Warning!!! This stupid class is OOM-unsafe!!!
* This is not an issue as far as you don't show any of it's instance to the user of libregexis024 */
struct ColoredCodesetSegmentList {
struct ColoredCodesetSegmentList {
ColoredCodesetSegment* first = NULL;
ColoredCodesetSegmentList();
@ -38,10 +38,10 @@ struct ColoredCodesetSegmentList {
~ColoredCodesetSegmentList();
ColoredCodesetSegmentList& operator=(const ColoredCodesetSegmentList& other);
};
};
/* Highly unoptimized algorithm on this data structure O(C^2) time*/
class ColoredCodeset {
/* Highly unoptimized algorithm on this data structure O(C^2) time*/
class ColoredCodeset {
ColoredCodesetSegmentList list;
/* Size of this vector is equal to the number of colors */
std::vector<std::vector<size_t>> requests;
@ -49,7 +49,7 @@ class ColoredCodeset {
size_t nxt_request_id = 0;
void split_phase(const codeset_t& X);
public:
public:
/* First dummy_n split requests will be viewed as 'dummy requests', when complete map of splits is requested,
* colors that are registed indide only dummy requests won't be returned. */
ColoredCodeset(uint64_t dummy_n);
@ -60,7 +60,7 @@ public:
/* Returned 'requests' mapping will feature request id's with DummyN substituted from them */
void get_splits_of_non_dummy(std::vector<codeset_t>& res_input,
std::vector<std::vector<size_t>>& res_color_to_requests);
};
};
}
#endif

View File

@ -7,7 +7,8 @@
// #include <debugging_regexis024/debug_through_graphviz.h>
// #endif
REGEX_IS024_FA_FirstStageFixInfo first_stage_fix_fa(FA_Container& sourceFa, FA_Container& resultFa) {
namespace regexis024 {
REGEX_IS024_FA_FirstStageFixInfo first_stage_fix_fa(FA_Container& sourceFa, FA_Container& resultFa) {
assert(sourceFa.start);
REGEX_IS024_FA_FirstStageFixInfo info;
@ -140,19 +141,19 @@ REGEX_IS024_FA_FirstStageFixInfo first_stage_fix_fa(FA_Container& sourceFa, FA_C
d->search_mark = -1;
}
return info;
}
}
FA_NodeOfOneCharRead* generate_alt_ending(const codeset_t& restriction, FA_Container& fa){
FA_NodeOfOneCharRead* generate_alt_ending(const codeset_t& restriction, FA_Container& fa){
FA_NodeOfOneCharRead* n1 = fa.makeOneCharRead(restriction, true);
FA_NodeOfMatch* n2 = fa.makeMatch();
n2->ext_filter_added = true; // Won't actually be used
reattach_fa_node_edge(&(n1->nxt_node), n2);
return n1;
}
}
void regular_second_stage_fix(const FA_Container& sourceFa, FA_Container& resultFa,
void regular_second_stage_fix(const FA_Container& sourceFa, FA_Container& resultFa,
const REGEX_IS024_FA_FirstStageFixInfo &info1)
{
{
assert(resultFa.all.empty() && !resultFa.start);
if (!sourceFa.start)
return;
@ -192,4 +193,5 @@ void regular_second_stage_fix(const FA_Container& sourceFa, FA_Container& result
FA_NodeOfOneCharRead* ns = resultFa.makeOneCharRead(codeset_of_all, true);
yay_new_start(resultFa, ns);
}
}
}

View File

@ -3,16 +3,18 @@
#include "finite_automaton.h"
struct REGEX_IS024_FA_FirstStageFixInfo{
namespace regexis024 {
struct REGEX_IS024_FA_FirstStageFixInfo{
bool fed_chars_extend_one_left = false;
bool fed_chars_extend_one_right = false;
};
};
/* Will look for look_one_ahead nodes and apply their filter to reading filters ahead *
/* Will look for look_one_ahead nodes and apply their filter to reading filters ahead *
* sourceFa will be ruined. The output will be in resultFa */
REGEX_IS024_FA_FirstStageFixInfo first_stage_fix_fa(FA_Container& sourceFa, FA_Container& resultFa);
REGEX_IS024_FA_FirstStageFixInfo first_stage_fix_fa(FA_Container& sourceFa, FA_Container& resultFa);
void regular_second_stage_fix(const FA_Container& sourceFa, FA_Container& resultFa,
void regular_second_stage_fix(const FA_Container& sourceFa, FA_Container& resultFa,
const REGEX_IS024_FA_FirstStageFixInfo &info1);
}
#endif //LIBREGEXIS024_FA_FIRST_STAGE_FIX_H

View File

@ -1,6 +1,5 @@
#include <libregexis024fa/fa_make_deterministic.h>
#include <libregexis024fa/misc_fa_funcs.h>
#include <libregexis024vm/utils.h> /* to get exitf */
#include <assert.h>
#include <libregexis024fa/tracking_fa_nodes.h>
#include <vector>
@ -17,38 +16,38 @@
#define PR_DEB
#endif
/* debug nonsence */
void input_fa_assert(const FA_Container& fa){
namespace regexis024 {
/* debug nonsence */
void input_fa_assert(const FA_Container& fa){
assert(fa.start);
for (FA_Node* node: fa.all){
if (node->type == one_char_read){
assert(!dynamic_cast<FA_NodeOfOneCharRead*>(node)->second_ns);
} else if (node->type == look_one_ahead ||
node->type == det_char_crossroads){
exitf("not allowed at this stage\n");
} else if (node->type == look_one_ahead || node->type == det_char_crossroads) {
assert(false);
}
}
}
}
struct OperHistoryNodeTransition {
struct OperHistoryNodeTransition {
TrackingOperationInFa op;
size_t u;
OperHistoryNodeTransition(const TrackingOperationInFa &op, size_t u): op(op), u(u) {}
};
};
struct OperHistoryNode {
struct OperHistoryNode {
std::vector<OperHistoryNodeTransition> next;
/* When it is part of clean history, this */
std::vector<uint64_t> compressed_selarr;
std::vector<uint64_t> raisin;
OperHistoryNode() = default;
};
};
/* This object can describe an empty superstate (needed to describe clean history nodes without raisin)
/* This object can describe an empty superstate (needed to describe clean history nodes without raisin)
* If det_stops is empty, interpret it as empty superstate */
struct SuperState {
struct SuperState {
std::vector<uint64_t> sorted_raisin;
std::vector<uint64_t> double_compressed_selarr;
@ -74,33 +73,33 @@ struct SuperState {
return "sorted_raisin: {" + f1_raisin + "}, double_comp_selarr: {" + f2_selarr + "}";
}
#endif
};
};
struct CleanOperHistoryNode {
struct CleanOperHistoryNode {
std::vector<OperHistoryNodeTransition> next;
SuperState exit;
};
};
struct SelarrCompressionScheme {
struct SelarrCompressionScheme {
size_t SN1, SN2 = 0, SN3 = 0;
std::vector<int32_t> S1_to_S2;
std::vector<regex_tai_t> S2_to_sifter;
std::vector<regex_tai_t> S3_to_sifter;
std::vector<tai_t> S2_to_sifter;
std::vector<tai_t> S3_to_sifter;
const RegexPriorityTable& sifter;
SelarrCompressionScheme(size_t sn1, const RegexPriorityTable &sifter) : SN1(sn1), sifter(sifter) {
assert(sifter.size() <= UINT32_MAX);
S1_to_S2.assign(SN1, -1);
for (regex_tai_t i = 0; i < sifter.size(); i++) {
for (tai_t i = 0; i < sifter.size(); i++) {
auto& act = sifter[i].pos;
regex_tai_t first_on_s2 = S2_to_sifter.size();
tai_t first_on_s2 = S2_to_sifter.size();
S2_to_sifter.push_back(i);
S1_to_S2[act.first] = first_on_s2;
if (act.type != tracking_var_types::dot_cur_pos) {
S3_to_sifter.push_back(i);
}
if (act.type == tracking_var_types::range) {
regex_tai_t second_on_s2 = S2_to_sifter.size();
tai_t second_on_s2 = S2_to_sifter.size();
S2_to_sifter.push_back(i);
S1_to_S2[act.second] = second_on_s2;
}
@ -110,9 +109,9 @@ struct SelarrCompressionScheme {
assert(SN3 <= SN2 && SN2 <= SN1 && SN1 <= UINT16_MAX);
}
};
};
std::vector<uint64_t> compress_compressed_selarr(const std::vector<uint64_t>& S2,
std::vector<uint64_t> compress_compressed_selarr(const std::vector<uint64_t>& S2,
const SelarrCompressionScheme& cmp) {
std::vector<uint64_t> S3(cmp.SN3);
for (size_t i = 0; i < cmp.SN3; i++) {
@ -127,9 +126,9 @@ std::vector<uint64_t> compress_compressed_selarr(const std::vector<uint64_t>& S2
}
}
return S3;
}
}
bool compressed_selarr_A_outranks_B(const std::vector<uint64_t>& A, const std::vector<uint64_t>& B,
bool compressed_selarr_A_outranks_B(const std::vector<uint64_t>& A, const std::vector<uint64_t>& B,
const SelarrCompressionScheme& cmp) {
for (const RegexPriorityTableAction& act: cmp.sifter) {
uint64_t valA = A[cmp.S1_to_S2[act.pos.first]];
@ -145,11 +144,11 @@ bool compressed_selarr_A_outranks_B(const std::vector<uint64_t>& A, const std::v
return (valA < valB) == act.minimize;
}
return false;
}
}
/* Beacuse of the way wash_history_bush builds this structure, root is te last node.
/* Beacuse of the way wash_history_bush builds this structure, root is te last node.
* rankdir is from left to right (guaranteed). Can be empty if original history contained no raisin */
struct RaisinBush {
struct RaisinBush {
std::vector<CleanOperHistoryNode> clean_history;
ssize_t start = -1;
@ -201,9 +200,9 @@ struct RaisinBush {
printLines(wrapWithBox(text));
}
#endif
};
};
void wash_history_bush(const std::vector<OperHistoryNode>& history, RaisinBush& answer,
void wash_history_bush(const std::vector<OperHistoryNode>& history, RaisinBush& answer,
const SelarrCompressionScheme& cmp) {
assert(!history.empty());
std::vector<bool> has_raisin(history.size());
@ -252,14 +251,14 @@ void wash_history_bush(const std::vector<OperHistoryNode>& history, RaisinBush&
answer.start = dirty_to_clean[0];
}
}
}
/* If is_it_after_read is false, unknown selarr range variable border and cur pos are evaluated to 0.
/* If is_it_after_read is false, unknown selarr range variable border and cur pos are evaluated to 0.
* Otherwise, cur pos considered to be greater than previous values of selarr ange variable boundaries */
void building_detour(const SelarrCompressionScheme& cmp,
void building_detour(const SelarrCompressionScheme& cmp,
const std::vector<uint64_t>& outer_selarr, const std::vector<FA_Node*>& zeroeps, const codeset_t& I,
RaisinBush& answer, bool is_it_after_read)
{
{
#ifdef PR_DEB
printf("Det Debug: build_detour started with zeroeps:{");
for (FA_Node* node: zeroeps)
@ -368,7 +367,7 @@ void building_detour(const SelarrCompressionScheme& cmp,
val2[key_s2] = tv->imm_value;
}
}
add_history_update(TrackingOperationInFa(tv->operation, tv->key, tv->imm_value), Hop, Hv);
add_history_update(TrackingOperationInFa{tv->operation, tv->key, tv->imm_value}, Hop, Hv);
} else if (v->type == track_array_mov_halfinvariant) {
FA_NodeOfTrackArrayMovHalfinvariant* tv = dynamic_cast<FA_NodeOfTrackArrayMovHalfinvariant*>(v);
if (isSelarrOpcode(tv->operation)) {
@ -383,7 +382,7 @@ void building_detour(const SelarrCompressionScheme& cmp,
}
}
}
add_history_update(TrackingOperationInFa(tv->operation, tv->key), Hop, Hv);
add_history_update(TrackingOperationInFa{tv->operation, tv->key}, Hop, Hv);
}
} else if (v->type == match || v->type == one_char_read) {
// Determinization stop
@ -414,22 +413,22 @@ void building_detour(const SelarrCompressionScheme& cmp,
}
/* Packaging the answer (we do a little bit of dfs here) */
wash_history_bush(history, answer, cmp);
}
}
void update_had_to_fork_status(const RaisinBush& bush, int& had_to_fork) {
void update_had_to_fork_status(const RaisinBush& bush, int& had_to_fork) {
for (const CleanOperHistoryNode& node: bush.clean_history) {
if (node.next.size() > 1 || (!node.next.empty() && !node.exit.empty())) {
had_to_fork = 1;
return;
}
}
}
}
typedef size_t superstate_id_t;
typedef size_t superstate_id_t;
typedef std::vector<std::pair<FA_Node**, superstate_id_t>> homework_t;
typedef std::vector<std::pair<FA_Node**, superstate_id_t>> homework_t;
struct LessSuperState {
struct LessSuperState {
bool operator()(const SuperState& A, const SuperState& B) const {
std::less<std::vector<uint64_t>> f1L;
if (f1L(A.sorted_raisin, B.sorted_raisin))
@ -438,17 +437,17 @@ struct LessSuperState {
return false;
return f1L(A.double_compressed_selarr, B.double_compressed_selarr);
}
};
};
struct GlobalDetourProgress {
struct GlobalDetourProgress {
std::map<SuperState, superstate_id_t, LessSuperState> superstates;
/* Each element is a root of some megabush in resFa */
std::vector<FA_Node*> superstate_megabush_constructed;
std::vector<SuperState> todo_superstaes;
};
};
/* If x was not previously achieved, it will also add it to t o d o list of global detour */
superstate_id_t convertSuperstateToId(const SuperState& x, GlobalDetourProgress& gdp) {
/* If x was not previously achieved, it will also add it to t o d o list of global detour */
superstate_id_t convertSuperstateToId(const SuperState& x, GlobalDetourProgress& gdp) {
if (gdp.superstates.count(x)) {
return gdp.superstates[x];
}
@ -457,13 +456,13 @@ superstate_id_t convertSuperstateToId(const SuperState& x, GlobalDetourProgress&
gdp.todo_superstaes.push_back(x);
gdp.superstate_megabush_constructed.push_back(NULL);
return n;
}
}
FA_Node* build_dead_end(FA_Container& resFa) {
FA_Node* build_dead_end(FA_Container& resFa) {
return resFa.makeForking();
}
}
void build_bush(const RaisinBush& alpha, FA_Node** sowing_location, FA_Container& resFa,
void build_bush(const RaisinBush& alpha, FA_Node** sowing_location, FA_Container& resFa,
homework_t& homework, GlobalDetourProgress& gdp) {
size_t n = alpha.clean_history.size();
if (n == 0) {
@ -504,9 +503,9 @@ void build_bush(const RaisinBush& alpha, FA_Node** sowing_location, FA_Container
reattach_fa_node_edge(sl, forker);
}
}
}
}
ColoredCodeset get_pretreated_cc(FA_Container& sourceFa) {
ColoredCodeset get_pretreated_cc(FA_Container& sourceFa) {
std::set<codeset_t> little_insects;
for (FA_Node* v: sourceFa.all) {
if (v->type == look_one_behind) {
@ -518,12 +517,12 @@ ColoredCodeset get_pretreated_cc(FA_Container& sourceFa) {
pretreated_cc.apply_divisor(cs);
}
return pretreated_cc;
}
}
// todo add a check on size of dfa
void try_determinize_fa(FA_Container &sourceFa, const RegexPriorityTable &sifter, regex_tai_t selarr_sz,
// todo add a check on size of dfa
void try_determinize_fa(FA_Container &sourceFa, const RegexPriorityTable &sifter, tai_t selarr_sz,
const REGEX_IS024_FA_FirstStageFixInfo &info1, FA_Container &resFa, int &error, int& had_to_fork)
{
{
/* During execuion, i will create pointers to field res.start and store them (inside the scope of this function)
* Luckily res argument is already immovable in this scope. */
error = 0;
@ -661,5 +660,5 @@ void try_determinize_fa(FA_Container &sourceFa, const RegexPriorityTable &sifter
for (auto& p: homework) {
reattach_fa_node_edge(p.first, gdp.superstate_megabush_constructed[p.second]);
}
}
}

View File

@ -4,7 +4,9 @@
#include <libregexis024fa/fa_first_stage_fix.h>
#include <libregexis024fa/selarr_priority_table.h>
void try_determinize_fa(FA_Container &sourceFa, const RegexPriorityTable &sifter, regex_tai_t selarr_sz,
namespace regexis024 {
void try_determinize_fa(FA_Container &sourceFa, const RegexPriorityTable &sifter, tai_t selarr_sz,
const REGEX_IS024_FA_FirstStageFixInfo &info1, FA_Container &resFa, int &error, int& had_to_fork);
}
#endif //LIBREGEXIS024_FA_MAKE_DETERMINISTIC_H

View File

@ -2,140 +2,142 @@
#include <libregexis024vm/utils.h>
#include <assert.h>
bool FA_Node::empty() {
namespace regexis024 {
bool FA_Node::empty() {
return type != one_char_read && type != det_char_crossroads;
}
}
void FA_Node::apply_lookahead_restriction(const codeset_t &restriction) {}
void FA_Node::apply_lookahead_restriction(const codeset_t &restriction) {}
void FA_Node::reAdd_references() {
void FA_Node::reAdd_references() {
for (FA_Node** nxtPtr: get_all_transitions()){
if (*nxtPtr)
(**nxtPtr).refs++;
}
}
}
std::vector<FA_Node **> FA_Node::get_all_transitions() {
std::vector<FA_Node **> FA_Node::get_all_transitions() {
return {};
}
}
std::vector<FA_Node **> FA_Node::get_all_empty_valid_transitions() {
std::vector<FA_Node **> FA_Node::get_all_empty_valid_transitions() {
return {};
}
}
std::vector<FA_Node **> FA_NodePathPart::get_all_transitions() {
std::vector<FA_Node **> FA_NodePathPart::get_all_transitions() {
return {&nxt_node};
}
}
std::vector<FA_Node **> FA_NodePathPart::get_all_empty_valid_transitions() {
std::vector<FA_Node **> FA_NodePathPart::get_all_empty_valid_transitions() {
if (nxt_node)
return {&nxt_node};
return {};
}
}
FA_NodeOfMatch::FA_NodeOfMatch() {type = match;}
FA_NodeOfMatch::FA_NodeOfMatch() {type = match;}
void FA_NodeOfMatch::apply_lookahead_restriction(const codeset_t &restriction) {
void FA_NodeOfMatch::apply_lookahead_restriction(const codeset_t &restriction) {
ext_filter_added = true;
pending_filter = restriction;
}
}
FA_NodeOfOneCharRead::FA_NodeOfOneCharRead(const codeset_t &filter, bool second_namespace) : filter(filter),
FA_NodeOfOneCharRead::FA_NodeOfOneCharRead(const codeset_t &filter, bool second_namespace) : filter(filter),
second_ns(second_namespace) { type = one_char_read;}
void FA_NodeOfOneCharRead::apply_lookahead_restriction(const codeset_t &restriction) {
void FA_NodeOfOneCharRead::apply_lookahead_restriction(const codeset_t &restriction) {
filter = intersect_sets(filter, restriction);
}
}
std::vector<FA_Node **> FA_NodeOfOneCharRead::get_all_empty_valid_transitions() {
std::vector<FA_Node **> FA_NodeOfOneCharRead::get_all_empty_valid_transitions() {
return {};
}
}
FA_NodeOfForking::FA_NodeOfForking() {type = forking;}
FA_NodeOfForking::FA_NodeOfForking() {type = forking;}
std::vector<FA_Node **> FA_NodeOfForking::get_all_empty_valid_transitions() {
std::vector<FA_Node **> FA_NodeOfForking::get_all_empty_valid_transitions() {
std::vector<FA_Node**> res;
for (size_t i = 0; i < nxt_options.size(); i++)
if (nxt_options[i])
res.push_back(&nxt_options[i]);
return res;
}
}
std::vector<FA_Node **> FA_NodeOfForking::get_all_transitions() {
std::vector<FA_Node **> FA_NodeOfForking::get_all_transitions() {
std::vector<FA_Node**> res;
for (size_t i = 0; i < nxt_options.size(); i++)
res.push_back(&nxt_options[i]);
return res;
}
}
FA_NodeOfLookOneBehind::FA_NodeOfLookOneBehind(const codeset_t &filter) : filter(filter) {type = look_one_behind;}
FA_NodeOfLookOneBehind::FA_NodeOfLookOneBehind(const codeset_t &filter) : filter(filter) {type = look_one_behind;}
FA_NodeOfLookOneAhead::FA_NodeOfLookOneAhead(const codeset_t &restriction) : restriction(restriction) {
FA_NodeOfLookOneAhead::FA_NodeOfLookOneAhead(const codeset_t &restriction) : restriction(restriction) {
type = look_one_ahead;
}
}
FA_NodeOfTrackArrayMovImm::FA_NodeOfTrackArrayMovImm(regex024_opcode operation, uint16_t key, uint64_t immValue) :
FA_NodeOfTrackArrayMovImm::FA_NodeOfTrackArrayMovImm(opcode_t operation, uint16_t key, uint64_t immValue) :
operation(operation), key(key), imm_value(immValue) {type = track_array_mov_imm;}
//
//
FA_NodeOfTrackArrayMovHalfinvariant::FA_NodeOfTrackArrayMovHalfinvariant(regex024_opcode operation, uint16_t key):
FA_NodeOfTrackArrayMovHalfinvariant::FA_NodeOfTrackArrayMovHalfinvariant(opcode_t operation, uint16_t key):
operation(operation), key(key){type = track_array_mov_halfinvariant;}
//
//
void FA_NodeOfDetCharCrossroads::apply_lookahead_restriction(const codeset_t &restriction) {
exitf("What?? Oh, no, no. I am NOT doing it");
}
void FA_NodeOfDetCharCrossroads::apply_lookahead_restriction(const codeset_t &restriction) {
assert(false);
}
FA_NodeOfDetCharCrossroads::FA_NodeOfDetCharCrossroads(const std::vector<DFA_CrossroadPath> &crossroads)
FA_NodeOfDetCharCrossroads::FA_NodeOfDetCharCrossroads(const std::vector<DFA_CrossroadPath> &crossroads)
: crossroads(crossroads) {type = det_char_crossroads;}
std::vector<FA_Node **> FA_NodeOfDetCharCrossroads::get_all_empty_valid_transitions() {
std::vector<FA_Node **> FA_NodeOfDetCharCrossroads::get_all_empty_valid_transitions() {
return {};
}
}
std::vector<FA_Node **> FA_NodeOfDetCharCrossroads::get_all_transitions() {
std::vector<FA_Node **> FA_NodeOfDetCharCrossroads::get_all_transitions() {
std::vector<FA_Node**> res;
for (auto& tr: crossroads)
res.push_back(&tr.nxt_node);
return res;
}
}
/* If transferring ownership of node to container has failed, node is freed (which means it is ivalidated)
/* If transferring ownership of node to container has failed, node is freed (which means it is ivalidated)
* If this semi-ownership transfer succeded (no std::bad_alloc), then node is still valid to use, and at the end
* of FA_Container lifetime it is guaranteed to be deleted
*/
void FA_Container::registerNew(FA_Node *node) {
void FA_Container::registerNew(FA_Node *node) {
try {
node->nodeId = (int64_t)all.size();
all.push_back(node);
} catch (const std::bad_alloc& ba) {
} catch (const std::exception& ba) {
delete node;
throw;
}
}
}
DFA_CrossroadPath::DFA_CrossroadPath(const codeset_t &input, FA_Node *nxt_node): input(input),nxt_node(nxt_node) {}
//
DFA_CrossroadPath::DFA_CrossroadPath(const codeset_t &input, FA_Node *nxt_node): input(input),nxt_node(nxt_node) {}
//
FA_Container::~FA_Container() {
FA_Container::~FA_Container() {
for (FA_Node* n: all)
delete n;
}
}
#define bs(name, args, params) \
FA_NodeOf ## name *FA_Container::make ## name(args) { \
FA_NodeOf ## name *node = new FA_NodeOf ## name(params); \
registerNew(node); \
return node; \
}
FA_NodeOf ## name *FA_Container::make ## name(args) { \
FA_NodeOf ## name *node = new FA_NodeOf ## name(params); \
registerNew(node); \
return node; \
}
#define COMMA ,
bs(Match, , )
bs(OneCharRead, const codeset_t& filter COMMA bool second_namespace, filter COMMA second_namespace)
bs(Forking, , )
bs(LookOneBehind, const codeset_t& filter, filter)
bs(LookOneAhead, const codeset_t& filter, filter)
bs(TrackArrayMovImm, regex024_opcode operation COMMA uint16_t key COMMA uint64_t immValue,
bs(Match, , )
bs(OneCharRead, const codeset_t& filter COMMA bool second_namespace, filter COMMA second_namespace)
bs(Forking, , )
bs(LookOneBehind, const codeset_t& filter, filter)
bs(LookOneAhead, const codeset_t& filter, filter)
bs(TrackArrayMovImm, opcode_t operation COMMA uint16_t key COMMA uint64_t immValue,
operation COMMA key COMMA immValue)
bs(TrackArrayMovHalfinvariant, regex024_opcode operation COMMA uint16_t key, operation COMMA key)
bs(DetCharCrossroads, ,{})
bs(TrackArrayMovHalfinvariant, opcode_t operation COMMA uint16_t key, operation COMMA key)
bs(DetCharCrossroads, ,{})
}

