diff --git a/Doc/whatsnew/3.11.rst b/Doc/whatsnew/3.11.rst index 9f7f6f52a8e..a5a52682b50 100644 --- a/Doc/whatsnew/3.11.rst +++ b/Doc/whatsnew/3.11.rst @@ -520,6 +520,12 @@ Optimizations becomes 272 bytes from 352 bytes on 64bit platform. (Contributed by Inada Naoki in :issue:`46845`.) +* :mod:`re`'s regular expression matching engine has been partially refactored, + and now uses computed gotos (or "threaded code") on supported platforms. As a + result, Python 3.11 executes the `pyperformance regular expression benchmarks + `_ up to 10% + faster than Python 3.10. + Faster CPython ============== diff --git a/Makefile.pre.in b/Makefile.pre.in index d89886ddc76..f6c8c72bbc2 100644 --- a/Makefile.pre.in +++ b/Makefile.pre.in @@ -1351,11 +1351,12 @@ regen-stdlib-module-names: build_all Programs/_testembed $(UPDATE_FILE) $(srcdir)/Python/stdlib_module_names.h $(srcdir)/Python/stdlib_module_names.h.new regen-sre: - # Regenerate Modules/_sre/sre_constants.h from Lib/re/_constants.py - # using Tools/scripts/generate_sre_constants.py + # Regenerate Modules/_sre/sre_constants.h and Modules/_sre/sre_targets.h + # from Lib/re/_constants.py using Tools/scripts/generate_sre_constants.py $(PYTHON_FOR_REGEN) $(srcdir)/Tools/scripts/generate_sre_constants.py \ $(srcdir)/Lib/re/_constants.py \ - $(srcdir)/Modules/_sre/sre_constants.h + $(srcdir)/Modules/_sre/sre_constants.h \ + $(srcdir)/Modules/_sre/sre_targets.h Python/compile.o Python/symtable.o Python/ast_unparse.o Python/ast.o Python/future.o: $(srcdir)/Include/internal/pycore_ast.h diff --git a/Misc/NEWS.d/next/Library/2022-04-12-19-08-13.gh-issue-91404.zjqYHo.rst b/Misc/NEWS.d/next/Library/2022-04-12-19-08-13.gh-issue-91404.zjqYHo.rst new file mode 100644 index 00000000000..58464fceeba --- /dev/null +++ b/Misc/NEWS.d/next/Library/2022-04-12-19-08-13.gh-issue-91404.zjqYHo.rst @@ -0,0 +1,3 @@ +Improve the performance of :mod:`re` matching by using computed gotos (or +"threaded code") on supported platforms and removing expensive pointer +indirections. diff --git a/Modules/_sre/sre_lib.h b/Modules/_sre/sre_lib.h index 34cd0552532..3472e65b87a 100644 --- a/Modules/_sre/sre_lib.h +++ b/Modules/_sre/sre_lib.h @@ -485,16 +485,20 @@ do { \ #define JUMP_ATOMIC_GROUP 16 #define DO_JUMPX(jumpvalue, jumplabel, nextpattern, toplevel_) \ + ctx->pattern = pattern; \ + ctx->ptr = ptr; \ DATA_ALLOC(SRE(match_context), nextctx); \ nextctx->last_ctx_pos = ctx_pos; \ nextctx->jump = jumpvalue; \ nextctx->pattern = nextpattern; \ nextctx->toplevel = toplevel_; \ + pattern = nextpattern; \ ctx_pos = alloc_pos; \ ctx = nextctx; \ goto entrance; \ jumplabel: \ - while (0) /* gcc doesn't like labels at end of scopes */ \ + pattern = ctx->pattern; \ + ptr = ctx->ptr; #define DO_JUMP(jumpvalue, jumplabel, nextpattern) \ DO_JUMPX(jumpvalue, jumplabel, nextpattern, ctx->toplevel) @@ -517,6 +521,36 @@ typedef struct { int toplevel; } SRE(match_context); +#define MAYBE_CHECK_SIGNALS \ + do { \ + if ((0 == (++sigcount & 0xfff)) && PyErr_CheckSignals()) { \ + RETURN_ERROR(SRE_ERROR_INTERRUPTED); \ + } \ + } while (0) + +#ifdef HAVE_COMPUTED_GOTOS + #ifndef USE_COMPUTED_GOTOS + #define