View File

@ -6,7 +6,8 @@
#include <libregexis024fa/codeset.h>
#include <libregexis024vm/vm_opcodes.h>
enum FA_Node_type: uint8_t {
namespace regexis024 {
enum FA_Node_type: uint8_t {
match,
one_char_read,
forking,
@ -16,9 +17,9 @@ enum FA_Node_type: uint8_t {
track_array_mov_halfinvariant,
/* Used for DFA */
det_char_crossroads,
};
};
struct FA_Node{
struct FA_Node{
size_t refs = 0;
/* If node is not in searched subset (at least yet), `search mark == -1`, otherwise
* it is an index (for that particular node) in the vector that captures all nodes in
@ -33,35 +34,35 @@ struct FA_Node{
void reAdd_references();
virtual ~FA_Node() = default;
virtual std::vector<FA_Node**> get_all_transitions();
};
};
struct FA_NodePathPart: public FA_Node{
struct FA_NodePathPart: public FA_Node{
FA_Node* nxt_node = NULL;
std::vector<FA_Node **> get_all_empty_valid_transitions() override;
std::vector<FA_Node **> get_all_transitions() override;
};
};
struct FA_NodeOfMatch: public FA_Node{
struct FA_NodeOfMatch: public FA_Node{
bool ext_filter_added = false;
codeset_t pending_filter;
explicit FA_NodeOfMatch();
void apply_lookahead_restriction(const codeset_t &restriction) override;
};
};
/* .type == one_char_read */
struct FA_NodeOfOneCharRead: public FA_NodePathPart{
/* .type == one_char_read */
struct FA_NodeOfOneCharRead: public FA_NodePathPart{
codeset_t filter;
bool second_ns = false;
FA_NodeOfOneCharRead(const codeset_t &filter, bool second_namespace);
void apply_lookahead_restriction(const codeset_t &restriction) override;
std::vector<FA_Node **> get_all_empty_valid_transitions() override;
};
};
/* .type == forking */
struct FA_NodeOfForking: public FA_Node{
/* .type == forking */
struct FA_NodeOfForking: public FA_Node{
/* Won't be modified after init (in regexp compilation into NFA) */
std::vector<FA_Node*> nxt_options;
int64_t stopId = -1;
@ -69,51 +70,51 @@ struct FA_NodeOfForking: public FA_Node{
explicit FA_NodeOfForking();
std::vector<FA_Node **> get_all_empty_valid_transitions() override;
std::vector<FA_Node **> get_all_transitions() override;
};
};
/* .type == look_one_behind */
struct FA_NodeOfLookOneBehind: public FA_NodePathPart{
/* .type == look_one_behind */
struct FA_NodeOfLookOneBehind: public FA_NodePathPart{
/* [0; UINT32_MAX] is equivalent to no filter */
codeset_t filter;
explicit FA_NodeOfLookOneBehind(const codeset_t &filter);
};
};
/* .type == look_one_ahead */
struct FA_NodeOfLookOneAhead: public FA_NodePathPart{
/* .type == look_one_ahead */
struct FA_NodeOfLookOneAhead: public FA_NodePathPart{
/* [0; UINT32_MAX] is equivalent to no restriction */
codeset_t restriction;
explicit FA_NodeOfLookOneAhead(const codeset_t &restriction);
};
};
/* .type == track_array_mov_imm */
struct FA_NodeOfTrackArrayMovImm: public FA_NodePathPart{
regex024_opcode operation;
/* .type == track_array_mov_imm */
struct FA_NodeOfTrackArrayMovImm: public FA_NodePathPart{
opcode_t operation;
uint16_t key;
uint64_t imm_value;
FA_NodeOfTrackArrayMovImm(regex024_opcode operation, uint16_t key, uint64_t immValue);
};
FA_NodeOfTrackArrayMovImm(opcode_t operation, uint16_t key, uint64_t immValue);
};
/* .type == track_array_mov_halfinvariant */
struct FA_NodeOfTrackArrayMovHalfinvariant: public FA_NodePathPart{
regex024_opcode operation;
/* .type == track_array_mov_halfinvariant */
struct FA_NodeOfTrackArrayMovHalfinvariant: public FA_NodePathPart{
opcode_t operation;
uint16_t key;
FA_NodeOfTrackArrayMovHalfinvariant(regex024_opcode operation, uint16_t key);
};
FA_NodeOfTrackArrayMovHalfinvariant(opcode_t operation, uint16_t key);
};
struct DFA_CrossroadPath{
struct DFA_CrossroadPath{
codeset_t input;
FA_Node* nxt_node = NULL;
DFA_CrossroadPath(const codeset_t &input, FA_Node *nxt_node);
DFA_CrossroadPath() = default;
};
};
/* .type == det_char_crossroads */
struct FA_NodeOfDetCharCrossroads: public FA_Node{
/* .type == det_char_crossroads */
struct FA_NodeOfDetCharCrossroads: public FA_Node{
std::vector<DFA_CrossroadPath> crossroads;
bool matching = false;
bool second_ns = false;
@ -122,9 +123,9 @@ struct FA_NodeOfDetCharCrossroads: public FA_Node{
void apply_lookahead_restriction(const codeset_t &restriction) override;
std::vector<FA_Node **> get_all_empty_valid_transitions() override;
std::vector<FA_Node **> get_all_transitions() override;
};
};
struct FA_Container{
struct FA_Container{
FA_Container(const FA_Container&) = delete;
FA_Container& operator=(const FA_Container&) = delete;
FA_Container() = default;
@ -139,11 +140,12 @@ struct FA_Container{
FA_NodeOfForking* makeForking();
FA_NodeOfLookOneBehind* makeLookOneBehind(const codeset_t& filter);
FA_NodeOfLookOneAhead* makeLookOneAhead(const codeset_t& filter);
FA_NodeOfTrackArrayMovImm* makeTrackArrayMovImm(regex024_opcode operation, uint16_t key, uint64_t immValue);
FA_NodeOfTrackArrayMovHalfinvariant* makeTrackArrayMovHalfinvariant(regex024_opcode operation, uint16_t key);
FA_NodeOfTrackArrayMovImm* makeTrackArrayMovImm(opcode_t operation, uint16_t key, uint64_t immValue);
FA_NodeOfTrackArrayMovHalfinvariant* makeTrackArrayMovHalfinvariant(opcode_t operation, uint16_t key);
FA_NodeOfDetCharCrossroads* makeDetCharCrossroads();
~FA_Container();
};
};
}
#endif //LIBREGEXIS024_FINITE_AUTOMATON_H

View File

@ -5,11 +5,11 @@
#include <libregexis024fa/graph_to_bytecode/filter.h>
namespace regexis024 {
#define nonthrowing_assert(expr) if (!(expr)) {error = -1; return; }
void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_bookmarks& bookmark_manager,
void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_bookmarks& bookmark_manager,
size_t& first_read_ns, size_t& second_read_ns, size_t& fork_ss_ns, int& error)
{
{
bookmark_id_t node_start_bm_offset = bookmark_manager.new_range_of_bookmarks(fa.all.size());
std::vector<size_t> not_yet_dedicated_second_read_ns_ssids;
first_read_ns = 0;
@ -72,7 +72,7 @@ void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_b
}
if (nxt_options.size() >= 2) {
nonthrowing_assert(fork_ss_ns < UINT32_MAX);
regex_sslot_id_t sslot = fork_ss_ns++;
sslot_id_t sslot = fork_ss_ns++;
for (size_t i = 0; i + 1 < nxt_options.size(); i++) {
cmd_FORK(result, bookmark_manager, sslot, nodesBookmark(nxt_options[i]));
addBranching(nxt_options[i]);
@ -114,4 +114,5 @@ void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_b
for (size_t j = 0; j < not_yet_dedicated_second_read_ns_ssids.size(); j++) {
belated_sslot_id(result, not_yet_dedicated_second_read_ns_ssids[j], j + first_read_ns);
}
}
}

View File

@ -4,7 +4,9 @@
#include <libregexis024fa/finite_automaton.h>
#include <libregexis024fa/graph_to_bytecode/natural_compiler_utils.h>
void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_bookmarks& bookmark_manager,
namespace regexis024 {
void compilation_core(std::vector<uint8_t>& result, FA_Container& fa, explicit_bookmarks& bookmark_manager,
size_t& first_read_ns, size_t& second_read_ns, size_t& fork_ss_ns, int& error);
}
#endif

View File

@ -7,74 +7,75 @@
#include <libregexis024fa/graph_to_bytecode/core.h>
void write_priority_table_actions(std::vector<uint8_t>& result, RegexPriorityTable &priority_table) {
namespace regexis024 {
void write_priority_table_actions(std::vector<uint8_t>& result, RegexPriorityTable &priority_table) {
for (RegexPriorityTableAction& act: priority_table) {
if (act.pos.isForRange()) {
write_byte(result, regex024_opcodes::DDIST_RABX_SELARR);
write_byte(result, opcodes::DDIST_RABX_SELARR);
write_tai(result, act.pos.first);
write_tai(result, act.pos.second);
} else {
write_byte(result, regex024_opcodes::DMOV_RABX_SELARR);
write_byte(result, opcodes::DMOV_RABX_SELARR);
write_tai(result, act.pos.first);
}
write_byte(result, act.minimize ?
regex024_opcodes::SIFTPRIOR_MIN_RABX :
regex024_opcodes::SIFTPRIOR_MAX_RABX);
opcodes::SIFTPRIOR_MIN_RABX :
opcodes::SIFTPRIOR_MAX_RABX);
}
write_byte(result, opcodes::SIFT_DONE);
}
write_byte(result, regex024_opcodes::SIFT_DONE);
}
struct belate_initialization_parameters {
struct belate_initialization_parameters {
size_t todo_pos_read_ss_n;
size_t todo_pos_fork_ss_n;
size_t todo_pos_second_ns_size;
void complete_it(std::vector<uint8_t>& result,
regex_sslot_id_t first_read_ns, regex_sslot_id_t second_read_ns, regex_sslot_id_t fork_ss_ns)
sslot_id_t first_read_ns, sslot_id_t second_read_ns, sslot_id_t fork_ss_ns)
{
assert((uint64_t)first_read_ns + (uint64_t)second_read_ns <= UINT32_MAX);
belated_sslot_id(result, todo_pos_read_ss_n , first_read_ns + second_read_ns);
belated_sslot_id(result, todo_pos_fork_ss_n, fork_ss_ns);
belated_sslot_id(result, todo_pos_second_ns_size, second_read_ns);
}
};
};
/* when I compile initializational part of program, I don't yet know what to put in
/* when I compile initializational part of program, I don't yet know what to put in
* PARAM_READ_SS_NUMBER, PARAM_FORK_SS_NUMBER and MSG_FED_INPUT_EXTENDED (second namespace size).
* These values are belate. */
belate_initialization_parameters write_some_normal_initialization(std::vector<uint8_t>& result,
belate_initialization_parameters write_some_normal_initialization(std::vector<uint8_t>& result,
size_t selarr_size, const REGEX_IS024_FA_FirstStageFixInfo& info1)
{
{
belate_initialization_parameters todo;
write_byte(result, regex024_opcodes::PARAM_READ_SS_NUMBER);
write_byte(result, opcodes::PARAM_READ_SS_NUMBER);
todo.todo_pos_read_ss_n = result.size();
write_sslot_id(result, 0); // Belate
write_byte(result, regex024_opcodes::PARAM_FORK_SS_NUMBER);
write_byte(result, opcodes::PARAM_FORK_SS_NUMBER);
todo.todo_pos_fork_ss_n = result.size();
write_sslot_id(result, 0); // Belate
write_byte(result, regex024_opcodes::PARAM_SELARR_LEN);
write_byte(result, opcodes::PARAM_SELARR_LEN);
write_tai(result, selarr_size);
write_byte(result, regex024_opcodes::MSG_MULTISTART_ALLOWED);
write_byte(result, opcodes::MSG_MULTISTART_ALLOWED);
write_byte(result, 1);
write_byte(result, regex024_opcodes::MSG_FED_INPUT_EXTENDED);
write_byte(result, opcodes::MSG_FED_INPUT_EXTENDED);
write_byte(result, info1.fed_chars_extend_one_left ? 1 : 0);
write_byte(result, info1.fed_chars_extend_one_right ? 1 : 0);
todo.todo_pos_second_ns_size = result.size();
write_sslot_id(result, 0); // Belate
write_byte(result, regex024_opcodes::INIT);
write_byte(result, opcodes::INIT);
return todo;
}
}
void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result,
void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result,
FA_Container &fa, RegexPriorityTable &priority_table,
size_t selarr_size, const REGEX_IS024_FA_FirstStageFixInfo& info1, int& error)
{
{
error = 0;
explicit_bookmarks bookmark_manager;
@ -87,7 +88,7 @@ void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result,
write_priority_table_actions(result, priority_table);
bookmark_manager.land_bookmark(result, BM_after_sift);
write_byte(result, regex024_opcodes::PARAM_COLSIFTFUNC_SET);
write_byte(result, opcodes::PARAM_COLSIFTFUNC_SET);
bookmark_manager.write_unresolved_reference(result, BM_sift_function);
}
@ -99,4 +100,5 @@ void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result,
return;
init_param_todo.complete_it(result, first_read_ns, second_read_ns, fork_ss_ns);
bookmark_manager.finish(result);
}
}

View File

@ -7,8 +7,10 @@
#include <libregexis024fa/selarr_priority_table.h>
#include <libregexis024fa/fa_first_stage_fix.h>
void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result, FA_Container& fa, RegexPriorityTable& priority_table,
namespace regexis024 {
void compile_fa_to_regexis024_bytecode(std::vector<uint8_t>& result, FA_Container& fa, RegexPriorityTable& priority_table,
size_t selarr_size, const REGEX_IS024_FA_FirstStageFixInfo& info1, int& error);
}
#endif

View File

@ -4,42 +4,43 @@
#include <algorithm>
#include <libregexis024fa/graph_to_bytecode/writing_commands.h>
std::vector<FilterSegment> convert_to_compSeg(const std::vector<codeset_t>& crossroad_codesets)
{
namespace regexis024 {
std::vector<FilterSegment> convert_to_compSeg(const std::vector<codeset_t>& crossroad_codesets)
{
std::vector<FilterSegment> compSeg;
std::vector<FilterSegment> seg;
for (size_t i = 0; i < crossroad_codesets.size(); i++) {
for (auto& p: crossroad_codesets[i]) {
seg.emplace_back(i, p.first, p.second);
seg.push_back({(ssize_t)i, p.first, p.second});
}
}
std::sort(seg.begin(), seg.end(),
[](const FilterSegment& a, const FilterSegment& b)->bool{return a.L < b.L;});
if (seg.empty()) {
compSeg.emplace_back(-1, 0, UINT32_MAX);
compSeg.push_back({-1, 0, UINT32_MAX});
} else {
if (seg[0].L > 0)
compSeg.emplace_back(-1, 0, seg[0].L - 1);
compSeg.push_back({-1, 0, seg[0].L - 1});
size_t N = seg.size();
for (size_t i = 0; i + 1 < N; i++) {
compSeg.push_back(seg[i]);
assert(seg[i].R < seg[i + 1].L);
if (seg[i].R + 1 < seg[i + 1].L)
compSeg.emplace_back(-1, seg[i].R + 1, seg[i + 1].L - 1);
compSeg.push_back({-1, seg[i].R + 1, seg[i + 1].L - 1});
}
compSeg.push_back(seg.back());
if (seg.back().R < UINT32_MAX)
compSeg.emplace_back(-1, seg[N - 1].R + 1, UINT32_MAX);
compSeg.push_back({-1, seg[N - 1].R + 1, UINT32_MAX});
}
assert(!compSeg.empty());
return compSeg;
}
}
/* Return whether the resulting bytecode relies on me placing [0]'th node at the end */
void write_filter_exit(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
/* Return whether the resulting bytecode relies on me placing [0]'th node at the end */
void write_filter_exit(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
const std::vector<bookmark_id_t>& crossroad_marks,
ssize_t color, bool at_the_end, bool& relies_on_proper_ending)
{
{
if (color < 0) {
cmd_DIE(result);
} else if (color != 0 || !at_the_end) {
@ -47,12 +48,12 @@ void write_filter_exit(std::vector<uint8_t>& result, explicit_bookmarks& bookmar
} else {
relies_on_proper_ending = true;
}
}
}
// todo: use return value of this function
bool write_filter(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
// todo: use return value of this function
bool write_filter(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
const std::vector<codeset_t>& crossroad_codesets, const std::vector<bookmark_id_t>& crossroad_marks)
{
{
bool relies_on_proper_ending = false;
std::vector<FilterSegment> compSeg = convert_to_compSeg(crossroad_codesets);
@ -114,7 +115,5 @@ bool write_filter(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_man
}
}
return relies_on_proper_ending;
}
}
FilterSegment::FilterSegment(ssize_t color, uint32_t l, uint32_t r): color(color), L(l), R(r) {}
//

View File

@ -6,16 +6,17 @@
#include <libregexis024fa/codeset.h>
#include <libregexis024fa/graph_to_bytecode/natural_compiler_utils.h>
struct FilterSegment {
namespace regexis024 {
struct FilterSegment {
ssize_t color;
uint32_t L, R;
uint32_t L;
uint32_t R;
};
FilterSegment(ssize_t color, uint32_t l, uint32_t r);
};
/* Return whether user of function must place [0]'th option after the filter
/* Return whether user of function must place [0]'th option after the filter
* The filter can end up being written in such a way that the end will never be reached */
bool write_filter(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
bool write_filter(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager,
const std::vector<codeset_t>& crossroad_codesets, const std::vector<bookmark_id_t>& crossroad_marks);
}
#endif

View File

@ -2,93 +2,93 @@
#include <assert.h>
#include <libregexis024vm/vm_opcodes.h>
namespace regexis024 {
#define push_to_res_least_signif result.push_back(x & 0xffLU); x >>= 8
void write_byte(std::vector<uint8_t>& result, uint8_t x) {
void write_byte(std::vector<uint8_t>& result, uint8_t x) {
result.push_back(x);
}
}
void write_word(std::vector<uint8_t>& result, uint16_t x) {
void write_word(std::vector<uint8_t>& result, uint16_t x) {
push_to_res_least_signif; push_to_res_least_signif;
}
}
void write_doubleword(std::vector<uint8_t>& result, uint32_t x) {
void write_doubleword(std::vector<uint8_t>& result, uint32_t x) {
push_to_res_least_signif; push_to_res_least_signif; push_to_res_least_signif; push_to_res_least_signif;
}
}
void write_quadword(std::vector<uint8_t>& result, uint64_t x) {
void write_quadword(std::vector<uint8_t>& result, uint64_t x) {
for (int i = 0; i < 8; i++) {
push_to_res_least_signif;
}
}
}
#undef push_to_res_least_signif
#define put_belated_to_res assert(result[pos] == 0); result[pos++] = value & 0xffLU; value >>= 8
void belated_byte(std::vector<uint8_t>& result, size_t pos, uint8_t value) {
void belated_byte(std::vector<uint8_t>& result, size_t pos, uint8_t value) {
assert(pos < result.size());
result[pos] = value;
}
}
void belated_word(std::vector<uint8_t>& result, size_t pos, uint16_t value) {
void belated_word(std::vector<uint8_t>& result, size_t pos, uint16_t value) {
assert(pos + 2 <= result.size());
put_belated_to_res; put_belated_to_res;
}
}
void belated_doubleword(std::vector<uint8_t>& result, size_t pos, uint32_t value) {
void belated_doubleword(std::vector<uint8_t>& result, size_t pos, uint32_t value) {
assert(pos + 4 <= result.size());
put_belated_to_res; put_belated_to_res; put_belated_to_res; put_belated_to_res;
}
}
void belated_quadword(std::vector<uint8_t>& result, size_t pos, uint64_t value) {
void belated_quadword(std::vector<uint8_t>& result, size_t pos, uint64_t value) {
assert(pos + 8 <= result.size());
for (int i = 0; i < 8; i++) {
put_belated_to_res;
}
}
}
#undef put_belated_to_res
void write_sslot_id(std::vector<uint8_t>& result, regex_sslot_id_t x) {
void write_sslot_id(std::vector<uint8_t>& result, sslot_id_t x) {
write_doubleword(result, x);
}
}
void write_tai(std::vector<uint8_t>& result, regex_tai_t x) {
void write_tai(std::vector<uint8_t>& result, tai_t x) {
write_word(result, x);
}
}
void write_near_ptr(std::vector<uint8_t>& result, regex_near_ptr_t x) {
void write_near_ptr(std::vector<uint8_t>& result, near_ptr_t x) {
write_quadword(result, x);
}
}
void belated_sslot_id(std::vector<uint8_t>& result, size_t pos, regex_sslot_id_t value) {
void belated_sslot_id(std::vector<uint8_t>& result, size_t pos, sslot_id_t value) {
belated_doubleword(result, pos, value);
}
}
void belated_tai(std::vector<uint8_t>& result, size_t pos, regex_tai_t value) {
void belated_tai(std::vector<uint8_t>& result, size_t pos, tai_t value) {
belated_word(result, pos, value);
}
}
void belated_near_ptr(std::vector<uint8_t>& result, size_t pos, regex_near_ptr_t value) {
void belated_near_ptr(std::vector<uint8_t>& result, size_t pos, near_ptr_t value) {
belated_quadword(result, pos, value);
}
}
bookmark_id_t explicit_bookmarks::new_bookmark() {
bookmark_id_t explicit_bookmarks::new_bookmark() {
pile.emplace_back();
return free_bid++;
}
}
void explicit_bookmarks::write_unresolved_reference(std::vector<uint8_t> &result, bookmark_id_t bm) {
void explicit_bookmarks::write_unresolved_reference(std::vector<uint8_t> &result, bookmark_id_t bm) {
size_t where_to_fill_later = result.size();
write_near_ptr(result, 0);
pile[bm].positions_of_belated_refs.push_back(where_to_fill_later);
}
}
void explicit_bookmarks::land_bookmark(std::vector<uint8_t> &result, bookmark_id_t bm) {
void explicit_bookmarks::land_bookmark(std::vector<uint8_t> &result, bookmark_id_t bm) {
assert(!pile[bm].placed_somewhere);
pile[bm].placed_somewhere = true;
pile[bm].actual_position = result.size();
}
}
void explicit_bookmarks::finish(std::vector<uint8_t> &result) {
void explicit_bookmarks::finish(std::vector<uint8_t> &result) {
for (explicit_bookmark_info& bmi: pile) {
assert(bmi.positions_of_belated_refs.empty() || bmi.placed_somewhere);
if (bmi.placed_somewhere) {
@ -97,19 +97,20 @@ void explicit_bookmarks::finish(std::vector<uint8_t> &result) {
}
}
}
}
}
bookmark_id_t explicit_bookmarks::new_range_of_bookmarks(size_t n) {
bookmark_id_t explicit_bookmarks::new_range_of_bookmarks(size_t n) {
bookmark_id_t offset = free_bid;
free_bid += n;
for (size_t i = 0; i < n; i++) {
pile.emplace_back();
}
return offset;
}
}
bool explicit_bookmarks::has_landed(bookmark_id_t bm) {
bool explicit_bookmarks::has_landed(bookmark_id_t bm) {
return pile[bm].placed_somewhere;
}
}
#undef put_belated_to_res

View File

@ -4,40 +4,40 @@
#include <stdint.h>
#include <libregexis024vm/vm_opcodes_types.h>
#include <vector>
namespace regexis024 {
void write_byte(std::vector<uint8_t>& result, uint8_t x);
void write_word(std::vector<uint8_t>& result, uint16_t x);
void write_doubleword(std::vector<uint8_t>& result, uint32_t x);
void write_quadword(std::vector<uint8_t>& result, uint64_t x);
void write_byte(std::vector<uint8_t>& result, uint8_t x);
void write_word(std::vector<uint8_t>& result, uint16_t x);
void write_doubleword(std::vector<uint8_t>& result, uint32_t x);
void write_quadword(std::vector<uint8_t>& result, uint64_t x);
void belated_byte(std::vector<uint8_t>& result, size_t pos, uint8_t value);
void belated_word(std::vector<uint8_t>& result, size_t pos, uint16_t value);
void belated_doubleword(std::vector<uint8_t>& result, size_t pos, uint32_t value);
void belated_quadword(std::vector<uint8_t>& result, size_t pos, uint64_t value);
void belated_byte(std::vector<uint8_t>& result, size_t pos, uint8_t value);
void belated_word(std::vector<uint8_t>& result, size_t pos, uint16_t value);
void belated_doubleword(std::vector<uint8_t>& result, size_t pos, uint32_t value);
void belated_quadword(std::vector<uint8_t>& result, size_t pos, uint64_t value);
void write_sslot_id(std::vector<uint8_t>& result, regex_sslot_id_t x);
void write_tai(std::vector<uint8_t>& result, regex_tai_t x);
void write_near_ptr(std::vector<uint8_t>& result, regex_near_ptr_t x);
void write_sslot_id(std::vector<uint8_t>& result, sslot_id_t x);
void write_tai(std::vector<uint8_t>& result, tai_t x);
void write_near_ptr(std::vector<uint8_t>& result, near_ptr_t x);
void belated_sslot_id(std::vector<uint8_t>& result, size_t pos, regex_sslot_id_t value);
void belated_tai(std::vector<uint8_t>& result, size_t pos, regex_tai_t value);
void belated_near_ptr(std::vector<uint8_t>& result, size_t pos, regex_near_ptr_t value);
void belated_sslot_id(std::vector<uint8_t>& result, size_t pos, sslot_id_t value);
void belated_tai(std::vector<uint8_t>& result, size_t pos, tai_t value);
void belated_near_ptr(std::vector<uint8_t>& result, size_t pos, near_ptr_t value);
// constexpr uint64_t INSTRUCTION_SZ = REGEX024_BYTECODE_INSTRUCTION_SZ;
// constexpr uint64_t SSLOT_ID_SZ = REGEX024_BYTECODE_SSLOT_ID_SZ;
// constexpr uint64_t TRACK_ARRAY_INDEX_ID_SZ = REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ;
// constexpr uint64_t NEAR_POINTER_SZ = REGEX024_BYTECODE_NEAR_POINTER_SZ;
// constexpr uint64_t INSTRUCTION_SZ = REGEX024_BYTECODE_INSTRUCTION_SZ;
// constexpr uint64_t SSLOT_ID_SZ = REGEX024_BYTECODE_SSLOT_ID_SZ;
// constexpr uint64_t TRACK_ARRAY_INDEX_ID_SZ = REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ;
// constexpr uint64_t NEAR_POINTER_SZ = REGEX024_BYTECODE_NEAR_POINTER_SZ;
typedef size_t bookmark_id_t;
typedef size_t bookmark_id_t;
struct explicit_bookmark_info {
struct explicit_bookmark_info {
std::vector<size_t> positions_of_belated_refs;
bool placed_somewhere = false;
size_t actual_position;
};
};
struct explicit_bookmarks {
struct explicit_bookmarks {
bookmark_id_t free_bid = 0;
/* For each named explicit bookmark there is an element in PILE */
std::vector<explicit_bookmark_info> pile;
@ -57,7 +57,7 @@ struct explicit_bookmarks {
bookmark_id_t new_range_of_bookmarks(size_t n);
bool has_landed(bookmark_id_t bm);
};
};
}
#endif

View File

@ -2,22 +2,23 @@
#include <libregexis024vm/vm_opcodes.h>
#include <assert.h>
void cmd_JUMP(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, bookmark_id_t dest) {
write_byte(result, regex024_opcodes::JUMP);
namespace regexis024 {
void cmd_JUMP(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, bookmark_id_t dest) {
write_byte(result, opcodes::JUMP);
bookmark_manager.write_unresolved_reference(result, dest);
}
}
constexpr regex024_opcode cmp_EQUAL[4] = {regex024_opcodes::JCEQUAL_B, regex024_opcodes::JCEQUAL_W,
regex024_opcodes::JCEQUAL_DW, regex024_opcodes::JCEQUAL_QW};
constexpr regex024_opcode cmp_LESS[4] = {regex024_opcodes::JCLESS_B, regex024_opcodes::JCLESS_W,
regex024_opcodes::JCLESS_DW, regex024_opcodes::JCLESS_QW};
constexpr regex024_opcode cmp_GRTR[4] = {regex024_opcodes::JCGRTR_B, regex024_opcodes::JCGRTR_W,
regex024_opcodes::JCGRTR_DW, regex024_opcodes::JCGRTR_QW};
constexpr opcode_t cmp_EQUAL[4] = {opcodes::JCEQUAL_B, opcodes::JCEQUAL_W,
opcodes::JCEQUAL_DW, opcodes::JCEQUAL_QW};
constexpr opcode_t cmp_LESS[4] = {opcodes::JCLESS_B, opcodes::JCLESS_W,
opcodes::JCLESS_DW, opcodes::JCLESS_QW};
constexpr opcode_t cmp_GRTR[4] = {opcodes::JCGRTR_B, opcodes::JCGRTR_W,
opcodes::JCGRTR_DW, opcodes::JCGRTR_QW};
void cmd_JC(const regex024_opcode cmpT[4],
void cmd_JC(const opcode_t cmpT[4],
std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest)
{
{
if (val <= UINT8_MAX) {
write_byte(result, cmpT[0]);
write_byte(result, static_cast<uint8_t>(val));
@ -32,44 +33,45 @@ void cmd_JC(const regex024_opcode cmpT[4],
write_quadword(result, val);
}
bookmark_manager.write_unresolved_reference(result, dest);
}
}
void cmd_JCEQUAL(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
void cmd_JCEQUAL(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
cmd_JC(cmp_EQUAL, result, bookmark_manager, val, dest);
}
}
void cmd_JCLESS(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
void cmd_JCLESS(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
cmd_JC(cmp_LESS, result, bookmark_manager, val, dest);
}
}
void cmd_JCGRTR(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
void cmd_JCGRTR(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest) {
cmd_JC(cmp_GRTR, result, bookmark_manager, val, dest);
}
}
void cmd_DIE(std::vector<uint8_t> &result) {
write_byte(result, regex024_opcodes::DIE);
}
void cmd_DIE(std::vector<uint8_t> &result) {
write_byte(result, opcodes::DIE);
}
void cmd_MATCH(std::vector<uint8_t> &result) {
write_byte(result, regex024_opcodes::MATCH);
}
void cmd_MATCH(std::vector<uint8_t> &result) {
write_byte(result, opcodes::MATCH);
}
void cmd_READ_first_ns(std::vector<uint8_t>& result, size_t slot) {
void cmd_READ_first_ns(std::vector<uint8_t>& result, size_t slot) {
assert(slot <= UINT32_MAX);
write_byte(result, regex024_opcodes::READ);
write_byte(result, opcodes::READ);
write_sslot_id(result, slot);
}
}
void cmd_FORK(std::vector<uint8_t> &result, explicit_bookmarks& bookmark_manager, size_t slot, bookmark_id_t dest) {
void cmd_FORK(std::vector<uint8_t> &result, explicit_bookmarks& bookmark_manager, size_t slot, bookmark_id_t dest) {
assert(slot <= UINT32_MAX);
write_byte(result, regex024_opcodes::FORK);
write_byte(result, opcodes::FORK);
write_sslot_id(result, slot);
bookmark_manager.write_unresolved_reference(result, dest);
}
}
void cmd_READ_second_ns(std::vector<uint8_t>& result, std::vector<size_t>& belate_second_read_ns_slot_args) {
write_byte(result, regex024_opcodes::READ);
void cmd_READ_second_ns(std::vector<uint8_t>& result, std::vector<size_t>& belate_second_read_ns_slot_args) {
write_byte(result, opcodes::READ);
belate_second_read_ns_slot_args.push_back(result.size());
write_sslot_id(result, 0);
}
}

View File

@ -4,17 +4,19 @@
#include <libregexis024fa/graph_to_bytecode/natural_compiler_utils.h>
#include <libregexis024vm/vm_opcodes.h>
void cmd_JUMP(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, bookmark_id_t dest);
namespace regexis024 {
void cmd_JUMP(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, bookmark_id_t dest);
void cmd_JCEQUAL(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_JCLESS(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_JCGRTR(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_JCEQUAL(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_JCLESS(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_JCGRTR(std::vector<uint8_t>& result, explicit_bookmarks& bookmark_manager, uint64_t val, bookmark_id_t dest);
void cmd_DIE(std::vector<uint8_t>& result);
void cmd_MATCH(std::vector<uint8_t>& result);
void cmd_DIE(std::vector<uint8_t>& result);
void cmd_MATCH(std::vector<uint8_t>& result);
void cmd_READ_first_ns(std::vector<uint8_t>& result, size_t slot);
void cmd_READ_second_ns(std::vector<uint8_t>& result, std::vector<size_t>& belate_second_read_ns_slot_args);
void cmd_FORK(std::vector<uint8_t> &result, explicit_bookmarks& bookmark_manager, size_t slot, bookmark_id_t dest);
void cmd_READ_first_ns(std::vector<uint8_t>& result, size_t slot);
void cmd_READ_second_ns(std::vector<uint8_t>& result, std::vector<size_t>& belate_second_read_ns_slot_args);
void cmd_FORK(std::vector<uint8_t> &result, explicit_bookmarks& bookmark_manager, size_t slot, bookmark_id_t dest);
}
#endif

View File

@ -3,7 +3,8 @@
#include <assert.h>
#include <libregexis024vm/utils.h>
void reattach_fa_node_edge(FA_Node **old_node_ptr, FA_Node *new_node) {
namespace regexis024 {
void reattach_fa_node_edge(FA_Node **old_node_ptr, FA_Node *new_node) {
assert(old_node_ptr);
if (*old_node_ptr){
assert((**old_node_ptr).refs);
@ -12,28 +13,27 @@ void reattach_fa_node_edge(FA_Node **old_node_ptr, FA_Node *new_node) {
if (new_node)
new_node->refs++;
*old_node_ptr = new_node;
}
}
/* We basically reattch fa.start to node */
void yay_new_start(FA_Container &fa, FA_NodePathPart *node) {
/* We basically reattch fa.start to node */
void yay_new_start(FA_Container &fa, FA_NodePathPart *node) {
assert(node);
node->refs++;
node->nxt_node = fa.start;
fa.start = node;
}
}
void add_option_to_fork_node(FA_NodeOfForking *fnode, FA_Node *transition_dest) {
void add_option_to_fork_node(FA_NodeOfForking *fnode, FA_Node *transition_dest) {
fnode->nxt_options.push_back(transition_dest);
if(transition_dest)
transition_dest->refs++;
}
}
void reattach_nxt_node(FA_NodePathPart *node, FA_Node *dest) {
void reattach_nxt_node(FA_NodePathPart *node, FA_Node *dest) {
reattach_fa_node_edge(&(node->nxt_node), dest);
}
}
// todo: get rid of exitf in the whole project
FA_Node* copy_node_no_container_adjustments(FA_Node& node){
FA_Node* copy_node_no_container_adjustments(FA_Node& node){
FA_Node* res;
/* Using implicitly defined copy constructors */
#define typeCase(etype, ctype) case etype: res = new ctype((ctype&)node); break;
@ -53,19 +53,20 @@ FA_Node* copy_node_no_container_adjustments(FA_Node& node){
res->refs = 0;
res->search_mark = -1;
return res;
}
}
/* In case when transferring the ownership of this new raw pointer has failed, node is destroyed, exception is thrown */
FA_Node *copy_fa_node(FA_Node& node, FA_Container &fa) {
/* In case when transferring the ownership of this new raw pointer has failed, node is destroyed, exception is thrown */
FA_Node *copy_fa_node(FA_Node& node, FA_Container &fa) {
FA_Node* res = copy_node_no_container_adjustments(node);
/* Can invalidate ponter res (in which case it also throws exeption, so none of this matters in the end) */
fa.registerNew(res);
res->reAdd_references();
return res;
}
}
FA_Node *copy_fa_node_to_another_fa(FA_Node& node, FA_Container &resultFa) {
FA_Node *copy_fa_node_to_another_fa(FA_Node& node, FA_Container &resultFa) {
FA_Node* res = copy_node_no_container_adjustments(node);
resultFa.registerNew(res);
return res;
}
}

View File

@ -4,14 +4,16 @@
#include "finite_automaton.h"
#include "fa_first_stage_fix.h"
FA_Node* copy_fa_node(FA_Node& node, FA_Container& fa);
void yay_new_start(FA_Container& fa, FA_NodePathPart* node);
void reattach_fa_node_edge(FA_Node** old_node_ptr, FA_Node* new_node);
void add_option_to_fork_node(FA_NodeOfForking* fnode, FA_Node* transition_dest);
void reattach_nxt_node(FA_NodePathPart* node, FA_Node* dest);
namespace regexis024 {
FA_Node* copy_fa_node(FA_Node& node, FA_Container& fa);
void yay_new_start(FA_Container& fa, FA_NodePathPart* node);
void reattach_fa_node_edge(FA_Node** old_node_ptr, FA_Node* new_node);
void add_option_to_fork_node(FA_NodeOfForking* fnode, FA_Node* transition_dest);
void reattach_nxt_node(FA_NodePathPart* node, FA_Node* dest);
/* This is a one weird operation. New node in resultFa will still point to nodes in sourceFa,
/* This is a one weird operation. New node in resultFa will still point to nodes in sourceFa,
* without increasing refcount of those nodes. YOU HAVE TO FIX IT ASAP */
FA_Node* copy_fa_node_to_another_fa(FA_Node& node, FA_Container& resultFa);
FA_Node* copy_fa_node_to_another_fa(FA_Node& node, FA_Container& resultFa);
}
#endif //LIBREGEXIS024_MISC_FA_FUNCS_H

View File

@ -1,15 +1,16 @@
#include <libregexis024fa/selarr_priority_table.h>
#include <assert.h>
bool RegexPriorityTableAction_Pos::isForRange() const {
namespace regexis024 {
bool RegexPriorityTableAction_Pos::isForRange() const {
return second >= 0;
}
}
RegexPriorityTableAction_Pos::RegexPriorityTableAction_Pos(int first, int second, tracking_var_type type):
RegexPriorityTableAction_Pos::RegexPriorityTableAction_Pos(int first, int second, tracking_var_type_t type):
first(first),second(second), type(type) {}
//
//
RegexPriorityTableAction::RegexPriorityTableAction(bool minimize, int first, int second, tracking_var_type type):
RegexPriorityTableAction::RegexPriorityTableAction(bool minimize, int first, int second, tracking_var_type_t type):
minimize(minimize), pos(first, second, type) {}
//
//
}

View File

@ -5,22 +5,24 @@
#include <vector>
#include <libregexis024fa/tracking_variables.h>
struct RegexPriorityTableAction_Pos{
namespace regexis024 {
struct RegexPriorityTableAction_Pos{
/* first and second are indexes in selarr (but second can be -1 if it is unused) */
int first;
int second;
tracking_var_type type;
tracking_var_type_t type;
bool isForRange() const;
RegexPriorityTableAction_Pos(int first, int second, tracking_var_type type);
};
RegexPriorityTableAction_Pos(int first, int second, tracking_var_type_t type);
};
struct RegexPriorityTableAction{
struct RegexPriorityTableAction{
bool minimize;
RegexPriorityTableAction_Pos pos;
RegexPriorityTableAction(bool minimize, int first, int second, tracking_var_type type);
};
RegexPriorityTableAction(bool minimize, int first, int second, tracking_var_type_t type);
};
typedef std::vector<RegexPriorityTableAction> RegexPriorityTable;
typedef std::vector<RegexPriorityTableAction> RegexPriorityTable;
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024FA_SELARR_PRIORITY_TABLE_H

View File

@ -1,53 +1,48 @@
#include <libregexis024fa/tracking_fa_nodes.h>
#include <assert.h>
bool isImmMovOpcode(regex024_opcode inst) {
return inst == regex024_opcodes::MOV_COLARR_IMM || inst == regex024_opcodes::MOV_SELARR_IMM;
}
namespace regexis024 {
bool isImmMovOpcode(opcode_t inst) {
return inst == opcodes::MOV_COLARR_IMM || inst == opcodes::MOV_SELARR_IMM;
}
bool isCurPosMovOpcode(regex024_opcode inst) {
return inst == regex024_opcodes::MOV_COLARR_BTPOS || inst == regex024_opcodes::MOV_SELARR_CHPOS;
}
bool isCurPosMovOpcode(opcode_t inst) {
return inst == opcodes::MOV_COLARR_BTPOS || inst == opcodes::MOV_SELARR_CHPOS;
}
bool isColarrOpcode(regex024_opcode inst) {
return inst == regex024_opcodes::MOV_COLARR_IMM || inst == regex024_opcodes::MOV_COLARR_BTPOS;
}
bool isColarrOpcode(opcode_t inst) {
return inst == opcodes::MOV_COLARR_IMM || inst == opcodes::MOV_COLARR_BTPOS;
}
bool isSelarrOpcode(regex024_opcode inst) {
return inst == regex024_opcodes::MOV_SELARR_IMM || inst == regex024_opcodes::MOV_SELARR_CHPOS;
}
bool isSelarrOpcode(opcode_t inst) {
return inst == opcodes::MOV_SELARR_IMM || inst == opcodes::MOV_SELARR_CHPOS;
}
bool isTrackingFaNode(const FA_Node *n) {
bool isTrackingFaNode(const FA_Node *n) {
return n->type == track_array_mov_imm || n->type == track_array_mov_halfinvariant;
}
}
TrackingOperationInFa::TrackingOperationInFa(regex024_opcode opcode, regex_tai_t key, uint64_t imm_value)
: opcode(opcode), key(key), immValue(imm_value) {}
TrackingOperationInFa::TrackingOperationInFa(regex024_opcode opcode, regex_tai_t key)
: opcode(opcode), key(key) {}
std::string TrackingOperationInFa::toString() const {
std::string TrackingOperationInFa::toString() const {
switch (opcode){
case regex024_opcodes::MOV_COLARR_IMM:
case opcodes::MOV_COLARR_IMM:
return "colarr[" + std::to_string(key) + "] := " + std::to_string(immValue);
case regex024_opcodes::MOV_SELARR_IMM:
case opcodes::MOV_SELARR_IMM:
return "selarr[" + std::to_string(key) + "] := " + std::to_string(immValue);
case regex024_opcodes::MOV_COLARR_BTPOS:
case opcodes::MOV_COLARR_BTPOS:
return "colarr[" + std::to_string(key) + "] := cur byte position";
case regex024_opcodes::MOV_SELARR_CHPOS:
case opcodes::MOV_SELARR_CHPOS:
return "selarr[" + std::to_string(key) + "] := cur char position";
default:
return "wrong collection operation";
}
}
}
FA_NodePathPart* convert_to_node(const TrackingOperationInFa& op, FA_Container& fa) {
FA_NodePathPart* convert_to_node(const TrackingOperationInFa& op, FA_Container& fa) {
if (isImmMovOpcode(op.opcode)) {
return fa.makeTrackArrayMovImm(op.opcode, op.key, op.immValue);
}
assert(isCurPosMovOpcode(op.opcode));
return fa.makeTrackArrayMovHalfinvariant(op.opcode, op.key);
}
}

View File

@ -5,27 +5,24 @@
#include <libregexis024fa/finite_automaton.h>
#include <string>
bool isImmMovOpcode(regex024_opcode inst);
bool isCurPosMovOpcode(regex024_opcode inst);
bool isColarrOpcode(regex024_opcode inst);
bool isSelarrOpcode(regex024_opcode inst);
namespace regexis024 {
bool isImmMovOpcode(opcode_t inst);
bool isCurPosMovOpcode(opcode_t inst);
bool isColarrOpcode(opcode_t inst);
bool isSelarrOpcode(opcode_t inst);
bool isTrackingFaNode(const FA_Node* n);
bool isTrackingFaNode(const FA_Node* n);
struct TrackingOperationInFa {
regex024_opcode opcode;
regex_tai_t key;
struct TrackingOperationInFa {
opcode_t opcode;
tai_t key;
/* Not needed for halfinvariant operations */
uint64_t immValue;
TrackingOperationInFa(regex024_opcode opcode, regex_tai_t key, uint64_t imm_value);
TrackingOperationInFa(regex024_opcode opcode, regex_tai_t key);
std::string toString() const;
};
FA_NodePathPart* convert_to_node(const TrackingOperationInFa& op, FA_Container& fa);
};
FA_NodePathPart* convert_to_node(const TrackingOperationInFa& op, FA_Container& fa);
}
#endif

View File

@ -1,14 +1,16 @@
#ifndef LIBREGEXIS024_SRC_LIBREGEXIS024FA_TRACKING_VARIABLES_H
#define LIBREGEXIS024_SRC_LIBREGEXIS024FA_TRACKING_VARIABLES_H
namespace tracking_var_types {
namespace regexis024 {
namespace tracking_var_types {
enum tracking_var_type_I {
range,
dot_cur_pos,
dot_immediate
};
}
typedef tracking_var_types::tracking_var_type_I tracking_var_type_t;
}
typedef tracking_var_types::tracking_var_type_I tracking_var_type;
#endif

View File

@ -2,7 +2,8 @@
#include <libregexis024sol/sol_misc_base.h>
#include <assert.h>
uint32_t read_hex(REGEX_IS024_MeaningContext& ctx, int sz){
namespace regexis024 {
uint32_t read_hex(REGEX_IS024_MeaningContext& ctx, int sz){
uint32_t res = 0;
for (int i = 0; i < sz; i++){
int32_t ch = peep(ctx);
@ -19,19 +20,19 @@ uint32_t read_hex(REGEX_IS024_MeaningContext& ctx, int sz){
readChar(ctx);
}
return res;
}
}
void unicode_in_bs_case(REGEX_IS024_MeaningContext &ctx, bool &ret_is_multicode, codeset_t &ret_set, int sz){
void unicode_in_bs_case(REGEX_IS024_MeaningContext &ctx, bool &ret_is_multicode, codeset_t &ret_set, int sz){
ret_is_multicode = false;
readChar(ctx);
uint32_t hc = read_hex(ctx, sz); // Might create an error
ret_set = codeset_of_one_char(hc);
}
}
void
backslash_expression_parsing_try_regular(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc,
void
backslash_expression_parsing_try_regular(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc,
bool &ret_is_multicode, codeset_t &ret_set)
{
{
int32_t leader = peep(ctx);
if (ctx.error)
return;
@ -59,4 +60,5 @@ backslash_expression_parsing_try_regular(REGEX_IS024_MeaningContext &ctx, const
report(ctx, "backslash in the wrong place");
}
}
}
}

View File

@ -5,32 +5,33 @@
#include <assert.h>
#include <memory>
struct ParseCall{
namespace regexis024 {
struct ParseCall{
virtual ~ParseCall() = default;
virtual std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext& ctx) { assert(false); }
virtual std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext& ctx) { assert(false); }
};
};
struct Top_ParseCall: public ParseCall{
struct Top_ParseCall: public ParseCall{
Command& res;
explicit Top_ParseCall(Command &res) : res(res) {}
std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext &ctx) override;
std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext &ctx) override;
};
};
struct Bracker_ParseCall: public ParseCall{
struct Bracker_ParseCall: public ParseCall{
std::vector<CommandArgument>& res;
bool closingBraceEnded = false;
explicit Bracker_ParseCall(std::vector<CommandArgument> &res) : res(res) {}
std::unique_ptr<ParseCall> argReadProc(REGEX_IS024_MeaningContext& ctx);
std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext &ctx) override;
std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext &ctx) override;
};
};
#define call_ERROR_CHECK do { if (ctx.error) { return NULL; } } while (0)
#define call_THROW(str) do { report(ctx, "command expression: " str); return NULL; } while (0)
std::unique_ptr<ParseCall> Top_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
std::unique_ptr<ParseCall> Top_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
assert(readChar(ctx) == U'!');
int32_t ch = peep(ctx); call_ERROR_CHECK;
if (ch == U'~'){
@ -50,23 +51,23 @@ std::unique_ptr<ParseCall> Top_ParseCall::firstTime(REGEX_IS024_MeaningContext &
return std::make_unique<Bracker_ParseCall>(res.arguments);
}
call_THROW("top lvl: command call should be ended with ';' or '{...}'");
}
}
std::unique_ptr<ParseCall> Top_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
std::unique_ptr<ParseCall> Top_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
return NULL;
}
}
std::unique_ptr<ParseCall> Bracker_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
std::unique_ptr<ParseCall> Bracker_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
assert(readChar(ctx) == U'{');
return argReadProc(ctx);
}
}
std::unique_ptr<ParseCall> Bracker_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
std::unique_ptr<ParseCall> Bracker_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
closingBraceEnded = true;
return argReadProc(ctx);
}
}
std::unique_ptr<ParseCall> Bracker_ParseCall::argReadProc(REGEX_IS024_MeaningContext &ctx) {
std::unique_ptr<ParseCall> Bracker_ParseCall::argReadProc(REGEX_IS024_MeaningContext &ctx) {
repeat:
int32_t ch = peep(ctx); call_ERROR_CHECK;
if (ch == U';'){
@ -98,9 +99,9 @@ std::unique_ptr<ParseCall> Bracker_ParseCall::argReadProc(REGEX_IS024_MeaningCon
call_THROW("brace lvl: argument ends with ';' or {...}");
}
call_THROW("brace lvl: argument starts with ';' or it's name");
}
}
Command command_expr_parse(REGEX_IS024_MeaningContext &ctx) {
Command command_expr_parse(REGEX_IS024_MeaningContext &ctx) {
std::vector<std::unique_ptr<ParseCall>> callStack;
Command res;
callStack.push_back(std::make_unique<Top_ParseCall>(res));
@ -118,16 +119,16 @@ Command command_expr_parse(REGEX_IS024_MeaningContext &ctx) {
}
}
return res;
}
}
const char* commands_for_codesets[] = {"word", "space", "digit", "variable", "any", "A", NULL};
const char* commands_for_codesets[] = {"word", "space", "digit", "variable", "any", "A", NULL};
bool is_command_for_charset(const Command &cmd) {
bool is_command_for_charset(const Command &cmd) {
return !cmd.tilda && cmd.arguments.empty() && is_string_in_stringset(cmd.name.c_str(), commands_for_codesets);
}
}
void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command &cmd, codeset_t& ret)
{
void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command &cmd, codeset_t& ret)
{
if (cmd.name == "word")
ret = cc.word_constituents;
else if (cmd.name == "space")
@ -140,4 +141,5 @@ void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command
ret = codeset_of_all;
else
assert(false);
}
}

View File

@ -1,6 +1,7 @@
#include <libregexis024sol/common_codesets.h>
CommonCodesets::CommonCodesets() {
namespace regexis024 {
CommonCodesets::CommonCodesets() {
spaces = set_add_char(spaces, U'\n');
spaces = set_add_char(spaces, U' ');
spaces = set_add_char(spaces, U'\t');
@ -10,4 +11,5 @@ CommonCodesets::CommonCodesets() {
digits = codeset_t({{'0', '9'}});
variable_constituents = set_add_char(word_constituents, U'-');
variable_constituents = merge_sets(variable_constituents, digits);
}
}