USE_COMPUTED_GOTOS 1 + #endif +#elif defined(USE_COMPUTED_GOTOS) && USE_COMPUTED_GOTOS + #error "Computed gotos are not supported on this compiler." +#else + #undef USE_COMPUTED_GOTOS + #define USE_COMPUTED_GOTOS 0 +#endif + +#if USE_COMPUTED_GOTOS + #define TARGET(OP) TARGET_ ## OP + #define DISPATCH \ + do { \ + MAYBE_CHECK_SIGNALS; \ + goto *sre_targets[*pattern++]; \ + } while (0) +#else + #define TARGET(OP) case OP + #define DISPATCH goto dispatch +#endif + /* check if string matches the given pattern. returns <0 for error, 0 for failure, and 1 for success */ LOCAL(Py_ssize_t) @@ -536,38 +570,44 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) DATA_ALLOC(SRE(match_context), ctx); ctx->last_ctx_pos = -1; ctx->jump = JUMP_NONE; - ctx->pattern = pattern; ctx->toplevel = toplevel; ctx_pos = alloc_pos; +#if USE_COMPUTED_GOTOS +#include "sre_targets.h" +#endif + entrance: - ctx->ptr = (SRE_CHAR *)state->ptr; + ; // Fashion statement. + const SRE_CHAR *ptr = (SRE_CHAR *)state->ptr; - if (ctx->pattern[0] == SRE_OP_INFO) { + if (pattern[0] == SRE_OP_INFO) { /* optimization info block */ /* <1=skip> <2=flags> <3=min> ... */ - if (ctx->pattern[3] && (uintptr_t)(end - ctx->ptr) < ctx->pattern[3]) { + if (pattern[3] && (uintptr_t)(end - ptr) < pattern[3]) { TRACE(("reject (got %zd chars, need %zd)\n", - end - ctx->ptr, (Py_ssize_t) ctx->pattern[3])); + end - ptr, (Py_ssize_t) pattern[3])); RETURN_FAILURE; } - ctx->pattern += ctx->pattern[1] + 1; + pattern += pattern[1] + 1; } - for (;;) { - ++sigcount; - if ((0 == (sigcount & 0xfff)) && PyErr_CheckSignals()) - RETURN_ERROR(SRE_ERROR_INTERRUPTED); +#if USE_COMPUTED_GOTOS + DISPATCH; +#else +dispatch: + MAYBE_CHECK_SIGNALS; + switch (*pattern++) +#endif + { - switch (*ctx->pattern++) { - - case SRE_OP_MARK: + TARGET(SRE_OP_MARK): /* set mark */ /* */ - TRACE(("|%p|%p|MARK %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - i = ctx->pattern[0]; + TRACE(("|%p|%p|MARK %d\n", pattern, + ptr, pattern[0])); + i = pattern[0]; if (i & 1) state->lastindex = i/2 + 1; if (i > state->lastmark) { @@ -580,210 +620,210 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) state->mark[j++] = NULL; state->lastmark = i; } - state->mark[i] = ctx->ptr; - ctx->pattern++; - break; + state->mark[i] = ptr; + pattern++; + DISPATCH; - case SRE_OP_LITERAL: + TARGET(SRE_OP_LITERAL): /* match literal string */ /* */ - TRACE(("|%p|%p|LITERAL %d\n", ctx->pattern, - ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end || (SRE_CODE) ctx->ptr[0] != ctx->pattern[0]) + TRACE(("|%p|%p|LITERAL %d\n", pattern, + ptr, *pattern)); + if (ptr >= end || (SRE_CODE) ptr[0] != pattern[0]) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_NOT_LITERAL: + TARGET(SRE_OP_NOT_LITERAL): /* match anything that is not literal character */ /* */ - TRACE(("|%p|%p|NOT_LITERAL %d\n", ctx->pattern, - ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end || (SRE_CODE) ctx->ptr[0] == ctx->pattern[0]) + TRACE(("|%p|%p|NOT_LITERAL %d\n", pattern, + ptr, *pattern)); + if (ptr >= end || (SRE_CODE) ptr[0] == pattern[0]) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_SUCCESS: + TARGET(SRE_OP_SUCCESS): /* end