View File

@ -3,12 +3,14 @@
#include <libregexis024fa/codeset.h>
struct CommonCodesets {
namespace regexis024 {
struct CommonCodesets {
codeset_t spaces;
codeset_t word_constituents;
codeset_t digits;
codeset_t variable_constituents;
CommonCodesets();
};
};
}
#endif

View File

@ -23,12 +23,13 @@
#define aux_ERROR_CHECK do { if (ctx.error) { return; } } while (0)
#define aux_THROW(str) do { report(ctx, "regex: " str); return; } while (0)
/* ****************************** Top */
namespace regexis024 {
/* ****************************** Top */
const char* dfa_arg_aliases_condone[] = {"forgive", "condone", "okay", "optional", "nonimportant", "ifpossible", NULL};
const char* dfa_arg_aliases_acerbic[] = {"acerbic", "angry", "pedantic", "nofork", "pure", "important", "fierce", NULL};
const char* dfa_arg_aliases_condone[] = {"forgive", "condone", "okay", "optional", "nonimportant", "ifpossible", NULL};
const char* dfa_arg_aliases_acerbic[] = {"acerbic", "angry", "pedantic", "nofork", "pure", "important", "fierce", NULL};
void dfa_command_processing(REGEX_IS024_MeaningContext &ctx, ParsingContext& pctx, const Command& cmdBuf){
void dfa_command_processing(REGEX_IS024_MeaningContext &ctx, ParsingContext& pctx, const Command& cmdBuf){
if (pctx.dfa_cmd_activated){
report(ctx, "repeating !dfa command");
return;
@ -48,9 +49,9 @@ void dfa_command_processing(REGEX_IS024_MeaningContext &ctx, ParsingContext& pct
}
}
report(ctx, "wrong arguments in !dfa command");
}
}
void select_command_processing(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, const Command& cmdBuf){
void select_command_processing(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, const Command& cmdBuf){
if (pctx.select_cmd_encountered)
aux_THROW("repeating !select command");
pctx.select_cmd_encountered = true;
@ -89,9 +90,9 @@ void select_command_processing(REGEX_IS024_MeaningContext& ctx, ParsingContext&
pctx.is_inside_of_these_sa_subexpressions.assign(ctx.ktr.retrieval_info.size(), false);
/* Other info will be filled once a tracking-unit with such name will be actually found in regex */
}
}
}
void jump_into_madness(ctx_t& ctx, ParsingContext& pctx, FA_Container &fa, int hn){
void jump_into_madness(ctx_t& ctx, ParsingContext& pctx, FA_Container &fa, int hn){
while (true){
int32_t pch = peep(ctx); aux_ERROR_CHECK;
if (pch != U'!'){
@ -115,26 +116,26 @@ void jump_into_madness(ctx_t& ctx, ParsingContext& pctx, FA_Container &fa, int h
break;
}
}
}
}
chekushka TopLvl_ParseCall::firstTime(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) {
chekushka TopLvl_ParseCall::firstTime(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) {
result.assertDefault();
jump_into_madness(ctx, pctx, fa, 1);
if (ctx.have_comment_tail)
return NULL;
return std::make_unique<ForkLvl_ParseCall>(result);
}
}
chekushka TopLvl_ParseCall::afterReceive(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) {
chekushka TopLvl_ParseCall::afterReceive(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) {
jump_into_madness(ctx, pctx, fa, 2);
if (!isEnd(ctx))
call_THROW("top lvl: EOF expected");
return NULL;
}
}
/* ********************************* Bracket */
/* ********************************* Bracket */
chekushka BracketLvl_ParseCall::firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
chekushka BracketLvl_ParseCall::firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
result.assertDefault();
assert(readChar(ctx) == U'(');
/* sequence lvl already took care about resolving name and configuring SubtrackingNameInfo */
@ -148,9 +149,9 @@ chekushka BracketLvl_ParseCall::firstTime(REGEX_IS024_MeaningContext& ctx, Parsi
}
}
return std::make_unique<ForkLvl_ParseCall>(tmp_ret_buff);
}
}
chekushka BracketLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
chekushka BracketLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
if (peep(ctx) != U')')
call_THROW("missing ')'");
readChar(ctx);
@ -161,32 +162,32 @@ chekushka BracketLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, Pa
assert(tai_slots.colarr_first >= 0 && tai_slots.colarr_first < UINT16_MAX);
assert(tai_slots.colarr_second >= 0 && tai_slots.colarr_second < UINT16_MAX);
result = join(subexpression_from_path(fa.makeTrackArrayMovHalfinvariant(
regex024_opcodes::MOV_COLARR_BTPOS, tai_slots.colarr_first)), result);
opcodes::MOV_COLARR_BTPOS, tai_slots.colarr_first)), result);
result = join(result, subexpression_from_path(fa.makeTrackArrayMovHalfinvariant(
regex024_opcodes::MOV_COLARR_BTPOS, tai_slots.colarr_second)));
opcodes::MOV_COLARR_BTPOS, tai_slots.colarr_second)));
}
if (tai_slots.stored_in_sa){
assert(tai_slots.selarr_first >= 0 && tai_slots.selarr_first < UINT16_MAX);
assert(tai_slots.selarr_second >= 0 && tai_slots.selarr_second < UINT16_MAX);
result = join(subexpression_from_path(fa.makeTrackArrayMovHalfinvariant(
regex024_opcodes::MOV_SELARR_CHPOS, tai_slots.selarr_first)), result);
opcodes::MOV_SELARR_CHPOS, tai_slots.selarr_first)), result);
result = join(result, subexpression_from_path(fa.makeTrackArrayMovHalfinvariant(
regex024_opcodes::MOV_SELARR_CHPOS, tai_slots.selarr_second)));
opcodes::MOV_SELARR_CHPOS, tai_slots.selarr_second)));
pctx.is_inside_of_these_sa_subexpressions[namedSubexpressionId] = false;
}
}
return NULL;
}
}
/* ******************************* Fork */
/* ******************************* Fork */
chekushka ForkLvl_ParseCall::firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
chekushka ForkLvl_ParseCall::firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
result.assertDefault();
options.emplace_back(); // Default one contains nothing. It will be overwritten
return std::make_unique<Sequence_ParseCall>(options.back());
}
}
chekushka ForkLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
chekushka ForkLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) {
int32_t end_reason = peep(ctx); call_ERROR_CHECK;
if (end_reason == U'|'){
readChar(ctx);
@ -194,9 +195,9 @@ chekushka ForkLvl_ParseCall::afterReceive(REGEX_IS024_MeaningContext& ctx, Parsi
}
result = forkify(options, fa);
return NULL;
}
}
void parseBody(REGEX_IS024_MeaningContext& ctx, FA_Container& fa, SubExprCompiled& result, ParsingContext& pctx){
void parseBody(REGEX_IS024_MeaningContext& ctx, FA_Container& fa, SubExprCompiled& result, ParsingContext& pctx){
std::vector<std::shared_ptr<ParseCall>> callStack;
callStack.push_back(std::make_unique<TopLvl_ParseCall>(result));
bool first_time = true;
@ -222,10 +223,10 @@ void parseBody(REGEX_IS024_MeaningContext& ctx, FA_Container& fa, SubExprCompile
pctx.priority_table.emplace_back(sni.minimizing, sni.selarr_first, sni.selarr_second, sni.type);
}
}
}
}
REGEX_IS024_MeaningContext::REGEX_IS024_MeaningContext(size_t inputSize, const char *input) : input_size(inputSize),
input(reinterpret_cast<const uint8_t *>(input)) {
REGEX_IS024_MeaningContext::REGEX_IS024_MeaningContext(size_t inputSize, const char *input) : input_size(inputSize),
input(input) {
CommonCodesets codeset_collection;
FA_Container fa;
FA_Container fa_1f;
@ -277,4 +278,5 @@ REGEX_IS024_MeaningContext::REGEX_IS024_MeaningContext(size_t inputSize, const c
report(*this, "Failed to compile graph representation to bytecode representation");
return;
}
}
}

View File

@ -5,14 +5,11 @@
#include <vector>
#include <stdint.h>
// todo: SUPER HIGHT PRIORITY: MOVE all this spaces digits variable_constituents junk out of this class
// todo: also PLEEEASE, write static before literally nearly every single one little stupid function in this library
#include <libregexis024sol/part_of_expr_that_tracks.h>
struct REGEX_IS024_MeaningContext{
namespace regexis024 {
struct REGEX_IS024_MeaningContext{
size_t input_size;
const uint8_t* input;
const char* input;
bool error = false;
std::string error_msg;
@ -29,6 +26,6 @@ struct REGEX_IS024_MeaningContext{
uint16_t free_colarr_tai = 0;
REGEX_IS024_MeaningContext(size_t inputSize, const char *input);
};
};
}
#endif //LIBREGEXIS024_EXPR_COMPILER_H

View File

@ -6,23 +6,24 @@
#define aux_ERROR_CHECK do { if (ctx.error) { return; } } while (0)
#define aux_THROW(str) do { report(ctx, "regex: " str); return; } while (0)
const char* header_command_dfa_names[] = {"dfa", "determinize", NULL};
namespace regexis024 {
const char* header_command_dfa_names[] = {"dfa", "determinize", NULL};
const char* header_command_select_names[] = {"s", "select", "selarr", "selectional", NULL};
const char* header_command_select_names[] = {"s", "select", "selarr", "selectional", NULL};
bool is_header_cmd(const Command &cmd) {
bool is_header_cmd(const Command &cmd) {
return cmd.tilda || is_header_dfa_cmd(cmd), is_header_dfa_cmd(cmd);
}
}
bool is_header_dfa_cmd(const Command &cmd) {
bool is_header_dfa_cmd(const Command &cmd) {
return is_string_in_stringset(cmd.name.c_str(), header_command_dfa_names);
}
}
bool is_header_select_cmd(const Command &cmd) {
bool is_header_select_cmd(const Command &cmd) {
return is_string_in_stringset(cmd.name.c_str(), header_command_select_names);
}
}
void int_parse_with_limit_concern(const std::string &str, REGEX_IS024_MeaningContext &ctx, size_t &res, int lim) {
void int_parse_with_limit_concern(const std::string &str, REGEX_IS024_MeaningContext &ctx, size_t &res, int lim) {
res = 0;
for (char ch: str){
if (!('0' <= ch && ch <= '9'))
@ -31,4 +32,5 @@ void int_parse_with_limit_concern(const std::string &str, REGEX_IS024_MeaningCon
if (res > (size_t)lim)
aux_THROW("integer is too big");
}
}
}

View File

@ -4,10 +4,11 @@
#include <libregexis024sol/special_terminals.h>
bool is_header_cmd(const Command& cmd);
bool is_header_dfa_cmd(const Command& cmd);
bool is_header_select_cmd(const Command& cmd);
void int_parse_with_limit_concern(const std::string& str, REGEX_IS024_MeaningContext &ctx, size_t& res, int lim);
namespace regexis024 {
bool is_header_cmd(const Command& cmd);
bool is_header_dfa_cmd(const Command& cmd);
bool is_header_select_cmd(const Command& cmd);
void int_parse_with_limit_concern(const std::string& str, REGEX_IS024_MeaningContext &ctx, size_t& res, int lim);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_EXPR_PARSE_FUNCTIONS_COMMAND_RECOGNITION_H

View File

@ -14,9 +14,10 @@
#define aux_ERROR_CHECK do { if (ctx.error) { return; } } while (0)
#define aux_THROW(str) do { report(ctx, "regex: " str); return; } while (0)
/* **************************** Sequence */
namespace regexis024 {
/* **************************** Sequence */
void in_case_of_backslash(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc, FA_Container &fa, SubExprCompiled& backPart) {
void in_case_of_backslash(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc, FA_Container &fa, SubExprCompiled& backPart) {
assert(readChar(ctx) == U'\\');
int32_t leader = peep(ctx); aux_ERROR_CHECK;
if (leader == U'b'){
@ -62,9 +63,9 @@ void in_case_of_backslash(REGEX_IS024_MeaningContext &ctx, const CommonCodesets&
return; // To avoid reading leader again (it gets read in the end)
}
readChar(ctx);
}
}
void repeat_stuff_with_check(REGEX_IS024_MeaningContext& ctx,
void repeat_stuff_with_check(REGEX_IS024_MeaningContext& ctx,
SubExprCompiled &patient, FA_Container& fa, size_t min_allowed, size_t max_allowed){
if (min_allowed > max_allowed)
aux_THROW("repeat operation: min > max");
@ -75,9 +76,9 @@ void repeat_stuff_with_check(REGEX_IS024_MeaningContext& ctx,
"выражение корректно и не вызвает бесконечного цикла, напишите об этом в жалобную книгу: "
"По ссылке: file:///dev/null Ваши предложения по улучшению libregexis024 обязательно будут рассмотрены.");
apply_repeat_to_subexpression(patient, fa, min_allowed, max_allowed);
}
}
void repeat_command_processing(REGEX_IS024_MeaningContext &ctx, FA_Container &fa, std::vector<SubExprCompiled>& parts,
void repeat_command_processing(REGEX_IS024_MeaningContext &ctx, FA_Container &fa, std::vector<SubExprCompiled>& parts,
const Command& cmd){
if (parts.empty())
aux_THROW("no subexpression before !repeat command");
@ -107,10 +108,10 @@ void repeat_command_processing(REGEX_IS024_MeaningContext &ctx, FA_Container &fa
aux_THROW("!repeat: min > max");
repeat_stuff_with_check(ctx, parts.back(), fa, min_allowed, max_allowed); aux_ERROR_CHECK;
}
}
}
chekushka Sequence_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx, ParsingContext &pctx, FA_Container &fa) {
chekushka Sequence_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx, ParsingContext &pctx, FA_Container &fa) {
while (true) {
int32_t fst = peep(ctx);
call_ERROR_CHECK;
@ -179,22 +180,22 @@ chekushka Sequence_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx, Parsing
readChar(ctx);
if (ctx.ktr.retrieval_info[id].stored_in_sa)
parts.emplace_back(subexpression_from_path(
fa.makeTrackArrayMovImm(regex024_opcodes::MOV_SELARR_IMM,
fa.makeTrackArrayMovImm(opcodes::MOV_SELARR_IMM,
ctx.ktr.retrieval_info[id].selarr_first, value)));
if (ctx.ktr.retrieval_info[id].stored_in_ca)
parts.emplace_back(subexpression_from_path(
fa.makeTrackArrayMovImm(regex024_opcodes::MOV_COLARR_IMM,
fa.makeTrackArrayMovImm(opcodes::MOV_COLARR_IMM,
ctx.ktr.retrieval_info[id].colarr_first, value)));
} else if (typeDet == U';'){
ensure_space_for_track_unit(ctx, name, tracking_var_types::dot_cur_pos); call_ERROR_CHECK;
readChar(ctx);
if (ctx.ktr.retrieval_info[id].stored_in_sa)
parts.emplace_back(subexpression_from_path(
fa.makeTrackArrayMovHalfinvariant(regex024_opcodes::MOV_SELARR_CHPOS,
fa.makeTrackArrayMovHalfinvariant(opcodes::MOV_SELARR_CHPOS,
ctx.ktr.retrieval_info[id].selarr_first)));
if (ctx.ktr.retrieval_info[id].stored_in_ca)
parts.emplace_back(subexpression_from_path(
fa.makeTrackArrayMovHalfinvariant(regex024_opcodes::MOV_COLARR_BTPOS,
fa.makeTrackArrayMovHalfinvariant(opcodes::MOV_COLARR_BTPOS,
ctx.ktr.retrieval_info[id].colarr_first)));
} else
call_THROW("Missing ; or ( in the beginning of tracking unit");
@ -214,9 +215,10 @@ chekushka Sequence_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx, Parsing
for (SubExprCompiled& part: parts)
result = join(result, part);
return NULL;
}
}
chekushka Sequence_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx, ParsingContext &pctx, FA_Container &fa) {
chekushka Sequence_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx, ParsingContext &pctx, FA_Container &fa) {
// This is possible only if I received a bracket expression
return firstTime(ctx, pctx, fa);
}
}

View File

@ -10,7 +10,8 @@
#include <assert.h>
#include <libregexis024fa/selarr_priority_table.h>
struct ParsingContext{
namespace regexis024 {
struct ParsingContext{
/* Those subexpressions, that are tracket by s`a are forbidden from nesting inside themselves */
std::vector<bool> is_inside_of_these_sa_subexpressions;
bool select_cmd_encountered = false;
@ -27,24 +28,24 @@ struct ParsingContext{
explicit ParsingContext(const CommonCodesets& cc_): cc(cc_){}
};
typedef REGEX_IS024_MeaningContext ctx_t;
struct ParseCall;
typedef std::unique_ptr<ParseCall> chekushka;
struct ParseCall{
typedef REGEX_IS024_MeaningContext ctx_t;
struct ParseCall;
typedef std::unique_ptr<ParseCall> chekushka;
struct ParseCall{
SubExprCompiled& result;
explicit ParseCall(SubExprCompiled &result) : result(result) {}
virtual ~ParseCall() = default;
virtual chekushka afterReceive(ctx_t& ctx, ParsingContext& pctx, FA_Container& fa) { assert(false); }
virtual chekushka firstTime(ctx_t& ctx, ParsingContext& pctx, FA_Container& fa) { assert(false); }
};
};
struct TopLvl_ParseCall: public ParseCall{
struct TopLvl_ParseCall: public ParseCall{
explicit TopLvl_ParseCall(SubExprCompiled &result) : ParseCall(result) {}
chekushka afterReceive(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) override;
chekushka firstTime(ctx_t &ctx, ParsingContext &pctx, FA_Container &fa) override;
};
};
struct BracketLvl_ParseCall: public ParseCall{
struct BracketLvl_ParseCall: public ParseCall{
/* -1 if this is a normal bracket expression. Otherwise, it is an index in ctx.retrieval_info vector */
int64_t namedSubexpressionId;
SubExprCompiled tmp_ret_buff;
@ -52,23 +53,21 @@ struct BracketLvl_ParseCall: public ParseCall{
ParseCall(result), namedSubexpressionId(namedSubexpressionId) {}
chekushka afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) override;
chekushka firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa) override;
};
};
struct ForkLvl_ParseCall: public ParseCall{
struct ForkLvl_ParseCall: public ParseCall{
std::vector<SubExprCompiled> options;
explicit ForkLvl_ParseCall(SubExprCompiled &result) : ParseCall(result) {}
chekushka afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa);
chekushka firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa);
};
};
struct Sequence_ParseCall: public ParseCall{
struct Sequence_ParseCall: public ParseCall{
std::vector<SubExprCompiled> parts;
explicit Sequence_ParseCall(SubExprCompiled &result) :ParseCall(result) {}
chekushka afterReceive(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa);
chekushka firstTime(REGEX_IS024_MeaningContext& ctx, ParsingContext& pctx, FA_Container& fa);
};
/* Some auxilary functions */
};
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_EXPR_PARSE_FUNCTIONS_EPF_H

View File

@ -4,20 +4,20 @@
#define aux_THROW(str) do { report(ctx, "regex: " str); return; } while (0)
#define aux_ERROR_CHECK do { if (ctx.error) { return; } } while (0)
void for_one_type(REGEX_IS024_MeaningContext &ctx, uint16_t& free_ARR_tai, int& ARR_first, int& ARR_second,
const std::string& ARR_NAME, tracking_var_type type){
namespace regexis024 {
void for_one_type(REGEX_IS024_MeaningContext &ctx, uint16_t& free_ARR_tai, int& ARR_first, int& ARR_second,
const std::string& ARR_NAME, tracking_var_type_t type){
#define check_is_available() if (free_ARR_tai == UINT16_MAX) { \
report(ctx, ("regex: " + ARR_NAME + ": key namespace overflow").c_str()); return;}
report(ctx, ("regex: " + ARR_NAME + ": key namespace overflow").c_str()); return;}
check_is_available()
ARR_first = free_ARR_tai++;
if (type == tracking_var_types::range){
check_is_available()
ARR_second = free_ARR_tai++;
}
}
}
void ensure_space_for_track_unit(REGEX_IS024_MeaningContext &ctx, const std::string& name, tracking_var_type type) {
void ensure_space_for_track_unit(REGEX_IS024_MeaningContext &ctx, const std::string& name, tracking_var_type_t type) {
size_t id = ctx.ktr.track_names[name];
/* Size of this verctor won't be changed. THis is a safe reference */
SubtrackingNameInfo& info = ctx.ktr.retrieval_info[id];
@ -35,4 +35,5 @@ void ensure_space_for_track_unit(REGEX_IS024_MeaningContext &ctx, const std::str
} else if (info.type != type){
aux_THROW("tracking tool unit type mismatch");
}
}
}

View File

@ -4,7 +4,8 @@
#include <libregexis024sol/expr_compiler.h>
void ensure_space_for_track_unit(REGEX_IS024_MeaningContext &ctx, const std::string& name, tracking_var_type type);
namespace regexis024 {
void ensure_space_for_track_unit(REGEX_IS024_MeaningContext &ctx, const std::string& name, tracking_var_type_t type);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_TRACKING_UNITS_H

View File

@ -1,2 +0,0 @@
// #include <libregexis024sol/part_of_expr_that_tracks.h>

View File

@ -6,12 +6,13 @@
#include <string>
#include <libregexis024fa/tracking_variables.h>
struct SubtrackingNameInfo{
namespace regexis024 {
struct SubtrackingNameInfo{
bool stored_in_ca = true;
bool stored_in_sa = false;
bool discovered = false;
tracking_var_type type;
tracking_var_type_t type;
/* These fields will be -1 if unused */
int colarr_first = -1;
int colarr_second = -1;
@ -20,12 +21,12 @@ struct SubtrackingNameInfo{
bool minimizing = false;
int selarr_first = -1;
int selarr_second = -1;
};
};
struct KnownTrackingTools {
struct KnownTrackingTools {
std::map<std::string, int64_t> track_names;
std::vector<SubtrackingNameInfo> retrieval_info;
};
};
}
#endif //PART_OF_EXPR_THAT_TRACKS_H

View File

@ -1,19 +1,20 @@
#include <libregexis024sol/sol_misc_base.h>
#include <libregexis024vm/utils.h>
void report(REGEX_IS024_MeaningContext &ctx, const char *error) {
namespace regexis024 {
void report(REGEX_IS024_MeaningContext &ctx, const char *error) {
if (!ctx.error){
ctx.error = true;
ctx.error_msg = error;
}
}
}
bool isEnd(REGEX_IS024_MeaningContext &ctx) {
bool isEnd(REGEX_IS024_MeaningContext &ctx) {
return ctx.pos == ctx.input_size;
}
}
int32_t peep(REGEX_IS024_MeaningContext &ctx) {
// printf("pos = %lu\n", ctx.pos);
int32_t peep(REGEX_IS024_MeaningContext &ctx) {
// printf("pos = %lu\n", ctx.pos);
if (isEnd(ctx))
return -1; // This is probably the only place where getting negative return does not generate error
int32_t cp; size_t sz;
@ -21,10 +22,10 @@ int32_t peep(REGEX_IS024_MeaningContext &ctx) {
if (cp < 0)
report(ctx, "encoding error");
return cp;
}
}
int32_t readChar(REGEX_IS024_MeaningContext &ctx) {
// printf("READ pos = %lu\n", ctx.pos);
int32_t readChar(REGEX_IS024_MeaningContext &ctx) {
// printf("READ pos = %lu\n", ctx.pos);
int32_t cp; size_t sz;
utf8_string_iterat(cp, sz, ctx.pos, ctx.input, ctx.input_size);
if (cp >= 0)
@ -32,13 +33,13 @@ int32_t readChar(REGEX_IS024_MeaningContext &ctx) {
else
report(ctx, "bruh what?? How this even happened");
return cp;
}
}
bool is_REGEX024_nameConstituent(int32_t ch) {
bool is_REGEX024_nameConstituent(int32_t ch) {
return ('0' <= ch && ch <= '9') || ('a' <= ch && ch <= 'z') || ('A' <= ch && ch <= 'Z');
}
}
std::string tryRead_REGEX024_name(REGEX_IS024_MeaningContext &ctx) {
std::string tryRead_REGEX024_name(REGEX_IS024_MeaningContext &ctx) {
std::string res;
while (true){
int32_t ch = peep(ctx);
@ -50,6 +51,5 @@ std::string tryRead_REGEX024_name(REGEX_IS024_MeaningContext &ctx) {
}
}
return res;
}
}

View File

@ -5,16 +5,17 @@
#include <libregexis024sol/expr_compiler.h>
#include <string>
void report(REGEX_IS024_MeaningContext& ctx, const char* error);
namespace regexis024 {
void report(REGEX_IS024_MeaningContext& ctx, const char* error);
bool isEnd(REGEX_IS024_MeaningContext& ctx);
int32_t peep(REGEX_IS024_MeaningContext& ctx);
int32_t readChar(REGEX_IS024_MeaningContext& ctx);
bool isEnd(REGEX_IS024_MeaningContext& ctx);
int32_t peep(REGEX_IS024_MeaningContext& ctx);
int32_t readChar(REGEX_IS024_MeaningContext& ctx);
bool is_REGEX024_nameConstituent(int32_t ch);
/* Name in my library consists of [0-9a-zA-Z]. If the first peeped letter is not name constituent,
bool is_REGEX024_nameConstituent(int32_t ch);
/* Name in my library consists of [0-9a-zA-Z]. If the first peeped letter is not name constituent,
* empty string is returned */
std::string tryRead_REGEX024_name(REGEX_IS024_MeaningContext& ctx);
std::string tryRead_REGEX024_name(REGEX_IS024_MeaningContext& ctx);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_SOL_MISC_BASE_H

View File

@ -5,32 +5,34 @@
#include <libregexis024sol/expr_compiler.h>
#include <libregexis024sol/common_codesets.h>
/* This option of backslash usage should be checked last.
namespace regexis024 {
/* This option of backslash usage should be checked last.
* Function can generate error. Always check the error first */
void
backslash_expression_parsing_try_regular(REGEX_IS024_MeaningContext& ctx, const CommonCodesets& cc,
void
backslash_expression_parsing_try_regular(REGEX_IS024_MeaningContext& ctx, const CommonCodesets& cc,
bool& ret_is_multicode, codeset_t& ret_set);
struct CommandEntity;
struct Command;
struct CommandArgument;
struct CommandEntity;
struct Command;
struct CommandArgument;
struct CommandEntity{
struct CommandEntity{
std::string name;
std::vector<CommandArgument> arguments;
};
};
struct CommandArgument: CommandEntity{
struct CommandArgument: CommandEntity{
bool is_empty = true;
};
};
struct Command: CommandEntity{
struct Command: CommandEntity{
bool tilda = false;
};
};
/* Zlaya sobaka. Kidaet oshibki v context */
Command command_expr_parse(REGEX_IS024_MeaningContext& ctx);
bool is_command_for_charset(const Command& cmd);
void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command& cmd, codeset_t& ret);
/* Zlaya sobaka. Kidaet oshibki v context */
Command command_expr_parse(REGEX_IS024_MeaningContext& ctx);
bool is_command_for_charset(const Command& cmd);
void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command& cmd, codeset_t& ret);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_SPECIAL_TERMINALS_H

View File

@ -6,82 +6,83 @@
#include <memory>
#include <assert.h>
/* Can allow backslash (later should check that backslash expression is not multicharar or empty */
bool soundsLikeCharOrRangeStart(int32_t peeped) {
namespace regexis024 {
/* Can allow backslash (later should check that backslash expression is not multicharar or empty */
bool soundsLikeCharOrRangeStart(int32_t peeped) {
return peeped >= 0 && (peeped != U'[' && peeped != U']' && peeped != U'!' && \
peeped != '^' && peeped != '&' && peeped != '-');
}
}
typedef REGEX_IS024_MeaningContext ctx_t;
typedef REGEX_IS024_MeaningContext ctx_t;
struct ParseCall;
typedef std::shared_ptr<ParseCall> chekushka;
struct ParseCall;
typedef std::shared_ptr<ParseCall> chekushka;
struct ParseCall{
struct ParseCall{
codeset_t& result;
explicit ParseCall(codeset_t &result) : result(result) {}
virtual ~ParseCall() = default;
virtual chekushka afterReceive(ctx_t& ctx, const CommonCodesets& cc) { assert(false); }
virtual chekushka firstTime(ctx_t& ctx, const CommonCodesets& cc) { assert(false); }
};
};
#define call_ERROR_CHECK do { if (ctx.error) { return NULL; } } while (0)
#define call_THROW(str) do { report(ctx, "square bracket expression: " str); return NULL; } while (0)
/* [...] */
struct ZeroLvl_ParseCall: public ParseCall{
/* [...] */
struct ZeroLvl_ParseCall: public ParseCall{
explicit ZeroLvl_ParseCall(codeset_t &result) : ParseCall(result) {}
chekushka afterReceive(ctx_t &ctx, const CommonCodesets& cc) override;
chekushka firstTime(ctx_t &ctx, const CommonCodesets& cc) override;
};
};
/* ...&...&... */
struct FirstLvl_ParseCall: public ParseCall{
/* ...&...&... */
struct FirstLvl_ParseCall: public ParseCall{
codeset_t ret_buf_for_new;
bool got_one = false;
explicit FirstLvl_ParseCall(codeset_t& result) : ParseCall(result) {}
chekushka afterReceive(ctx_t &ctx, const CommonCodesets& cc) override;
chekushka firstTime(ctx_t &ctx, const CommonCodesets& cc) override;
};
};
/* ab[]vgd[]eyo[]zhz */
struct SecondLvl_ParseCall: public ParseCall{
/* ab[]vgd[]eyo[]zhz */
struct SecondLvl_ParseCall: public ParseCall{
codeset_t ret_buf_for_new;
explicit SecondLvl_ParseCall(codeset_t& result) : ParseCall(result) {}
chekushka afterReceive(ctx_t &ctx, const CommonCodesets& cc) override;
chekushka firstTime(ctx_t &ctx, const CommonCodesets& cc) override;
};
};
/* ^... */
struct CircumflexLvl_ParseCall: public ParseCall{
/* ^... */
struct CircumflexLvl_ParseCall: public ParseCall{
codeset_t ret_buf_for_new;
explicit CircumflexLvl_ParseCall(codeset_t& result) : ParseCall(result) {}
chekushka afterReceive(ctx_t &ctx, const CommonCodesets& cc) override;
chekushka firstTime(ctx_t &ctx, const CommonCodesets& cc) override;
};
};
/* ********* ZeroLvl_ParseCall ********** */
/* ********* ZeroLvl_ParseCall ********** */
chekushka ZeroLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
chekushka ZeroLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
assert(readChar(ctx) == U'[');
return std::make_shared<FirstLvl_ParseCall>(result);
}
}
chekushka ZeroLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
chekushka ZeroLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
if (peep(ctx) != U']')
call_THROW("lvl 0: missing ]");
readChar(ctx);
return NULL;
}
}
/* ********* FirstLvl_ParseCall ********** */
/* ********* FirstLvl_ParseCall ********** */
chekushka FirstLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
chekushka FirstLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
return std::make_shared<SecondLvl_ParseCall>(result);
}
}
chekushka FirstLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
chekushka FirstLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
if (got_one)
result = intersect_sets(result, ret_buf_for_new);
else
@ -91,11 +92,11 @@ chekushka FirstLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc)
return std::make_shared<SecondLvl_ParseCall>(ret_buf_for_new);
}
return NULL;
}
}
/* ********* SecondLvl_ParseCall ********** */
/* ********* SecondLvl_ParseCall ********** */
chekushka SecondLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
chekushka SecondLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
repeat:
int32_t ch = peep(ctx); call_ERROR_CHECK;
if (ch == U'^'){
@ -147,28 +148,28 @@ chekushka SecondLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
goto repeat;
}
return NULL;
}
}
chekushka SecondLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
chekushka SecondLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
result = merge_sets(result, ret_buf_for_new);
return firstTime(ctx, cc);
}
}
/* ********* CircumflexLvl_ParseCall ********* */
/* ********* CircumflexLvl_ParseCall ********* */
chekushka CircumflexLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
chekushka CircumflexLvl_ParseCall::firstTime(ctx_t &ctx, const CommonCodesets& cc) {
assert(readChar(ctx) == U'^');
return std::make_shared<FirstLvl_ParseCall>(ret_buf_for_new);
}
}
chekushka CircumflexLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
chekushka CircumflexLvl_ParseCall::afterReceive(ctx_t &ctx, const CommonCodesets& cc) {
result = invert_set(ret_buf_for_new);
return NULL;
}
}
/* Aaaaaaaaand... The function we have all been waiting for so long! */
codeset_t sq_bracket_expr_parse(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc) {
/* Aaaaaaaaand... The function we have all been waiting for so long! */
codeset_t sq_bracket_expr_parse(REGEX_IS024_MeaningContext &ctx, const CommonCodesets& cc) {
std::vector<std::shared_ptr<ParseCall>> callStack;
codeset_t res;
callStack.push_back(std::make_shared<ZeroLvl_ParseCall>(res));
@ -186,4 +187,5 @@ codeset_t sq_bracket_expr_parse(REGEX_IS024_MeaningContext &ctx, const CommonCod
}
}
return res;
}
}

View File

@ -5,6 +5,7 @@
#include <libregexis024sol/expr_compiler.h>
#include <libregexis024sol/common_codesets.h>
codeset_t sq_bracket_expr_parse(REGEX_IS024_MeaningContext& ctx, const CommonCodesets& cc);
namespace regexis024 {
codeset_t sq_bracket_expr_parse(REGEX_IS024_MeaningContext& ctx, const CommonCodesets& cc);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_SQUARE_BRACKET_EXPRESSION_H

View File

@ -3,11 +3,12 @@
#include <assert.h>
#include <stdio.h>
SubExprCompiled subexpr_charset_reading_filter(const codeset_t &codeset, FA_Container &fa) {
namespace regexis024 {
SubExprCompiled subexpr_charset_reading_filter(const codeset_t &codeset, FA_Container &fa) {
return subexpression_from_path(fa.makeOneCharRead(codeset, false));
}
}
SubExprCompiled join(const SubExprCompiled &A, const SubExprCompiled &B) {
SubExprCompiled join(const SubExprCompiled &A, const SubExprCompiled &B) {
if (!A.start)
return B;
if (!B.start)
@ -19,18 +20,18 @@ SubExprCompiled join(const SubExprCompiled &A, const SubExprCompiled &B) {
res.ends = B.ends;
res.can_be_empty = A.can_be_empty && B.can_be_empty;
return res;
}
}
SubExprCompiled subexpression_from_path(FA_NodePathPart *node) {
SubExprCompiled subexpression_from_path(FA_NodePathPart *node) {
SubExprCompiled res;
res.start = node;
res.ends.push_back(&(node->nxt_node));
/* There is only one char reading path node type */
res.can_be_empty = (node->type != one_char_read);
return res;
}
}
SubExprCompiled RobertAngier(const SubExprCompiled& source, FA_Container& fa) {
SubExprCompiled RobertAngier(const SubExprCompiled& source, FA_Container& fa) {
SubExprCompiled res;
if (!source.start)
return res;
@ -68,9 +69,9 @@ SubExprCompiled RobertAngier(const SubExprCompiled& source, FA_Container& fa) {
for (Marked& mrkd: searched)
mrkd.original->search_mark = -1;
return res;
}
}
void reattach_all_ends_to_one_node(SubExprCompiled& patient, FA_Node* node){
void reattach_all_ends_to_one_node(SubExprCompiled& patient, FA_Node* node){
assert(node);
assert(patient.start);
for (FA_Node** end: patient.ends){
@ -78,9 +79,9 @@ void reattach_all_ends_to_one_node(SubExprCompiled& patient, FA_Node* node){
printf("DEBUG %lu->->->->->%lu\n", patient.start->nodeId, node->nodeId);
reattach_fa_node_edge(end, node);
}
}
}
void apply_repeat_to_subexpression(SubExprCompiled &patient, FA_Container& fa, size_t min_allowed, size_t max_allowed) {
void apply_repeat_to_subexpression(SubExprCompiled &patient, FA_Container& fa, size_t min_allowed, size_t max_allowed) {
assert(min_allowed <= max_allowed && min_allowed <= REGEXIS024_MAX_REPEAT);
if (!patient.start)
return;
@ -144,9 +145,9 @@ void apply_repeat_to_subexpression(SubExprCompiled &patient, FA_Container& fa, s
}
if (min_allowed == 0)
patient.can_be_empty = true;
}
}
SubExprCompiled forkify(const std::vector<SubExprCompiled> &options, FA_Container& fa){
SubExprCompiled forkify(const std::vector<SubExprCompiled> &options, FA_Container& fa){
SubExprCompiled result;
size_t non_empty = 0;
result.can_be_empty = false;
@ -177,8 +178,9 @@ SubExprCompiled forkify(const std::vector<SubExprCompiled> &options, FA_Containe
}
}
return result;
}
}
void SubExprCompiled::assertDefault() {
void SubExprCompiled::assertDefault() {
assert(!start && ends.empty() && can_be_empty);
}
}

View File

@ -3,30 +3,31 @@
#include <libregexis024fa/finite_automaton.h>
struct SubExprCompiled{
namespace regexis024 {
struct SubExprCompiled{
FA_Node* start = NULL;
/* After putting there values from neighbour vectors in nodes, these vectors must not change size */
std::vector<FA_Node**> ends;
bool can_be_empty = true;
void assertDefault();
};
};
SubExprCompiled subexpr_charset_reading_filter(const codeset_t& codeset, FA_Container& fa);
SubExprCompiled subexpr_charset_reading_filter(const codeset_t& codeset, FA_Container& fa);
SubExprCompiled join(const SubExprCompiled& A, const SubExprCompiled& B);
SubExprCompiled join(const SubExprCompiled& A, const SubExprCompiled& B);
SubExprCompiled forkify(const std::vector<SubExprCompiled>& options, FA_Container& fa);
SubExprCompiled forkify(const std::vector<SubExprCompiled>& options, FA_Container& fa);
SubExprCompiled subexpression_from_path(FA_NodePathPart* node);
SubExprCompiled subexpression_from_path(FA_NodePathPart* node);
/* And then Robert Angier said `It's prestige time` and prestiged all over the place.
/* And then Robert Angier said `It's prestige time` and prestiged all over the place.
* If you still don't get it, this function copies section of NFA of regexp */
SubExprCompiled RobertAngier(const SubExprCompiled& source, FA_Container& fa);
SubExprCompiled RobertAngier(const SubExprCompiled& source, FA_Container& fa);
#define REGEXIS024_MAX_REPEAT 64
/* pass REGEXIS024_MAX_REPEAT + 1 as max_allowed to allow infinite repeat */
void apply_repeat_to_subexpression(SubExprCompiled& patient, FA_Container& fa, size_t min_allowed, size_t max_allowed);
/* pass REGEXIS024_MAX_REPEAT + 1 as max_allowed to allow infinite repeat */
void apply_repeat_to_subexpression(SubExprCompiled& patient, FA_Container& fa, size_t min_allowed, size_t max_allowed);
}
#endif //LIBREGEXIS024_SRC_LIBREGEXIS024SOL_SUBEXPR_FA_TRANSFORMED_H

View File

@ -11,15 +11,17 @@
#include <map>
#include <stdio.h>
using namespace regexis024;
struct assembler_context_bookmark{
regex_near_ptr_t pos_in_r024program;
near_ptr_t pos_in_r024program;
int LINE;
};
struct pending_bookmark{
/* Must fill this byte with pos of pos_in_r024program in assembler_context_bookmark
* In a sense, this is a pointer to a NULL pointer that is yet to become normal kinda pointer */
regex_near_ptr_t pos_in_r024program;
near_ptr_t pos_in_r024program;
const char* name;
/* LINE of the reference is needed in case of error */
int LINE;
@ -46,7 +48,7 @@ struct assembler_context{
}
/* pending bookmerk requests should be added only with beg_for_bookmark method,
* or else SEGFAULT will be your frequent guest */
*reinterpret_cast<regex_near_ptr_t *>(&result[br.pos_in_r024program]) = bookmarks[br.name].pos_in_r024program;
*reinterpret_cast<near_ptr_t *>(&result[br.pos_in_r024program]) = bookmarks[br.name].pos_in_r024program;
}
}

View File

@ -11,8 +11,9 @@
#include <map>
#include <stdio.h>
#include <inttypes.h>
#include <stdexcept>
// TODO: apply here my new change in near pointer size
using namespace regexis024;
struct landing_place_resolvance{
size_t name_id;
@ -34,12 +35,14 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
};
uint64_t used_names = 0;
/* From program position -> to names[ind] & */
std::map<regex_near_ptr_t, landing_place_resolvance> bookmarks;
regex_near_ptr_t IP = 0;
std::map<near_ptr_t, landing_place_resolvance> bookmarks;
near_ptr_t IP = 0;
auto check_inboundness = [&](int region){
if (!vmprog_check_inboundness(prgSize, IP, region))
exitf("This program can't be decomposed into commands in a trivial way");
if (!vmprog_check_inboundness(prgSize, IP, region)) {
fprintf(stderr, "This program can't be decomposed into commands in a trivial way");
std::terminate();
}
};
auto extract_b = [&]() -> uint8_t{
check_inboundness(1);
@ -60,19 +63,19 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
auto extract_instruction = [&]() -> uint8_t{
return extract_b();
};
auto extract_sslot_id = [&]() -> regex_sslot_id_t{
auto extract_sslot_id = [&]() -> sslot_id_t{
return extract_dw();
};
auto extract_near_pointer = [&]() -> regex_near_ptr_t{
auto extract_near_pointer = [&]() -> near_ptr_t{
return extract_qw();
};
auto extract_track_array_index = [&]() -> regex_tai_t{
auto extract_track_array_index = [&]() -> tai_t{
return extract_w();
};
bool second_phase = false;
auto fph_register_landing = [&](regex_near_ptr_t pos){
auto fph_register_landing = [&](near_ptr_t pos){
if (!second_phase){
if (bookmarks.count(pos) == 0){
if (used_names == names.size())
@ -83,15 +86,17 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
}
};
auto get_bookmark_in_2phase = [&](regex_near_ptr_t pos) -> std::string {
if (bookmarks.count(pos) == 0)
exitf("bruh");
auto get_bookmark_in_2phase = [&](near_ptr_t pos) -> std::string {
if (bookmarks.count(pos) == 0) {
fprintf(stderr, "Bruh\n");
std::terminate();
}
return names[bookmarks[pos].name_id];
};
auto one_reading = [&](){
while (IP < prgSize) {
regex_near_ptr_t start_pos = IP;
near_ptr_t start_pos = IP;
if (second_phase){
if (bookmarks.count(IP) != 0){
printf("%s:\n", get_bookmark_in_2phase(IP).c_str());
@ -102,11 +107,11 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
switch (opcode) {
#define secPrint(fmt, ...) if (second_phase) {printf("% 3lu) " fmt, start_pos, __VA_ARGS__);} } break;
#define secPrintNoArg(str) if (second_phase) {printf("% 3lu) " str, start_pos);} } break;
#define instCase(oper_code) case regex024_opcodes::oper_code: {
#define instCase(oper_code) case opcodes::oper_code: {
#define jcMess(cond, sz_uppercase, x_t, extract_method, printf_sign) \
instCase(JC ## cond ## _ ## sz_uppercase) \
x_t x = extract_method(); \
regex_near_ptr_t dest = extract_near_pointer(); \
near_ptr_t dest = extract_near_pointer(); \
fph_register_landing(dest); \
secPrint("JC" #cond "_" #sz_uppercase " %" printf_sign " $%s\n", x, get_bookmark_in_2phase(dest).c_str())
#define jcCacaphony(cond) \
@ -131,22 +136,22 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
instCase(FORK)
uint32_t ssid = extract_sslot_id();
regex_near_ptr_t dest = extract_near_pointer();
near_ptr_t dest = extract_near_pointer();
fph_register_landing(dest);
secPrint("FORK %u $%s\n", ssid, get_bookmark_in_2phase(dest).c_str())
simpleDimple(MATCH)
simpleDimple(DIE)
instCase(PARAM_READ_SS_NUMBER)
regex_sslot_id_t ssid_max_plus_one = extract_sslot_id();
sslot_id_t ssid_max_plus_one = extract_sslot_id();
secPrint("PARAM_READ_SS_NUMBER %u\n", ssid_max_plus_one)
instCase(PARAM_FORK_SS_NUMBER)
regex_sslot_id_t ssid_max_plus_one = extract_sslot_id();
sslot_id_t ssid_max_plus_one = extract_sslot_id();
secPrint("PARAM_FORK_SS_NUMBER %u\n", ssid_max_plus_one)
instCase(PARAM_SELARR_LEN)
regex_tai_t tai_max_plus_one = extract_track_array_index();
tai_t tai_max_plus_one = extract_track_array_index();
secPrint("PARAM_SELARR_LEN %hu\n", tai_max_plus_one)
instCase(PARAM_COLSIFTFUNC_SET)
regex_near_ptr_t entry = extract_near_pointer();
near_ptr_t entry = extract_near_pointer();
fph_register_landing(entry);
secPrint("PARAM_COLSIFTFUNC_SET $%s\n", get_bookmark_in_2phase(entry).c_str())
simpleDimple(PARAM_COLSIFTFUNC_WIPE)
@ -156,36 +161,37 @@ void print_disassembly(size_t prgSize, uint8_t* prg){
instCase(MSG_FED_INPUT_EXTENDED)
uint8_t left = extract_b();
uint8_t right = extract_b();
regex_sslot_id_t part = extract_sslot_id();
sslot_id_t part = extract_sslot_id();
secPrint("MSG_FED_INPUT_EXTENDED %hhu %hhu %u\n", left, right, part)
instCase(DMOV_RABX_SELARR)
regex_tai_t i = extract_track_array_index();
tai_t i = extract_track_array_index();
secPrint("DMOV_RABX_SELARR %hu\n", i)
instCase(DDIST_RABX_SELARR)
regex_tai_t s = extract_track_array_index();
regex_tai_t e = extract_track_array_index();
tai_t s = extract_track_array_index();
tai_t e = extract_track_array_index();
secPrint("DDIST_RABX_SELARR %hu %hu\n", s, e);
simpleDimple(SIFTPRIOR_MIN_RABX)
simpleDimple(SIFTPRIOR_MAX_RABX)
simpleDimple(SIFT_DONE)
instCase(MOV_COLARR_IMM)
regex_tai_t tai = extract_track_array_index();
tai_t tai = extract_track_array_index();
uint64_t imm = extract_qw();
secPrint("MOV_COLARR_IMM %hu %lu\n", tai, imm);
instCase(MOV_COLARR_BTPOS)
regex_tai_t tai = extract_track_array_index();
tai_t tai = extract_track_array_index();
secPrint("MOV_COLARR_BTPOS %hu\n", tai);
instCase(MOV_SELARR_IMM)
regex_tai_t tai = extract_track_array_index();
tai_t tai = extract_track_array_index();
uint64_t imm = extract_qw();
secPrint("MOV_SELARR_IMM %hu %lu\n", tai, imm);
instCase(MOV_SELARR_CHPOS)
regex_tai_t tai = extract_track_array_index();
tai_t tai = extract_track_array_index();
secPrint("MOV_SELARR_CHPOS %hu\n", tai);
simpleDimple(INIT)
simpleDimple(THROW)
default:
exitf("Bad opcode\n");
fprintf(stderr, "Bad opcode\n");
std::terminate();
#undef secPrint
#undef secPrintNoArg
#undef instCase

View File

@ -2,12 +2,16 @@
#include <libregexis024vm/utils.h>
#include <stdio.h>
using namespace regexis024;
void test_ccs_fnc(const codeset_t &got, const codeset_t &expected){
static int id = 1;
if (got == expected)
if (got == expected) {
printf("Test %d passed\n", id++);
else
exitf("Test %d failed\n", id);
} else {
printf("Test %d failed\n", id);
std::terminate();
}
}
void invert_test(const codeset_t& A, const codeset_t& C){

View File

@ -10,8 +10,8 @@ static int test_id = 0;
void do_test(const std::vector<uint8_t>& prg, const std::string& str, const std::vector<bool>& prefix_matching){
assert(str.size() + 1 == prefix_matching.size());
REGEX_IS024_CONTEXT ctx{prg.size(), prg.data(), 0, 0, 1000, 1000, 1000000};
regex024_error_code ret;
VMContext ctx{prg.size(), prg.data(), 0, 0, 1000, 1000, 1000000};
error_code_t ret;
// todo
printf("TEST %d passed\n", test_id);
test_id++;

View File

@ -1,6 +1,8 @@
#include <libregexis024sol/expr_compiler.h>
#include <libregexis024test/byte_code_disassembler.h>
using namespace regexis024;
int main(){
std::string regular_expression = "\\>1*";
REGEX_IS024_MeaningContext regex(regular_expression.size(), regular_expression.c_str());

View File

@ -7,6 +7,8 @@
#include <stdexcept>
#include <random>
using namespace regexis024;
struct test_id_t {
int test_id;
int subtest_id;

View File

@ -21,19 +21,54 @@ void test(const string& input, const string& pattern, const MatchInfo& right_ans
}
int main() {
test("11aa", "^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo({}, {}));
test("aa11", "^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo({}, {}));
test("a111", "^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo());
test("aa11", "^!A;\\B!A;\\B!any;\\B!any;$", MatchInfo());
test("1a11", "^!A;\\B!A;\\B!any;\\B!any;$", MatchInfo());
test("11aa", "!dfa;^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo({}, {}));
test("aa11", "!dfa;^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo({}, {}));
test("a111", "!dfa;^!A;\\B!A;\\b!any;\\B!any;$", MatchInfo());
test("aa11", "!dfa;^!A;\\B!A;\\B!any;\\B!any;$", MatchInfo());
test("1a11", "!dfa;^!A;\\B!A;\\B!any;\\B!any;$", MatchInfo());
test("LINE\r\nFirst:Second\r\nThird:12\r\n\r\n",
"!dfa;!select{fieldname{ca}fieldbody{ca}}^^^LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n$$$",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}, {0, 20}, {1, 25}, {2, 26}, {3, 28}}, {20, 25, 26, 28}));
test("LINE\r\nFirst:Second\r\nThird:12\r\n\r\n",
"!dfa;!select{fieldname{ca}fieldbody{ca}}^LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+\\>):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}, {0, 20}, {1, 25}, {2, 26}, {3, 28}}, {20, 25, 26, 28}));
test("LINE\r\nFirst:Second\r\nThird:12\r\n\r\n",
"!dfa;!select{fieldname{ca}fieldbody{ca}}^LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}, {0, 20}, {1, 25}, {2, 26}, {3, 28}}, {20, 25, 26, 28}));
test("LINE\r\nFirst:Second\r\n\r\n",
"!select{fieldname{ca}}LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}}, {6, 11}));
test("LINE\r\nFirst:Second\r\n\r\n",
"!select{fieldname}LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 12}, {1, 18}}, {6, 11}));
test("LINE\r\nFirst:Second\r\nThird:12\r\n\r\n",
"!select{fieldname{ca}fieldbody{ca}}LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}, {0, 20}, {1, 25}, {2, 26}, {3, 28}}, {20, 25, 26, 28}));
test("абвгд", "абвгд", MatchInfo({}, {}));
test("абвввввввгд", "абв*г", MatchInfo({}, {}));
test("абвввввввд", "абв*г", MatchInfo());
test("LINE\r\nFirst:Second\r\nThird:12\r\n\r\n",
"!dfa;^LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}, {0, 20}, {1, 25}, {2, 26}, {3, 28}}, {}));
test("LINE\r\nFirst:Second\r\n\r\n",
"LINE\r\n(#fieldname([\\u0021-\\u007E&^:]+):#fieldbody([\\u0000-\\u007F&^\r\n]*)\r\n)*\r\n",
MatchInfo({{0, 6}, {1, 11}, {2, 12}, {3, 18}}, {}));
test("C111111111111", "C\\>1*", MatchInfo({}, {}));
// return 0;
test("GET / HTTP/1.1\r\nHost: bibura sosat\r\nLos-es-raus: a\rfaafafdf\r\n\r\n",
test("GET / HTTP/1.1\r\nHost: example.com\r\nAAAAA: a\rfaafafdf\r\n\r\n",
"!dfa;(GET|POST) / HTTP/(1.1|1.0|0.9)\r\n([\\u0021-\\u007E&^:]+:([\\u0000-\\u007F&^\r\n])*\r\n)*\r\n",
MatchInfo());
test("\r24234\r\n", "[\\u0000-\\u007F&^\r\n]*\r\n", MatchInfo());
test("\n3432\r\n", "[\\u0000-\\u007F&^\r\n]*\r\n", MatchInfo());
test("3:::;;432\r\n", "[\\u0000-\\u007F&^\r\n]*\r\n", MatchInfo({}, {}));
test("3:::;;432 \r\n", "[\\u0000-\\u007F&^\r\n]*\r\n", MatchInfo({}, {}));
test("GET / HTTP/0.9\r\nHost: bibura sosat\r\nLos-es-raus: afaafafdf\r\n\r\n",
test("GET / HTTP/0.9\r\nHost: bibur at\r\nContent-type: html\r\n\r\n",
"^(GET|POST\\>) / HTTP/(1.1|1.0|0.9)\r\n([\\u0021-\\u007E&^:]+:([\\u0000-\\u007F&^\r\n])*\r\n)*\r\n",
MatchInfo({}, {}));
// return 0;
test("b", "#boba(b)", MatchInfo({{0, 0}, {1, 1}}, {}));
test("abc", "!selarr{boba{ca}}^a#boba(b)c$", MatchInfo({{0, 1}, {1, 2}}, {1, 2}));
for (int i = 0; i < 64; i++) {

View File

@ -7,7 +7,8 @@
// using namespace regexis024;
void convert(regexis024::TrackingVariableInfo& to, const SubtrackingNameInfo& from) {
namespace regexis024 {
void convert(TrackingVariableInfo& to, const SubtrackingNameInfo& from) {
#define plagiat(field) to.field = from.field;
plagiat(type);
plagiat(colarr_first);
@ -17,11 +18,11 @@ void convert(regexis024::TrackingVariableInfo& to, const SubtrackingNameInfo& fr
plagiat(selarr_second);
plagiat(stored_in_sa);
#undef plagiat
}
}
int regexis024::matchStrToRegexp(const std::string& input, const std::string& pattern,
int matchStrToRegexp(const std::string& input, const std::string& pattern,
MatchInfo& retMatchInfo, track_var_list& retTrackVarList, std::string& retStatus)
{
{
retTrackVarList = {};
retMatchInfo = MatchInfo();
retStatus = "";
@ -34,14 +35,14 @@ int regexis024::matchStrToRegexp(const std::string& input, const std::string& pa
for (auto& iip: regexp.ktr.track_names) {
convert(retTrackVarList[iip.first], regexp.ktr.retrieval_info[iip.second]);
}
REGEX_IS024_VirtualMachine vm(regexp.compiled_program.size(), regexp.compiled_program.data(),
VirtualMachine vm(regexp.compiled_program.size(), regexp.compiled_program.data(),
UINT64_MAX, UINT16_MAX,
UINT32_MAX, UINT32_MAX, UINT64_MAX);
auto getVMErrString = [&]() -> std::string {
return std::string(regex024_error_code_tostr(vm.getErrno()));
return std::string(error_code_to_str(vm.getErrno()));
};
if (vm.initialize() != regex024_error_codes::stable) {
if (vm.initialize() != error_codes::stable) {
retStatus = "Virtual machine initialization. " + getVMErrString();
return -1;
}
@ -51,11 +52,11 @@ int regexis024::matchStrToRegexp(const std::string& input, const std::string& pa
retStatus = "Unnatural extended input request.";
return -1;
}
if (vm.addNewMatchingThread() != regex024_error_codes::stable) {
if (vm.addNewMatchingThread() != error_codes::stable) {
retStatus = "Virtual machine first kick. " + getVMErrString();
}
if (left_ext_feed) {
if (vm.extendedFeedCharacter('\n') != regex024_error_codes::stable) {
if (vm.extendedFeedCharacter('\n') != error_codes::stable) {
retStatus = "VM left extended input. " + getVMErrString();
return -1;
}
@ -63,19 +64,19 @@ int regexis024::matchStrToRegexp(const std::string& input, const std::string& pa
for (size_t cur_text_pos = 0;cur_text_pos < input.size();) {
int32_t inp_code;
size_t adj;
utf8_string_iterat(inp_code, adj, cur_text_pos, reinterpret_cast<const uint8_t*>(input.data()), input.size());
utf8_string_iterat(inp_code, adj, cur_text_pos, input.data(), input.size());
if (inp_code < 0) {
retStatus = "Input string encoding error.";
return -1;
}
if (vm.feedCharacter(static_cast<uint64_t>(inp_code), adj) != regex024_error_codes::stable) {
if (vm.feedCharacter(static_cast<uint64_t>(inp_code), adj) != error_codes::stable) {
retStatus = "VM input. " + getVMErrString();
return -1;
}
cur_text_pos += adj;
}
if (right_ext_feed) {
if (vm.extendedFeedCharacter('\n') != regex024_error_codes::stable) {
if (vm.extendedFeedCharacter('\n') != error_codes::stable) {
retStatus = "VM right extended input. " + getVMErrString();
return -1;
}
@ -91,18 +92,19 @@ int regexis024::matchStrToRegexp(const std::string& input, const std::string& pa
std::reverse(retMatchInfo.ca_history.begin(), retMatchInfo.ca_history.end());
}
return 0;
}
}
bool regexis024::MatchInfo::operator==(const MatchInfo &other) const {
bool MatchInfo::operator==(const MatchInfo &other) const {
if (!have_match && !other.have_match)
return true;
return (have_match == other.have_match) && (sa == other.sa) && (ca_history == other.ca_history);
}
}
bool regexis024::MatchInfo::operator!=(const MatchInfo &other) const {
bool MatchInfo::operator!=(const MatchInfo &other) const {
return !(*this == other);
}
}
regexis024::MatchInfo::MatchInfo(const std::vector<REGEX_IS024_CAEvent> &ca_history, const std::vector<uint64_t> &sa):
MatchInfo::MatchInfo(const std::vector<CAEvent> &ca_history, const std::vector<uint64_t> &sa):
ca_history(ca_history), sa(sa), have_match(true) {
}
}

View File

@ -11,7 +11,7 @@ namespace regexis024 {
bool stored_in_ca = true;
bool stored_in_sa = false;
tracking_var_type type;
tracking_var_type_t type;
/* These fields will be -1 if unused */
int colarr_first = -1;
int colarr_second = -1;
@ -24,7 +24,7 @@ namespace regexis024 {
struct MatchInfo {
bool have_match = false;
std::vector<REGEX_IS024_CAEvent> ca_history;
std::vector<CAEvent> ca_history;
std::vector<uint64_t> sa;
bool operator==(const MatchInfo& other) const ;
@ -32,7 +32,7 @@ namespace regexis024 {
MatchInfo() = default;
MatchInfo(const std::vector<REGEX_IS024_CAEvent> &ca_history, const std::vector<uint64_t> &sa);
MatchInfo(const std::vector<CAEvent> &ca_history, const std::vector<uint64_t> &sa);
};
int matchStrToRegexp(const std::string& input, const std::string& pattern,

View File

@ -1,21 +1,22 @@
#include <libregexis024vm/instruction_implementation.h>
#include <stdexcept>
void swap_old_settled_and_new_active(REGEX_IS024_CONTEXT &ctx, REGEX_IS024_Thread& old_settled){
namespace regexis024 {
void swap_old_settled_and_new_active(VMContext &ctx, Thread& old_settled){
ctx_print_debug(ctx);
assert(old_settled.slot_occupation_status == SLOT_OCCUPIED_val);
REGEX_IS024_Thread temp = old_settled;
Thread temp = old_settled;
old_settled = ctx.active_thread;
old_settled.slot_occupation_status = SLOT_NEW_val;
ctx.active_thread = temp;
// slot_occupation_status & SLOT_OCCUPIED of actie thread is true, because it was retrieved from old_settled
}
// slot_occupation_status & SLOT_OCCUPIED of active thread is true, because it was retrieved from old_settled
}
void start_noncloning_conflict(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Thread& other){
void start_noncloning_conflict(VMContext& ctx, Thread& other){
ctx_print_debug(ctx);
if (ctx.have_sift_function){
ctx.sifting_with = &other;
ctx.who_started_sift = regex024_opcode::READ;
ctx.who_started_sift = opcode_t::READ;
ctx.intruder_IP = ctx.active_thread.IP;
ctx.active_thread.IP = ctx.sift_function;
ctx.RAX = ctx.RBX = 0;
@ -23,14 +24,14 @@ void start_noncloning_conflict(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Thread& oth
ctx.active_thread.delete_thread();
ctx.try_to_continue_scheduled();
}
}
}
/* The one that drops as an intruder here is current active.thread.IP */
void start_cloning_conflict(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Thread& other, regex_near_ptr_t clone_IP){
/* The one that drops as an intruder here is current active.thread.IP */
void start_cloning_conflict(VMContext& ctx, Thread& other, near_ptr_t clone_IP){
ctx_print_debug(ctx);
if (ctx.have_sift_function){
ctx.sifting_with = &other;
ctx.who_started_sift = regex024_opcode::FORK;
ctx.who_started_sift = opcode_t::FORK;
ctx.intruder_IP = ctx.active_thread.IP;
ctx.child_ret_IP = clone_IP;
ctx.active_thread.IP = ctx.sift_function;
@ -38,32 +39,32 @@ void start_cloning_conflict(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Thread& other,
} else {
ctx.active_thread.IP = clone_IP;
}
}
}
#define initialization_phase_check() if (ctx.initialized){ \
ctx.error = regex024_error_codes::too_late; return; }
ctx.error = error_codes::too_late; return; }
#define general_matching_mode_check() if (!ctx.initialized){ \
ctx.error = regex024_error_codes::too_early; return; } if(ctx.sifting_with){ \
ctx.error = regex024_error_codes::instruction_not_for_collision_thread; return; }
ctx.error = error_codes::too_early; return; } if(ctx.sifting_with){ \
ctx.error = error_codes::instruction_not_for_collision_thread; return; }
#define sift_mode_check() if (!ctx.sifting_with){ \
ctx.error = regex024_error_codes::instruction_not_for_collision_thread; return; }
ctx.error = error_codes::instruction_not_for_collision_thread; return; }
/* Can append to both read_halted+new stacks of context */
void read_halted_new_type_stacks_append(REGEX_IS024_CONTEXT &ctx, regex_sslot_id_t ssid){
/* Can append to both read_halted+new stacks of context */
void read_halted_new_type_stacks_append(VMContext &ctx, sslot_id_t ssid){
ctx_print_debug(ctx);
if (ssid < ctx.portion_of_FIRST_read_halt_ns){
ctx.READ_halted_stack_new_first.append(ssid);
} else {
ctx.READ_halted_stack_new_second.append(ssid);
}
}
}
void do_i_read(REGEX_IS024_CONTEXT &ctx, regex_sslot_id_t ssid) {
void do_i_read(VMContext &ctx, sslot_id_t ssid) {
ctx_print_debug(ctx);
general_matching_mode_check()
if (ssid >= ctx.read_slots_number)
smitsya(read_sslot_out_of_range);
REGEX_IS024_Thread& other = ctx.READ_halted_slots[ssid];
Thread& other = ctx.READ_halted_slots[ssid];
if (other.slot_occupation_status & SLOT_OCCUPIED){
if (other.slot_occupation_status & SLOT_NEW){
start_noncloning_conflict(ctx, other);
@ -79,60 +80,60 @@ void do_i_read(REGEX_IS024_CONTEXT &ctx, regex_sslot_id_t ssid) {
read_halted_new_type_stacks_append(ctx, ssid);
ctx.try_to_continue_scheduled();
}
}
}
void i_READ(REGEX_IS024_CONTEXT &ctx) {
void i_READ(VMContext &ctx) {
ctx_print_debug(ctx);
check_available_prg(REGEX024_BYTECODE_SSLOT_ID_SZ)
regex_sslot_id_t ssid = ctx.extract_sslot_id();
check_available_prg(BYTECODE_SSLOT_ID_SZ)
sslot_id_t ssid = ctx.extract_sslot_id();
do_i_read(ctx, ssid);
}
}
void i_READZ(REGEX_IS024_CONTEXT &ctx) {
void i_READZ(VMContext &ctx) {
ctx_print_debug(ctx);
do_i_read(ctx, 0);
}
}
void i_JUMP(REGEX_IS024_CONTEXT& ctx){
void i_JUMP(VMContext& ctx){
ctx_print_debug(ctx);
check_available_prg(REGEX024_BYTECODE_NEAR_POINTER_SZ)
check_available_prg(BYTECODE_NEAR_POINTER_SZ)
ctx.active_thread.IP = ctx.extract_near_pointer();
}
}
template<typename conditionT, typename immArgSzT>
void i_JC(REGEX_IS024_CONTEXT& ctx)
{
template<typename conditionT, typename immArgSzT>
void i_JC(VMContext& ctx)
{
ctx_print_debug(ctx);
check_available_prg(immArgSzT::byte_sz + REGEX024_BYTECODE_NEAR_POINTER_SZ);
check_available_prg(immArgSzT::byte_sz + BYTECODE_NEAR_POINTER_SZ);
uint64_t imm_val_B = immArgSzT::extract(ctx);
regex_near_ptr_t dest = ctx.extract_near_pointer();
near_ptr_t dest = ctx.extract_near_pointer();
uint64_t imm_val_A = ctx.INP;
if (conditionT::call(imm_val_A, imm_val_B))
ctx.active_thread.IP = dest;
}
}
struct condEqual{static bool call(uint64_t A, uint64_t B){return A == B;}};
struct condLess{static bool call(uint64_t A, uint64_t B){return A < B;}};
struct condGrtr{static bool call(uint64_t A, uint64_t B){return A > B;}};
struct condEqual{static bool call(uint64_t A, uint64_t B){return A == B;}};
struct condLess{static bool call(uint64_t A, uint64_t B){return A < B;}};
struct condGrtr{static bool call(uint64_t A, uint64_t B){return A > B;}};
struct immArgByte{
struct immArgByte{
static constexpr int byte_sz = 1;
static uint64_t extract(REGEX_IS024_CONTEXT& ctx){return ctx.extract_b();}
};
struct immArgWord{
static uint64_t extract(VMContext& ctx){return ctx.extract_b();}
};
struct immArgWord{
static constexpr int byte_sz = 2;
static uint64_t extract(REGEX_IS024_CONTEXT& ctx){return ctx.extract_w();}
};
struct immArgDoubleWord{
static uint64_t extract(VMContext& ctx){return ctx.extract_w();}
};
struct immArgDoubleWord{
static constexpr int byte_sz = 4;
static uint64_t extract(REGEX_IS024_CONTEXT& ctx){return ctx.extract_dw();}
};
struct immArgQuadWord{
static uint64_t extract(VMContext& ctx){return ctx.extract_dw();}
};
struct immArgQuadWord{
static constexpr int byte_sz = 8;
static uint64_t extract(REGEX_IS024_CONTEXT& ctx){return ctx.extract_qw();}
};
static uint64_t extract(VMContext& ctx){return ctx.extract_qw();}
};
void clone_thread_into_slot(REGEX_IS024_Thread& source, REGEX_IS024_Thread& vessel){
void clone_thread_into_slot(Thread& source, Thread& vessel){
thread_print_debug(source);
my_assert(!(vessel.slot_occupation_status & SLOT_OCCUPIED));
my_assert((source.slot_occupation_status & SLOT_OCCUPIED));
@ -143,18 +144,18 @@ void clone_thread_into_slot(REGEX_IS024_Thread& source, REGEX_IS024_Thread& vess
if (vessel.SAptr){
vessel.SAptr[0]++;
}
}
}
/* One FORK-slot governs the one single unique position in program: the next one after the fork */
void i_FORK(REGEX_IS024_CONTEXT& ctx){
/* One FORK-slot governs the one single unique position in program: the next one after the fork */
void i_FORK(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
check_available_prg(REGEX024_BYTECODE_SSLOT_ID_SZ + REGEX024_BYTECODE_NEAR_POINTER_SZ);
regex_sslot_id_t ssid = ctx.extract_sslot_id();
regex_near_ptr_t dest = ctx.extract_near_pointer();
check_available_prg(BYTECODE_SSLOT_ID_SZ + BYTECODE_NEAR_POINTER_SZ);
sslot_id_t ssid = ctx.extract_sslot_id();
near_ptr_t dest = ctx.extract_near_pointer();
if (ssid >= ctx.fork_slots_number)
smitsya(fork_sslot_out_of_range);
REGEX_IS024_Thread& other = ctx.FORK_halted_slots[ssid];
Thread& other = ctx.FORK_halted_slots[ssid];
if (other.slot_occupation_status & SLOT_OCCUPIED){
start_cloning_conflict(ctx, other, dest);
} else {
@ -162,9 +163,9 @@ void i_FORK(REGEX_IS024_CONTEXT& ctx){
ctx.active_thread.IP = dest;
ctx.FORK_halted_stack.append(ssid);
}
}
}
void i_MATCH(REGEX_IS024_CONTEXT& ctx){
void i_MATCH(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
if (ctx.matched_thread.slot_occupation_status & SLOT_OCCUPIED){
@ -172,108 +173,108 @@ void i_MATCH(REGEX_IS024_CONTEXT& ctx){
} else {
clone_thread_into_slot(ctx.active_thread, ctx.matched_thread);
}
}
}
void i_DIE(REGEX_IS024_CONTEXT& ctx){
void i_DIE(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
ctx.active_thread.delete_thread();
ctx.try_to_continue_scheduled();
}
}
void i_PARAM_READ_SS_NUMBER(REGEX_IS024_CONTEXT& ctx){
void i_PARAM_READ_SS_NUMBER(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(REGEX024_BYTECODE_SSLOT_ID_SZ)
regex_sslot_id_t read_slots_number = ctx.extract_sslot_id();
check_available_prg(BYTECODE_SSLOT_ID_SZ)
sslot_id_t read_slots_number = ctx.extract_sslot_id();
ctx.read_slots_number = read_slots_number;
}
}
void i_PARAM_FORK_SS_NUMBER(REGEX_IS024_CONTEXT& ctx){
void i_PARAM_FORK_SS_NUMBER(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(REGEX024_BYTECODE_SSLOT_ID_SZ)
regex_sslot_id_t fork_slots_number = ctx.extract_sslot_id();
check_available_prg(BYTECODE_SSLOT_ID_SZ)
sslot_id_t fork_slots_number = ctx.extract_sslot_id();
ctx.fork_slots_number = fork_slots_number;
}
}
void i_PARAM_SELARR_LEN(REGEX_IS024_CONTEXT& ctx){
void i_PARAM_SELARR_LEN(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
regex_tai_t selection_array_len = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
tai_t selection_array_len = ctx.extract_track_array_index();
ctx.selection_array_len = selection_array_len;
}
}
void i_PARAM_COLSIFTFUNC_SET(REGEX_IS024_CONTEXT& ctx){
void i_PARAM_COLSIFTFUNC_SET(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(REGEX024_BYTECODE_NEAR_POINTER_SZ)
regex_near_ptr_t sift_function = ctx.extract_near_pointer();
check_available_prg(BYTECODE_NEAR_POINTER_SZ)
near_ptr_t sift_function = ctx.extract_near_pointer();
ctx.have_sift_function = true;
ctx.sift_function = sift_function;
}
}
void i_PARAM_COLSIFTFUNC_WIPE(REGEX_IS024_CONTEXT& ctx){
void i_PARAM_COLSIFTFUNC_WIPE(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
ctx.have_sift_function = false;
}
}
void i_MSG_MULTISTART_ALLOWED(REGEX_IS024_CONTEXT& ctx){
void i_MSG_MULTISTART_ALLOWED(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(1)
ctx.allows_multistart = (bool)ctx.extract_b();
}
}
void i_MSG_FED_INPUT_EXTENDED(REGEX_IS024_CONTEXT& ctx){
void i_MSG_FED_INPUT_EXTENDED(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
check_available_prg(1 + 1 + REGEX024_BYTECODE_SSLOT_ID_SZ)
check_available_prg(1 + 1 + BYTECODE_SSLOT_ID_SZ)
ctx.fed_input_extends_left = ctx.extract_b();
ctx.fed_input_extends_right = ctx.extract_b();
ctx.portion_of_second_read_halt_ns = ctx.extract_sslot_id();
}
}
uint64_t get_el_from_selarr(uint64_t* sa, regex_near_ptr_t ind){
uint64_t get_el_from_selarr(uint64_t* sa, near_ptr_t ind){
return sa ? sa[1UL + ind] : 0;
}
}
void i_DMOV_RABX_SELARR(REGEX_IS024_CONTEXT& ctx){
void i_DMOV_RABX_SELARR(VMContext& ctx){
ctx_print_debug(ctx);
sift_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
regex_tai_t i1 = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
tai_t i1 = ctx.extract_track_array_index();
if (i1 >= ctx.selection_array_len)
smitsya(selection_arr_out_of_range);
ctx.RAX = get_el_from_selarr(ctx.active_thread.SAptr, i1);
ctx.RBX = get_el_from_selarr(ctx.sifting_with->SAptr, i1);
}
}
uint64_t get_selarr_el_dist(uint64_t* sa, uint16_t start, uint16_t end){
uint64_t get_selarr_el_dist(uint64_t* sa, uint16_t start, uint16_t end){
uint64_t v_start = get_el_from_selarr(sa, start);
uint64_t v_end = get_el_from_selarr(sa, end);
return v_end > v_start ? v_end - v_start : 0;
}
}
void i_DDIST_RABX_SELARR(REGEX_IS024_CONTEXT& ctx){
void i_DDIST_RABX_SELARR(VMContext& ctx){
ctx_print_debug(ctx);
sift_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ * 2)
regex_tai_t i_start = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ * 2)
tai_t i_start = ctx.extract_track_array_index();
if (i_start >= ctx.selection_array_len)
smitsya(selection_arr_out_of_range);
regex_tai_t i_end = ctx.extract_track_array_index();
tai_t i_end = ctx.extract_track_array_index();
if (i_end >= ctx.selection_array_len)
smitsya(selection_arr_out_of_range);
ctx.RAX = get_selarr_el_dist(ctx.active_thread.SAptr, i_start, i_end);
ctx.RBX = get_selarr_el_dist(ctx.sifting_with->SAptr, i_start, i_end);
}
}
void finish_conflict_homesteader_wins(REGEX_IS024_CONTEXT& ctx){
void finish_conflict_homesteader_wins(VMContext& ctx){
ctx_print_debug(ctx);
if (ctx.who_started_sift == regex024_opcodes::READ){
if (ctx.who_started_sift == opcodes::READ){
ctx.active_thread.delete_thread();
ctx.try_to_continue_scheduled();
} else {
@ -282,13 +283,13 @@ void finish_conflict_homesteader_wins(REGEX_IS024_CONTEXT& ctx){
ctx.active_thread.IP = ctx.child_ret_IP;
}
ctx.sifting_with = NULL;
}
}
void finish_conflict_intruder_wins(REGEX_IS024_CONTEXT& ctx){
void finish_conflict_intruder_wins(VMContext& ctx){
ctx_print_debug(ctx);
ctx.sifting_with->delete_thread();
ctx.active_thread.IP = ctx.intruder_IP;
if (ctx.who_started_sift == regex024_opcodes::READ){
if (ctx.who_started_sift == opcodes::READ){
/* noncloning conflict won by intruder+ */
*ctx.sifting_with = ctx.active_thread;
ctx.active_thread.slot_occupation_status = SLOT_EMPTY_val;
@ -299,9 +300,9 @@ void finish_conflict_intruder_wins(REGEX_IS024_CONTEXT& ctx){
ctx.active_thread.IP = ctx.child_ret_IP;
}
ctx.sifting_with = NULL;
}
}
void i_SIFTPRIOR_MIN_RABX(REGEX_IS024_CONTEXT& ctx){
void i_SIFTPRIOR_MIN_RABX(VMContext& ctx){
ctx_print_debug(ctx);
sift_mode_check()
if (ctx.RAX < ctx.RBX){
@ -309,9 +310,9 @@ void i_SIFTPRIOR_MIN_RABX(REGEX_IS024_CONTEXT& ctx){
} else if (ctx.RAX > ctx.RBX){
finish_conflict_homesteader_wins(ctx);
}
}
}
void i_SIFTPRIOR_MAX_RABX(REGEX_IS024_CONTEXT& ctx){
void i_SIFTPRIOR_MAX_RABX(VMContext& ctx){
ctx_print_debug(ctx);
sift_mode_check()
if (ctx.RAX > ctx.RBX){
@ -319,45 +320,45 @@ void i_SIFTPRIOR_MAX_RABX(REGEX_IS024_CONTEXT& ctx){
} else if (ctx.RAX < ctx.RBX){
finish_conflict_homesteader_wins(ctx);
}
}
}
void i_SIFT_DONE(REGEX_IS024_CONTEXT& ctx){
void i_SIFT_DONE(VMContext& ctx){
ctx_print_debug(ctx);
sift_mode_check()
finish_conflict_homesteader_wins(ctx);
}
}
/* Can give errors */
void ca_branch_new_node(REGEX_IS024_CONTEXT& ctx, regex_tai_t key, uint64_t val){
/* Can give errors */
void ca_branch_new_node(VMContext& ctx, tai_t key, uint64_t val){
ctx_print_debug(ctx);
if (ctx.CAN_total >= ctx.CA_TREE_LIMIT)
smitsya(ca_tree_limit_violation);
REGEX024_CollectionArrayNode* node = new REGEX024_CollectionArrayNode{key, val, ctx.active_thread.CAHptr, 1};
CollectionArrayNode* node = new CollectionArrayNode{key, val, ctx.active_thread.CAHptr, 1};
// if (ctx.active_thread.CAHptr)
// (ctx.active_thread.CAHptr->refs)++;
ctx.active_thread.CAHptr = node;
ctx.CAN_total++;
}
}
void i_MOV_COLARR_IMM(REGEX_IS024_CONTEXT& ctx){
void i_MOV_COLARR_IMM(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ + 8)
regex_tai_t ca_ind = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ + 8)
tai_t ca_ind = ctx.extract_track_array_index();
uint64_t imm = ctx.extract_qw();
ca_branch_new_node(ctx, ca_ind, imm);
}
}
void i_MOV_COLARR_BTPOS(REGEX_IS024_CONTEXT& ctx){
void i_MOV_COLARR_BTPOS(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
regex_tai_t ca_ind = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
tai_t ca_ind = ctx.extract_track_array_index();
ca_branch_new_node(ctx, ca_ind, ctx.passed_bytes);
}
}
/* Can throw error, should be placed at the end. Call ONLY in general matching mode */
void edit_selection_array(REGEX_IS024_CONTEXT& ctx, uint64_t key, uint64_t val){
/* Can throw error, should be placed at the end. Call ONLY in general matching mode */
void edit_selection_array(VMContext& ctx, uint64_t key, uint64_t val){
ctx_print_debug(ctx);
uint64_t N = ctx.selection_array_len;
if (key >= N)
@ -382,41 +383,42 @@ void edit_selection_array(REGEX_IS024_CONTEXT& ctx, uint64_t key, uint64_t val){
ctx.active_thread.SAptr[0]--;
ctx.active_thread.SAptr = sa_instance;
}
}
}
void i_MOV_SELARR_IMM(REGEX_IS024_CONTEXT& ctx){
void i_MOV_SELARR_IMM(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ + 8)
regex_tai_t sa_ind = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ + 8)
tai_t sa_ind = ctx.extract_track_array_index();
uint64_t imm = ctx.extract_qw();
edit_selection_array(ctx, sa_ind, imm);
}
}
void i_MOV_SELARR_CHPOS(REGEX_IS024_CONTEXT& ctx){
void i_MOV_SELARR_CHPOS(VMContext& ctx){
ctx_print_debug(ctx);
general_matching_mode_check()
check_available_prg(REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
regex_tai_t sa_ind = ctx.extract_track_array_index();
check_available_prg(BYTECODE_TRACK_ARRAY_INDEX_ID_SZ)
tai_t sa_ind = ctx.extract_track_array_index();
edit_selection_array(ctx, sa_ind, ctx.passed_chars);
}
}
void calloc_stack_slots(REGEX_IS024_Stack& stack, regex_sslot_id_t nmemb) {
assert(stack.sz == 0 && !stack.slots);
regex_sslot_id_t* storage = static_cast<regex_sslot_id_t *>(calloc(nmemb, sizeof(regex_sslot_id_t)));
void calloc_stack_slots(SSID_Stack& stack, sslot_id_t nmemb) {
assert(stack.max_size == 0 && stack.sz == 0 && !stack.slots);
sslot_id_t* storage = static_cast<sslot_id_t *>(calloc(nmemb, sizeof(sslot_id_t)));
if (!storage)
throw std::bad_alloc();
stack.slots = storage;
}
stack.max_size = nmemb;
}
REGEX_IS024_Thread* calloc_slots_array(regex_sslot_id_t nmemb) {
REGEX_IS024_Thread* ptr = static_cast<REGEX_IS024_Thread *>(calloc(nmemb, sizeof(REGEX_IS024_Thread)));
Thread* calloc_slots_array(sslot_id_t nmemb) {
Thread* ptr = static_cast<Thread *>(calloc(nmemb, sizeof(Thread)));
if (!ptr)
throw std::bad_alloc();
return ptr;
}
}
void i_INIT(REGEX_IS024_CONTEXT& ctx){
void i_INIT(VMContext& ctx){
ctx_print_debug(ctx);
initialization_phase_check()
if (ctx.selection_array_len > ctx.SA_LEN_LIMIT)
@ -440,21 +442,21 @@ void i_INIT(REGEX_IS024_CONTEXT& ctx){
ctx.initialized = true;
ctx.unnatural_started_thread_IP = ctx.active_thread.IP;
ctx.active_thread.delete_thread();
}
}
void i_THROW(REGEX_IS024_CONTEXT& ctx){
ctx.error = regex024_error_codes::program_throw;
}
void i_THROW(VMContext& ctx){
ctx.error = error_codes::program_throw;
}
void instruction_table(REGEX_IS024_CONTEXT &ctx) {
void instruction_table(VMContext &ctx) {
ctx_print_debug(ctx);
uint8_t opcode = ctx.extract_instruction();
#define rcase(inst) case regex024_opcodes::inst: return i_ ## inst (ctx);
#define jumpC(UN, st) case regex024_opcodes::JC ## UN ## _B: return i_JC<st, immArgByte>(ctx); \
case regex024_opcodes::JC ## UN ## _W: return i_JC<st, immArgWord>(ctx); \
case regex024_opcodes::JC ## UN ## _DW: return i_JC<st, immArgDoubleWord>(ctx); \
case regex024_opcodes::JC ## UN ## _QW: return i_JC<st, immArgQuadWord>(ctx);
#define rcase(inst) case opcodes::inst: return i_ ## inst (ctx);
#define jumpC(UN, st) case opcodes::JC ## UN ## _B: return i_JC<st, immArgByte>(ctx); \
case opcodes::JC ## UN ## _W: return i_JC<st, immArgWord>(ctx); \
case opcodes::JC ## UN ## _DW: return i_JC<st, immArgDoubleWord>(ctx); \
case opcodes::JC ## UN ## _QW: return i_JC<st, immArgQuadWord>(ctx);
switch (opcode) {
rcase(READ)
rcase(READZ)
@ -486,6 +488,7 @@ void instruction_table(REGEX_IS024_CONTEXT &ctx) {
rcase(INIT)
rcase(THROW)
default:
ctx.error = regex024_error_codes::invalid_opcode;
ctx.error = error_codes::invalid_opcode;
}
}
}

View File

@ -7,7 +7,7 @@
#include <libregexis024vm/vm_opcodes.h>
#include <assert.h>
#define smitsya(error_type) do {ctx.error = regex024_error_codes::error_type; return; } while (0)
#define smitsya(error_type) do {ctx.error = error_codes::error_type; return; } while (0)
#define SLOT_EMPTY_val 0
#define SLOT_OCCUPIED 1
@ -16,7 +16,7 @@
#define SLOT_NEW_val (SLOT_OCCUPIED | SLOT_NEW)
#define check_available_prg(regionSz) if (!ctx.check_inboundness(regionSz)){ \
ctx.error = regex024_error_codes::improper_finish; return; }
ctx.error = error_codes::improper_finish; return; }
#if defined(LIBREGEXIS024_DEBUG) && defined(LIBREGEXIS024_ALLOW_LOUD)
@ -30,6 +30,8 @@
#define thread_print_debug(thread)
#endif
void instruction_table(REGEX_IS024_CONTEXT& ctx);
namespace regexis024 {
void instruction_table(VMContext& ctx);
}
#endif //LIBREGEXIS024_INSTRUCTION_IMPLEMENTATION_H

View File

@ -1,10 +1,10 @@
#include <libregexis024vm/vm_opcodes.h>
#include <libregexis024vm/utils.h>
#define rcase(name) case regex024_opcodes::name: return #name;
const char *regex024_opcode_tostr(regex024_opcode x) {
namespace regexis024 {
const char *opcode_to_str(opcode_t x) {
switch (x) {
#define rcase(name) case opcodes::name: return #name;
rcase(READ)
rcase(READZ)
rcase(JUMP)
@ -44,4 +44,5 @@ const char *regex024_opcode_tostr(regex024_opcode x) {
default:
return "Invalid opcode";
}
}
}

View File

@ -2,13 +2,13 @@
#define LIBREGEXIS024_LIBREGEXIS024VM_H
/* This thing is bloated. And slow (Because I designed it imperfectly and because it is bloated).
* I could have halven the amount of bloat, but that would require me writing code in headers.
* I am gonna use it for KM, even more bloated project. So I thought that this design is on the spot.
* C++ is such a funny language. Code is divided into .cpp and .h files. But it only makes problems.
* All of my work on this C++ project was not serious from the beginning. It's all funny stuff. */
* I could have halven the amount of bloat, but that would require me writing code in headers.
* I am gonna use it for KM, even more bloated project. So I thought that this design is on the spot.
* C++ is such a funny language. Code is divided into .cpp and .h files. But it only makes problems.
* All of my work on this C++ project was not serious from the beginning. It's all funny stuff. */
/* Also, please, consider using libregexis024vm/libregexis024vm_interface.h
* Naming in this project is super inconsistent. I don't want it to trash your namespace */
* Naming in this project is super inconsistent. I don't want it to trash your namespace */
#include <libregexis024vm/vm_errno.h>
#include <libregexis024vm/utils.h>
@ -16,37 +16,38 @@
#include <stdlib.h>
#include <stdint.h>
struct REGEX_IS024_Stack{
regex_sslot_id_t* slots = NULL;
regex_sslot_id_t sz = 0;
namespace regexis024 {
struct SSID_Stack{
sslot_id_t* slots = NULL;
sslot_id_t max_size = 0;
sslot_id_t sz = 0;
regex_sslot_id_t pop();
void append(regex_sslot_id_t x);
sslot_id_t pop();
void append(sslot_id_t x);
bool empty() const;
bool non_empty() const;
REGEX_IS024_Stack(const REGEX_IS024_Stack&) = delete;
REGEX_IS024_Stack& operator=(const REGEX_IS024_Stack&) = delete;
REGEX_IS024_Stack() = default;
SSID_Stack(const SSID_Stack&) = delete;
SSID_Stack& operator=(const SSID_Stack&) = delete;
SSID_Stack() = default;
~REGEX_IS024_Stack();
};
~SSID_Stack();
};
struct REGEX024_CollectionArrayNode{
struct CollectionArrayNode{
/* Key is small for historical reasons I do not rememeber. Who cares anyway */
regex_tai_t key;
tai_t key;
uint64_t value;
/* NULL at the beginning */
REGEX024_CollectionArrayNode* prev;
CollectionArrayNode* prev;
/* Reference counting */
uint64_t refs = 0;
};
};
struct REGEX_IS024_Thread{
struct Thread{
/* First byte field is used only when thread is located in slot */
uint8_t slot_occupation_status = 0;
regex_near_ptr_t IP = 0;
REGEX024_CollectionArrayNode* CAHptr = NULL;
near_ptr_t IP = 0;
CollectionArrayNode* CAHptr = NULL;
/* Pointer to the seletion array. SA's are reference counted. Because of that every SA
* is elongated by one meta element in the beginning - reference counter. So the actual elements
* are enumerated starting from one. */
@ -54,20 +55,20 @@ struct REGEX_IS024_Thread{
void delete_thread() noexcept;
void debug_print(const char* place);
};
};
struct REGEX_IS024_CONTEXT{
REGEX_IS024_CONTEXT(size_t programSize, const uint8_t *data, uint64_t caTreeLimit, regex_tai_t saLenLimit,
regex_sslot_id_t readSsLimit, regex_sslot_id_t forkSsLimit, uint64_t timeTickLimit);
struct VMContext{
VMContext(size_t programSize, const uint8_t *data, uint64_t caTreeLimit, tai_t saLenLimit,
sslot_id_t readSsLimit, sslot_id_t forkSsLimit, uint64_t timeTickLimit);
regex024_error_code feedSOF();
error_code_t feedSOF();
/* You can safely pile up calls to this command, nothing bad will happen */
regex024_error_code startThread();
regex024_error_code extendedFeedCharacter(uint64_t input);
regex024_error_code feedCharacter(uint64_t INP, uint64_t corresponding_byte_amount);
error_code_t startThread();
error_code_t extendedFeedCharacter(uint64_t input);
error_code_t feedCharacter(uint64_t INP, uint64_t corresponding_byte_amount);
~REGEX_IS024_CONTEXT();
~VMContext();
/* Program size larger than 2^62 is forbidden */
size_t program_size = 0;
@ -78,9 +79,9 @@ struct REGEX_IS024_CONTEXT{
/* CA = Collecton array. */
uint64_t CA_TREE_LIMIT;
/* SA = Selection array */
regex_tai_t SA_LEN_LIMIT;
regex_sslot_id_t READ_SS_LIMIT;
regex_sslot_id_t FORK_SS_LIMIT;
tai_t SA_LEN_LIMIT;
sslot_id_t READ_SS_LIMIT;
sslot_id_t FORK_SS_LIMIT;
/* If time_tick_limit is non-zero, regex virtual machine will stop with error
* after this many ticks. This parameter set's the timeout.*/
@ -93,38 +94,38 @@ struct REGEX_IS024_CONTEXT{
uint64_t CAN_total = 0;
/* Program selects it */
regex_tai_t selection_array_len = 0;
regex_sslot_id_t read_slots_number = 0;
regex_sslot_id_t fork_slots_number = 0;
tai_t selection_array_len = 0;
sslot_id_t read_slots_number = 0;
sslot_id_t fork_slots_number = 0;
bool have_sift_function = false;
regex_near_ptr_t sift_function;
near_ptr_t sift_function;
bool allows_multistart = false;
uint8_t fed_input_extends_left = 0, fed_input_extends_right = 0;
regex_sslot_id_t portion_of_second_read_halt_ns = 0, portion_of_FIRST_read_halt_ns = 0;
sslot_id_t portion_of_second_read_halt_ns = 0, portion_of_FIRST_read_halt_ns = 0;
bool initialized = false;
regex_near_ptr_t unnatural_started_thread_IP = 1337;
regex024_error_code error = regex024_error_codes::stable;
near_ptr_t unnatural_started_thread_IP = 1337;
error_code_t error = error_codes::stable;
REGEX_IS024_Thread* READ_halted_slots;
REGEX_IS024_Stack READ_halted_stack_old;
REGEX_IS024_Stack READ_halted_stack_new_first;
REGEX_IS024_Stack READ_halted_stack_new_second;
REGEX_IS024_Thread* FORK_halted_slots;
REGEX_IS024_Stack FORK_halted_stack;
Thread* READ_halted_slots;
SSID_Stack READ_halted_stack_old;
SSID_Stack READ_halted_stack_new_first;
SSID_Stack READ_halted_stack_new_second;
Thread* FORK_halted_slots;
SSID_Stack FORK_halted_stack;
REGEX_IS024_Thread active_thread;
Thread active_thread;
/* Environment for sifting stuff */
REGEX_IS024_Thread* sifting_with = NULL;
Thread* sifting_with = NULL;
/* specifies the type of operation vm should do after shift (there are only two distinct options) */
uint8_t who_started_sift;
/* Sifting process uses IP field of active thread. Other data of thread is not modified or used during collision
* procudure. Old IP is stored there, if needed */
regex_near_ptr_t child_ret_IP;
regex_near_ptr_t intruder_IP;
near_ptr_t child_ret_IP;
near_ptr_t intruder_IP;
/* RAX corresponds to intruder. Its data is stored in active thread field*/
uint64_t RAX;
/* RBX corresponds to homesteader. Its data is accessible by `REGEX_IS024_Thread* sifting_with` pointer*/
@ -132,7 +133,7 @@ struct REGEX_IS024_CONTEXT{
/* Will be unoccupied if no threads matched. After each feed of character this field will be wiped
* User should take care of intermediate success himself */
REGEX_IS024_Thread matched_thread;
Thread matched_thread;
uint64_t INP = 0;
uint64_t passed_chars = 0;
@ -148,11 +149,11 @@ struct REGEX_IS024_CONTEXT{
uint64_t extract_qw();
uint8_t extract_instruction();
regex_sslot_id_t extract_sslot_id();
regex_near_ptr_t extract_near_pointer();
regex_tai_t extract_track_array_index();
sslot_id_t extract_sslot_id();
near_ptr_t extract_near_pointer();
tai_t extract_track_array_index();
void debug_print(const char* place);
};
};
}
#endif //LIBREGEXIS024_LIBREGEXIS024VM_H

View File

@ -1,54 +1,54 @@
#include <stdexcept>
#include <libregexis024vm/libregexis024vm.h>
#include <libregexis024vm/instruction_implementation.h>
#include <utility>
regex_sslot_id_t REGEX_IS024_Stack::pop() {
namespace regexis024 {
sslot_id_t SSID_Stack::pop() {
assert(sz != 0);
return slots[--sz];
}
}
void REGEX_IS024_Stack::append(regex_sslot_id_t x) {
void SSID_Stack::append(sslot_id_t x) {
assert(max_size > 0);
assert(slots);
assert(sz < max_size);
slots[sz] = x;
sz++;
}
}
bool REGEX_IS024_Stack::empty() const {
return !non_empty();
}
bool SSID_Stack::empty() const {
return sz == 0;
}
bool REGEX_IS024_Stack::non_empty() const {
return sz;
}
REGEX_IS024_Stack::~REGEX_IS024_Stack() {
SSID_Stack::~SSID_Stack() {
assert(empty());
free(slots);
}
}
REGEX_IS024_CONTEXT::REGEX_IS024_CONTEXT(size_t programSize, const uint8_t *data,
uint64_t caTreeLimit, regex_tai_t saLenLimit,
regex_sslot_id_t readSsLimit, regex_sslot_id_t forkSsLimit,
VMContext::VMContext(size_t programSize, const uint8_t *data,
uint64_t caTreeLimit, tai_t saLenLimit,
sslot_id_t readSsLimit, sslot_id_t forkSsLimit,
uint64_t timeTickLimit) :
program_size(programSize), prg(data), CA_TREE_LIMIT(caTreeLimit), SA_LEN_LIMIT(saLenLimit),
READ_SS_LIMIT(readSsLimit), FORK_SS_LIMIT(forkSsLimit), time_tick_limit(timeTickLimit)
{
{
if (program_size > (1UL << 62))
exitf("Program is too huge\n");
throw std::runtime_error("Program is too big");
active_thread.slot_occupation_status = SLOT_OCCUPIED;
}
}
/* No only will it launch a wave of deallocation in CA tree, but as a nice bonus it's
/* No only will it launch a wave of deallocation in CA tree, but as a nice bonus it's
* gonna deoccupy slot_occupation_status*/
void REGEX_IS024_Thread::delete_thread() noexcept {
void Thread::delete_thread() noexcept {
thread_print_debug(*this);
my_assert(slot_occupation_status & SLOT_OCCUPIED);
slot_occupation_status = SLOT_EMPTY_val;
REGEX024_CollectionArrayNode* cur_CAptr = CAHptr;
CollectionArrayNode* cur_CAptr = CAHptr;
while (cur_CAptr){
assert(cur_CAptr->refs > 0);
if (--(cur_CAptr->refs) == 0){
REGEX024_CollectionArrayNode* next_CAptr = cur_CAptr->prev;
CollectionArrayNode* next_CAptr = cur_CAptr->prev;
delete cur_CAptr;
cur_CAptr = next_CAptr;
} else
@ -58,33 +58,33 @@ void REGEX_IS024_Thread::delete_thread() noexcept {
if (--(SAptr[0]) == 0)
free(SAptr);
}
}
}
void emptify_one_of_new_read_halted_stacks(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Stack& type_new_stack){
while (type_new_stack.non_empty()){
REGEX_IS024_Thread& thread = ctx.READ_halted_slots[type_new_stack.pop()];
void emptify_one_of_new_read_halted_stacks(VMContext& ctx, SSID_Stack& type_new_stack){
while (!type_new_stack.empty()){
Thread& thread = ctx.READ_halted_slots[type_new_stack.pop()];
assert(thread.slot_occupation_status & SLOT_OCCUPIED);
thread.delete_thread();
}
}
}
/* First it will try to pop pending thread from FORK_halted_stack
/* First it will try to pop pending thread from FORK_halted_stack
* Then it will try popping thread from READ_halted_stack_old (checking if top
* thread here is not actually SLOT_NEW). If something succeded, corresponding slot will be deoccupied, and
* active slot will be occupied with it.
*
* try_to_continue_scheduled() assumes that active thread is unoccupied.*/
void REGEX_IS024_CONTEXT::try_to_continue_scheduled(){
void VMContext::try_to_continue_scheduled(){
ctx_print_debug(*this);
my_assert(!(active_thread.slot_occupation_status & SLOT_OCCUPIED));
if (FORK_halted_stack.sz){
regex_sslot_id_t ssid = FORK_halted_stack.pop();
sslot_id_t ssid = FORK_halted_stack.pop();
active_thread = FORK_halted_slots[ssid];
FORK_halted_slots[ssid].slot_occupation_status = SLOT_EMPTY_val;
return;
}
while (READ_halted_stack_old.sz){
regex_sslot_id_t ssid = READ_halted_stack_old.pop();
sslot_id_t ssid = READ_halted_stack_old.pop();
if (READ_halted_slots[ssid].slot_occupation_status & SLOT_NEW){
/* This is the case when old thread was silently replaced by settled new thread */
continue;
@ -94,30 +94,30 @@ void REGEX_IS024_CONTEXT::try_to_continue_scheduled(){
return;
}
/* Failure here will be detected. We started with unoccupied active thread. iterator inside kick will see it */
}
}
void kick(REGEX_IS024_CONTEXT& ctx) {
void kick(VMContext& ctx) {
ctx_print_debug(ctx);
while ((ctx.active_thread.slot_occupation_status & SLOT_OCCUPIED)
&& ctx.error == regex024_error_codes::stable){
&& ctx.error == error_codes::stable){
if (ctx.timer >= ctx.time_tick_limit)
smitsya(timeout);
ctx.timer++;
check_available_prg(REGEX024_BYTECODE_INSTRUCTION_SZ) // May return from kick(ctx)
check_available_prg(BYTECODE_INSTRUCTION_SZ) // May return from kick(ctx)
// smivanie from those instructions will be immediately detected. Everything is OK
instruction_table(ctx);
}
}
}
regex024_error_code REGEX_IS024_CONTEXT::feedSOF() {
error_code_t VMContext::feedSOF() {
ctx_print_debug(*this);
kick(*this);
return error;
}
}
regex024_error_code REGEX_IS024_CONTEXT::startThread() {
error_code_t VMContext::startThread() {
ctx_print_debug(*this);
active_thread.slot_occupation_status = SLOT_OCCUPIED;
active_thread.IP = unnatural_started_thread_IP;
@ -125,31 +125,22 @@ regex024_error_code REGEX_IS024_CONTEXT::startThread() {
active_thread.CAHptr = NULL;
kick(*this);
return error;
}
}
/* I hate C++ (aka antichrist), won't use move sementic (aka drink cornsyrup) */
void swap_stacks(REGEX_IS024_Stack& A, REGEX_IS024_Stack& B) {
std::swap(A.sz, B.sz);
std::swap(A.slots, B.slots);
}
void fill_empty_old_read_halted_stack(REGEX_IS024_CONTEXT& ctx, REGEX_IS024_Stack& read_halted_stack_new){
void fill_empty_old_read_halted_stack(VMContext& ctx, SSID_Stack& read_halted_stack_new){
ctx_print_debug(ctx);
my_assert(!ctx.READ_halted_stack_old.non_empty());
// Actually, READ_halted_stack_old is always empty in this case
assert(ctx.READ_halted_stack_old.empty());
swap_stacks(ctx.READ_halted_stack_old, read_halted_stack_new);
for (uint32_t i = 0; i < ctx.READ_halted_stack_old.sz; i++){
REGEX_IS024_Thread& slot = ctx.READ_halted_slots[ctx.READ_halted_stack_old.slots[i]];
/* Should get rid of 'NEW' qualifier */
assert(slot.slot_occupation_status & SLOT_OCCUPIED);
if (slot.slot_occupation_status & SLOT_OCCUPIED)
slot.slot_occupation_status = SLOT_OCCUPIED;
while (!read_halted_stack_new.empty()) {
sslot_id_t sr = read_halted_stack_new.pop();
Thread& slot = ctx.READ_halted_slots[sr];
assert(slot.slot_occupation_status & SLOT_NEW_val);
slot.slot_occupation_status = SLOT_OCCUPIED_val;
ctx.READ_halted_stack_old.append(sr);
}
}
}
regex024_error_code REGEX_IS024_CONTEXT::feedCharacter(uint64_t input, uint64_t corresponding_byte_amount) {
error_code_t VMContext::feedCharacter(uint64_t input, uint64_t corresponding_byte_amount) {
ctx_print_debug(*this);
if (matched_thread.slot_occupation_status & SLOT_OCCUPIED)
matched_thread.delete_thread();
@ -161,9 +152,9 @@ regex024_error_code REGEX_IS024_CONTEXT::feedCharacter(uint64_t input, uint64_t
try_to_continue_scheduled();
kick(*this);
return error;
}
}
regex024_error_code REGEX_IS024_CONTEXT::extendedFeedCharacter(uint64_t input) {
error_code_t VMContext::extendedFeedCharacter(uint64_t input) {
ctx_print_debug(*this);
if (matched_thread.slot_occupation_status & SLOT_OCCUPIED)
matched_thread.delete_thread();
@ -172,21 +163,21 @@ regex024_error_code REGEX_IS024_CONTEXT::extendedFeedCharacter(uint64_t input) {
try_to_continue_scheduled();
kick(*this);
return error;
}
}
REGEX_IS024_CONTEXT::~REGEX_IS024_CONTEXT() {
VMContext::~VMContext() {
ctx_print_debug(*this);
if (initialized){
emptify_one_of_new_read_halted_stacks(*this, READ_halted_stack_new_first);
emptify_one_of_new_read_halted_stacks(*this, READ_halted_stack_new_second);
while (READ_halted_stack_old.non_empty()){
REGEX_IS024_Thread& thread = READ_halted_slots[READ_halted_stack_old.pop()];
while (!READ_halted_stack_old.empty()){
Thread& thread = READ_halted_slots[READ_halted_stack_old.pop()];
assert(thread.slot_occupation_status & SLOT_OCCUPIED);
if (!(thread.slot_occupation_status & SLOT_NEW))
thread.delete_thread();
}
free(READ_halted_slots);
while (FORK_halted_stack.non_empty())
while (!FORK_halted_stack.empty())
FORK_halted_slots[FORK_halted_stack.pop()].delete_thread();
free(FORK_halted_slots);
@ -194,4 +185,5 @@ REGEX_IS024_CONTEXT::~REGEX_IS024_CONTEXT() {
matched_thread.delete_thread();
}
}
}
}

View File

@ -1,38 +1,40 @@
#include <libregexis024vm/libregexis024vm.h>
#include <libregexis024vm/vm_opcodes.h>
bool REGEX_IS024_CONTEXT::check_inboundness(int region){
namespace regexis024 {
bool VMContext::check_inboundness(int region){
return vmprog_check_inboundness(program_size, active_thread.IP, region);
}
}
uint8_t REGEX_IS024_CONTEXT::extract_b() {
uint8_t VMContext::extract_b() {
return vmprog_extract_b(&active_thread.IP, prg);
}
}
uint16_t REGEX_IS024_CONTEXT::extract_w() {
uint16_t VMContext::extract_w() {
return vmprog_extract_w(&active_thread.IP, prg);
}
}
uint32_t REGEX_IS024_CONTEXT::extract_dw() {
uint32_t VMContext::extract_dw() {
return vmprog_extract_dw(&active_thread.IP, prg);
}
}
uint64_t REGEX_IS024_CONTEXT::extract_qw() {
uint64_t VMContext::extract_qw() {
return vmprog_extract_qw(&active_thread.IP, prg);
}
}
uint8_t REGEX_IS024_CONTEXT::extract_instruction() {
uint8_t VMContext::extract_instruction() {
return extract_b();
}
}
regex_sslot_id_t REGEX_IS024_CONTEXT::extract_sslot_id() {
sslot_id_t VMContext::extract_sslot_id() {
return extract_dw();
}
}
regex_near_ptr_t REGEX_IS024_CONTEXT::extract_near_pointer() {
near_ptr_t VMContext::extract_near_pointer() {
return extract_qw();
}
}
regex_tai_t REGEX_IS024_CONTEXT::extract_track_array_index() {
tai_t VMContext::extract_track_array_index() {
return extract_w();
}
}

View File

@ -1,105 +1,106 @@
#include <stdexcept>
#include <libregexis024vm/libregexis024vm_interface.h>
#include <libregexis024vm/libregexis024vm.h>
#include <libregexis024vm/instruction_implementation.h>
bool REGEX_IS024_CAEvent::operator==(const REGEX_IS024_CAEvent &other) const {
namespace regexis024 {
bool CAEvent::operator==(const CAEvent &other) const {
return (key == other.key) && (value == other.value);
}
}
#define reveal ((REGEX_IS024_CONTEXT*)opaque)
#define reveal ((VMContext*)opaque)
REGEX_IS024_VirtualMachine::REGEX_IS024_VirtualMachine(size_t programSize, const uint8_t *data,
uint64_t caTreeLimit, regex_tai_t saLenLimit,
regex_sslot_id_t readSsLimit, regex_sslot_id_t forkSsLimit,
VirtualMachine::VirtualMachine(size_t programSize, const uint8_t *data,
uint64_t caTreeLimit, tai_t saLenLimit,
sslot_id_t readSsLimit, sslot_id_t forkSsLimit,
uint64_t timeTickLimit) {
opaque = new REGEX_IS024_CONTEXT(programSize, data, caTreeLimit, saLenLimit,
opaque = new VMContext(programSize, data, caTreeLimit, saLenLimit,
readSsLimit, forkSsLimit, timeTickLimit);
}
}
regex024_error_code REGEX_IS024_VirtualMachine::initialize() {
error_code_t VirtualMachine::initialize() {
if (gave_SOF)
exitf("double feedSOF\n");
throw std::runtime_error("double feedSOF\n");
gave_SOF = true;
return reveal->feedSOF();
}
}
bool REGEX_IS024_VirtualMachine::isInitialized() {
bool VirtualMachine::isInitialized() {
return reveal->initialized;
}
}
bool REGEX_IS024_VirtualMachine::isUsable() {
return isInitialized() && reveal->error == regex024_error_codes::stable;
}
bool VirtualMachine::isUsable() {
return isInitialized() && reveal->error == error_codes::stable;
}
REGEX_IS024_VirtualMachine::~REGEX_IS024_VirtualMachine() {
VirtualMachine::~VirtualMachine() {
delete reveal;
}
}
regex_tai_t REGEX_IS024_VirtualMachine::getSelectionArrayLength() {
tai_t VirtualMachine::getSelectionArrayLength() {
return isUsable() ? reveal->selection_array_len : 0;
}
}
bool REGEX_IS024_VirtualMachine::isAllowMultistart() {
bool VirtualMachine::isAllowMultistart() {
return isUsable() ? reveal->allows_multistart : false;
}
}
uint8_t REGEX_IS024_VirtualMachine::getInputLeftExtensionSize() {
uint8_t VirtualMachine::getInputLeftExtensionSize() {
return isUsable() ? reveal->fed_input_extends_left : 0;
}
}
uint8_t REGEX_IS024_VirtualMachine::getInputRightExtensionSize() {
uint8_t VirtualMachine::getInputRightExtensionSize() {
return isUsable() ? reveal->fed_input_extends_right : 0;
}
}
regex024_error_code REGEX_IS024_VirtualMachine::getErrno() {
error_code_t VirtualMachine::getErrno() {
return reveal->error;
}
}
/* Stupid kinda function. Checks if somebody is ready to continue reading the actual string */
bool REGEX_IS024_VirtualMachine::haveSurvivors() {
return isUsable() && (reveal->READ_halted_stack_new_first.non_empty());
}
/* Stupid kinda function. Checks if somebody is ready to continue reading the actual string or extended l-r input */
bool VirtualMachine::haveSurvivors() {
return isUsable() && (!reveal->READ_halted_stack_new_first.empty() || !reveal->READ_halted_stack_new_second.empty());
}
bool REGEX_IS024_VirtualMachine::isMatched() {
bool VirtualMachine::isMatched() {
return isUsable() && static_cast<bool>((reveal->matched_thread.slot_occupation_status & SLOT_OCCUPIED));
}
}
std::vector<REGEX_IS024_CAEvent> REGEX_IS024_VirtualMachine::getMatchedThreadCABranchReverse() {
std::vector<CAEvent> VirtualMachine::getMatchedThreadCABranchReverse() {
if (!isMatched())
return {};
std::vector<REGEX_IS024_CAEvent> res;
REGEX024_CollectionArrayNode* cur = reveal->matched_thread.CAHptr;
std::vector<CAEvent> res;
CollectionArrayNode* cur = reveal->matched_thread.CAHptr;
while (cur != NULL){
res.push_back({cur->key, cur->value});
cur = cur->prev;
}
return res;
}
}
uint64_t REGEX_IS024_VirtualMachine::getMatchedThreadSAValue(uint16_t key) {
uint64_t VirtualMachine::getMatchedThreadSAValue(uint16_t key) {
if (key >= getSelectionArrayLength())
return 0;
if (!isMatched())
return 0;
return reveal->matched_thread.SAptr ? reveal->matched_thread.SAptr[key + 1] : 0;
}
}
regex024_error_code REGEX_IS024_VirtualMachine::addNewMatchingThread() {
error_code_t VirtualMachine::addNewMatchingThread() {
if (!isUsable())
exitf("unusable\n");
// if (started_first_thread && !isAllowMultistart())
// exitf("Multistart is forbidden, bad usage of program\n");
throw std::runtime_error("unusable");
return reveal->startThread();
}
}
regex024_error_code REGEX_IS024_VirtualMachine::extendedFeedCharacter(uint64_t input) {
error_code_t VirtualMachine::extendedFeedCharacter(uint64_t input) {
if (!isUsable())
exitf("unusable\n");
throw std::runtime_error("unusable\n");
return reveal->extendedFeedCharacter(input);
}
}
regex024_error_code REGEX_IS024_VirtualMachine::feedCharacter(uint64_t input, uint64_t bytesResembled) {
error_code_t VirtualMachine::feedCharacter(uint64_t input, uint64_t bytesResembled) {
if (!isUsable())
exitf("unusable\n");
throw std::runtime_error("unusable\n");
return reveal->feedCharacter(input, bytesResembled);
}
}

View File

@ -6,41 +6,42 @@
#include <libregexis024vm/vm_errno.h>
#include <libregexis024vm/vm_opcodes_types.h>
struct REGEX_IS024_CAEvent{
regex_tai_t key;
namespace regexis024 {
struct CAEvent{
tai_t key;
uint64_t value;
bool operator==(const REGEX_IS024_CAEvent& other) const;
};
bool operator==(const CAEvent& other) const;
};
class REGEX_IS024_VirtualMachine{
public:
REGEX_IS024_VirtualMachine(size_t programSize, const uint8_t *data, uint64_t caTreeLimit, uint16_t saLenLimit,
struct VirtualMachine{
VirtualMachine(size_t programSize, const uint8_t *data, uint64_t caTreeLimit, uint16_t saLenLimit,
uint32_t readSsLimit, uint32_t forkSsLimit, uint64_t timeTickLimit);
REGEX_IS024_VirtualMachine(const REGEX_IS024_VirtualMachine& ) = delete;
REGEX_IS024_VirtualMachine& operator=(const REGEX_IS024_VirtualMachine&) = delete;
VirtualMachine(const VirtualMachine& ) = delete;
VirtualMachine& operator=(const VirtualMachine&) = delete;
regex024_error_code initialize();
error_code_t initialize();
bool isInitialized();
bool isUsable();
virtual ~REGEX_IS024_VirtualMachine();
regex_tai_t getSelectionArrayLength();
virtual ~VirtualMachine();
tai_t getSelectionArrayLength();
bool isAllowMultistart();
uint8_t getInputLeftExtensionSize();
uint8_t getInputRightExtensionSize();
regex024_error_code getErrno();
error_code_t getErrno();
bool haveSurvivors();
bool isMatched();
std::vector<REGEX_IS024_CAEvent> getMatchedThreadCABranchReverse();
std::vector<CAEvent> getMatchedThreadCABranchReverse();
uint64_t getMatchedThreadSAValue(uint16_t key);
regex024_error_code addNewMatchingThread();
regex024_error_code extendedFeedCharacter(uint64_t input);
regex024_error_code feedCharacter(uint64_t input, uint64_t bytesResembled);
error_code_t addNewMatchingThread();
error_code_t extendedFeedCharacter(uint64_t input);
error_code_t feedCharacter(uint64_t input, uint64_t bytesResembled);
private:
private:
bool gave_SOF = false;
void* opaque;
};
};
}
#endif //LIBREGEXIS024_LIBREGEXIS024VM_INTERFACE_H

View File

@ -10,35 +10,28 @@
#error "Big endian is currently unsupported"
#endif
void exitf(const char *fmt, ...) {
va_list va;
va_start(va, fmt);
vfprintf(stderr, fmt, va);
va_end(va);
exit(1);
}
int utf8_retrieve_size(uint8_t firstByte) {
if (!(firstByte & 0b10000000))
namespace regexis024 {
int utf8_retrieve_size(char firstByte) {
if (!((uint8_t)firstByte & 0b10000000))
return 1;
uint8_t a = 0b11000000;
uint8_t b = 0b00100000;
for (int i = 2; i <= 4; i++){
if ((firstByte & (a | b)) == a)
if (((uint8_t)firstByte & (a | b)) == a)
return i;
a |= b;
b >>= 1;
}
return -1;
}
}
int32_t utf8_retrieve_character(int sz, size_t pos, const uint8_t *string) {
int32_t utf8_retrieve_character(int sz, size_t pos, const char *string) {
if (sz == 1)
return string[pos];
uint32_t v = string[pos] & (0b01111111 >> sz);
return (uint8_t)string[pos];
uint32_t v = (uint8_t)string[pos] & (0b01111111 >> sz);
pos++;
for (int i = 1; i < sz; i++){
uint32_t th = string[pos];
uint32_t th = (uint8_t)string[pos];
if ((th & 0b11000000) != 0b10000000)
return -1;
v <<= 6;
@ -47,18 +40,16 @@ int32_t utf8_retrieve_character(int sz, size_t pos, const uint8_t *string) {
}
assert(v <= INT32_MAX);
return static_cast<int32_t>(v);
}
}
#define AAAAAA {cp = -1; return;}
void utf8_string_iterat(int32_t &cp, size_t &adj, size_t pos, const uint8_t *string, size_t string_size) {
if (pos >= string_size) AAAAAA
void utf8_string_iterat(int32_t &cp, size_t &adj, size_t pos, const char *string, size_t string_size) {
if (pos >= string_size) {cp = -1; return;}
adj = utf8_retrieve_size(string[pos]);
if (adj < 0 || pos + adj > string_size) AAAAAA
if ((cp = utf8_retrieve_character(adj, pos, string)) < 0) AAAAAA
}
if (adj < 0 || pos + adj > string_size) {cp = -1; return;}
if ((cp = utf8_retrieve_character(adj, pos, string)) < 0) {cp = -1;}
}
bool is_string_in_stringset(const char *strSample, const char **strSet) {
bool is_string_in_stringset(const char *strSample, const char **strSet) {
const char** cmpSubject = strSet;
while ((*cmpSubject) != NULL){
if (strcmp(strSample, *cmpSubject) == 0)
@ -66,4 +57,5 @@ bool is_string_in_stringset(const char *strSample, const char **strSet) {
cmpSubject++; // += 8 bytes
}
return false;
}
}

View File

@ -4,18 +4,19 @@
#include <stdint.h>
#include <stdlib.h>
void exitf(const char* fmt, ...);
// todo: move this file out from my eyes.
namespace regexis024 {
/* 1, 2, 3, 4 on success; -1 on error */
int utf8_retrieve_size(char firstByte);
/* 1, 2, 3, 4 on success; -1 on error */
int utf8_retrieve_size(uint8_t firstByte);
/* sz is a positive value returned by utf8_retrieve_size. Returns negative on error */
int32_t utf8_retrieve_character(int sz, size_t pos, const char* string);
/* sz is a positive value returned by utf8_retrieve_size. Returns negative on error */
int32_t utf8_retrieve_character(int sz, size_t pos, const uint8_t* string);
/* cp is negative on error. adj is the size of letter in bytes. Can be used to adjust pos.
/* cp is negative on error. adj is the size of letter in bytes. Can be used to adjust pos.
* All safety checks will be performed */
void utf8_string_iterat(int32_t& cp, size_t& adj, size_t pos, const uint8_t* string, size_t string_size);
void utf8_string_iterat(int32_t& cp, size_t& adj, size_t pos, const char* string, size_t string_size);
bool is_string_in_stringset(const char* strSample, const char* strSet[]);
bool is_string_in_stringset(const char* strSample, const char* strSet[]);
}
#endif //LIBREGEXIS024_UTILS_H

View File

@ -1,7 +1,8 @@
#include <libregexis024vm/vm_errno.h>
const char *regex024_error_code_tostr(regex024_error_code x) {
#define rcase(name) case regex024_error_codes::name: return #name;
namespace regexis024 {
const char *error_code_to_str(error_code_t x) {
#define rcase(name) case error_codes::name: return #name;
switch (x) {
rcase(stable)
rcase(ca_tree_limit_violation)
@ -23,4 +24,5 @@ const char *regex024_error_code_tostr(regex024_error_code x) {
default:
return "unknown_error_code";
}
}
}

View File

@ -3,7 +3,8 @@
#include <stdint.h>
namespace regex024_error_codes {
namespace regexis024 {
namespace error_codes {
enum regex024_error_code_I: int {
stable = 0,
ca_tree_limit_violation = -1,
@ -36,10 +37,11 @@ namespace regex024_error_codes {
/* O_o */
bad_alloc = -17,
};
}
typedef error_codes::regex024_error_code_I error_code_t;
const char* error_code_to_str(error_code_t x);
}
typedef regex024_error_codes::regex024_error_code_I regex024_error_code;
const char* regex024_error_code_tostr(regex024_error_code x);
#endif //LIBREGEXIS024_VM_ERRNO_H

View File

@ -3,7 +3,8 @@
#include <libregexis024vm/vm_opcodes_types.h>
namespace regex024_opcodes {
namespace regexis024 {
namespace opcodes {
enum regex024_opcode_I: uint8_t{
/* READ <Settlement ID> */
READ = 0,
@ -72,28 +73,27 @@ namespace regex024_opcodes {
THROW = 35,
regex024_opcode_greaterMax = 36
};
}
typedef opcodes::regex024_opcode_I opcode_t;
const char* opcode_to_str(opcode_t x);
constexpr uint64_t BYTECODE_INSTRUCTION_SZ = 1;
constexpr uint64_t BYTECODE_SSLOT_ID_SZ = 4;
constexpr uint64_t BYTECODE_TRACK_ARRAY_INDEX_ID_SZ = 2;
constexpr uint64_t BYTECODE_NEAR_POINTER_SZ = 8;
bool vmprog_check_inboundness(near_ptr_t prgSize, near_ptr_t IP, near_ptr_t region);
uint8_t vmprog_extract_b(near_ptr_t* IPptr, const uint8_t* prg);
uint16_t vmprog_extract_w(near_ptr_t* IPptr, const uint8_t* prg);
uint32_t vmprog_extract_dw(near_ptr_t* IPptr, const uint8_t* prg);
uint64_t vmprog_extract_qw(near_ptr_t* IPptr, const uint8_t* prg);
uint8_t vmprog_extract_instruction(near_ptr_t* IPptr, const uint8_t* prg);
sslot_id_t vmprog_extract_sslot_id(near_ptr_t* IPptr, const uint8_t* prg);
near_ptr_t vmprog_extract_near_pointer(near_ptr_t* IPptr, const uint8_t* prg);
tai_t vmprog_extrack_track_array_index(near_ptr_t* IPptr, const uint8_t* prg);
}
typedef regex024_opcodes::regex024_opcode_I regex024_opcode;
const char* regex024_opcode_tostr(regex024_opcode x);
constexpr uint64_t REGEX024_BYTECODE_INSTRUCTION_SZ = 1;
constexpr uint64_t REGEX024_BYTECODE_SSLOT_ID_SZ = 4;
constexpr uint64_t REGEX024_BYTECODE_TRACK_ARRAY_INDEX_ID_SZ = 2;
constexpr uint64_t REGEX024_BYTECODE_NEAR_POINTER_SZ = 8;
bool vmprog_check_inboundness(regex_near_ptr_t prgSize, regex_near_ptr_t IP, regex_near_ptr_t region);
uint8_t vmprog_extract_b(regex_near_ptr_t* IPptr, const uint8_t* prg);
uint16_t vmprog_extract_w(regex_near_ptr_t* IPptr, const uint8_t* prg);
uint32_t vmprog_extract_dw(regex_near_ptr_t* IPptr, const uint8_t* prg);
uint64_t vmprog_extract_qw(regex_near_ptr_t* IPptr, const uint8_t* prg);
uint8_t vmprog_extract_instruction(regex_near_ptr_t* IPptr, const uint8_t* prg);
regex_sslot_id_t vmprog_extract_sslot_id(regex_near_ptr_t* IPptr, const uint8_t* prg);
regex_near_ptr_t vmprog_extract_near_pointer(regex_near_ptr_t* IPptr, const uint8_t* prg);
regex_tai_t vmprog_extrack_track_array_index(regex_near_ptr_t* IPptr, const uint8_t* prg);
#endif //LIBREGEXIS024_VM_OPCODES_H

View File

@ -1,47 +1,54 @@
#include <libregexis024vm/vm_opcodes.h>
#ifndef __ORDER_LITTLE_ENDIAN__
#error "Big endian is currently unsupported"
#endif
bool vmprog_check_inboundness(regex_near_ptr_t prgSz, regex_near_ptr_t IP, regex_near_ptr_t region) {
namespace regexis024 {
bool vmprog_check_inboundness(near_ptr_t prgSz, near_ptr_t IP, near_ptr_t region) {
return IP + region <= prgSz;
}
}
uint8_t vmprog_extract_b(regex_near_ptr_t *IPptr, const uint8_t *prg) {
uint8_t vmprog_extract_b(near_ptr_t *IPptr, const uint8_t *prg) {
return prg[(*IPptr)++];
}
}
uint16_t vmprog_extract_w(regex_near_ptr_t *IPptr, const uint8_t *prg) {
uint16_t answer = *(uint16_t*)(&prg[*IPptr]);
*IPptr += 2;
uint16_t vmprog_extract_w(near_ptr_t *IPptr, const uint8_t *prg) {
uint16_t answer = 0;
(*IPptr) += 2;
for (int i = 1; i < 3; i++) {
answer <<= 8; answer |= prg[(*IPptr) - i];
}
return answer;
}
}
uint32_t vmprog_extract_dw(regex_near_ptr_t *IPptr, const uint8_t *prg) {
uint32_t answer = *(uint32_t *)(&prg[*IPptr]);
*IPptr += 4;
uint32_t vmprog_extract_dw(near_ptr_t *IPptr, const uint8_t *prg) {
uint32_t answer = 0;
(*IPptr) += 4;
for (int i = 1; i < 5; i++) {
answer <<= 8; answer |= prg[(*IPptr) - i];
}
return answer;
}
}
uint64_t vmprog_extract_qw(regex_near_ptr_t *IPptr, const uint8_t *prg) {
uint64_t answer = *(uint64_t *)(&prg[*IPptr]);
*IPptr += 8;
uint64_t vmprog_extract_qw(near_ptr_t *IPptr, const uint8_t *prg) {
uint64_t answer = 0;
(*IPptr) += 8;
for (int i = 1; i < 9; i++) {
answer <<= 8; answer |= prg[(*IPptr) - i];
}
return answer;
}
}
uint8_t vmprog_extract_instruction(regex_near_ptr_t *IPptr, const uint8_t *prg) {
uint8_t vmprog_extract_instruction(near_ptr_t *IPptr, const uint8_t *prg) {
return vmprog_extract_b(IPptr, prg);
}
}
regex_sslot_id_t vmprog_extract_sslot_id(regex_near_ptr_t *IPptr, const uint8_t *prg) {
sslot_id_t vmprog_extract_sslot_id(near_ptr_t *IPptr, const uint8_t *prg) {
return vmprog_extract_dw(IPptr, prg);
}
}
regex_near_ptr_t vmprog_extract_near_pointer(regex_near_ptr_t *IPptr, const uint8_t *prg) {
near_ptr_t vmprog_extract_near_pointer(near_ptr_t *IPptr, const uint8_t *prg) {
return vmprog_extract_qw(IPptr, prg);
}
}
regex_tai_t vmprog_extrack_track_array_index(regex_near_ptr_t *IPptr, const uint8_t *prg) {
tai_t vmprog_extrack_track_array_index(near_ptr_t *IPptr, const uint8_t *prg) {
return vmprog_extract_w(IPptr, prg);
}
}

View File

@ -3,9 +3,10 @@
#include <stdint.h>
typedef uint32_t regex_sslot_id_t;
typedef uint64_t regex_near_ptr_t;
typedef uint16_t regex_tai_t;
namespace regexis024 {
typedef uint32_t sslot_id_t;
typedef uint64_t near_ptr_t;
typedef uint16_t tai_t;
}
#endif //VM_OPCODES_TYPES_H