of pattern */ - TRACE(("|%p|%p|SUCCESS\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|SUCCESS\n", pattern, ptr)); if (ctx->toplevel && - ((state->match_all && ctx->ptr != state->end) || - (state->must_advance && ctx->ptr == state->start))) + ((state->match_all && ptr != state->end) || + (state->must_advance && ptr == state->start))) { RETURN_FAILURE; } - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_SUCCESS; - case SRE_OP_AT: + TARGET(SRE_OP_AT): /* match at given position */ /* */ - TRACE(("|%p|%p|AT %d\n", ctx->pattern, ctx->ptr, *ctx->pattern)); - if (!SRE(at)(state, ctx->ptr, *ctx->pattern)) + TRACE(("|%p|%p|AT %d\n", pattern, ptr, *pattern)); + if (!SRE(at)(state, ptr, *pattern)) RETURN_FAILURE; - ctx->pattern++; - break; + pattern++; + DISPATCH; - case SRE_OP_CATEGORY: + TARGET(SRE_OP_CATEGORY): /* match at given category */ /* */ - TRACE(("|%p|%p|CATEGORY %d\n", ctx->pattern, - ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end || !sre_category(ctx->pattern[0], ctx->ptr[0])) + TRACE(("|%p|%p|CATEGORY %d\n", pattern, + ptr, *pattern)); + if (ptr >= end || !sre_category(pattern[0], ptr[0])) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_ANY: + TARGET(SRE_OP_ANY): /* match anything (except a newline) */ /* */ - TRACE(("|%p|%p|ANY\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end || SRE_IS_LINEBREAK(ctx->ptr[0])) + TRACE(("|%p|%p|ANY\n", pattern, ptr)); + if (ptr >= end || SRE_IS_LINEBREAK(ptr[0])) RETURN_FAILURE; - ctx->ptr++; - break; + ptr++; + DISPATCH; - case SRE_OP_ANY_ALL: + TARGET(SRE_OP_ANY_ALL): /* match anything */ /* */ - TRACE(("|%p|%p|ANY_ALL\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end) + TRACE(("|%p|%p|ANY_ALL\n", pattern, ptr)); + if (ptr >= end) RETURN_FAILURE; - ctx->ptr++; - break; + ptr++; + DISPATCH; - case SRE_OP_IN: + TARGET(SRE_OP_IN): /* match set member (or non_member) */ /* */ - TRACE(("|%p|%p|IN\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end || - !SRE(charset)(state, ctx->pattern + 1, *ctx->ptr)) + TRACE(("|%p|%p|IN\n", pattern, ptr)); + if (ptr >= end || + !SRE(charset)(state, pattern + 1, *ptr)) RETURN_FAILURE; - ctx->pattern += ctx->pattern[0]; - ctx->ptr++; - break; + pattern += pattern[0]; + ptr++; + DISPATCH; - case SRE_OP_LITERAL_IGNORE: + TARGET(SRE_OP_LITERAL_IGNORE): TRACE(("|%p|%p|LITERAL_IGNORE %d\n", - ctx->pattern, ctx->ptr, ctx->pattern[0])); - if (ctx->ptr >= end || - sre_lower_ascii(*ctx->ptr) != *ctx->pattern) + pattern, ptr, pattern[0])); + if (ptr >= end || + sre_lower_ascii(*ptr) != *pattern) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_LITERAL_UNI_IGNORE: + TARGET(SRE_OP_LITERAL_UNI_IGNORE): TRACE(("|%p|%p|LITERAL_UNI_IGNORE %d\n", - ctx->pattern, ctx->ptr, ctx->pattern[0])); - if (ctx->ptr >= end || - sre_lower_unicode(*ctx->ptr) != *ctx->pattern) + pattern, ptr, pattern[0])); + if (ptr >= end || + sre_lower_unicode(*ptr) != *pattern) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_LITERAL_LOC_IGNORE: + TARGET(SRE_OP_LITERAL_LOC_IGNORE): TRACE(("|%p|%p|LITERAL_LOC_IGNORE %d\n", - ctx->pattern, ctx->ptr, ctx->pattern[0])); - if (ctx->ptr >= end - || !char_loc_ignore(*ctx->pattern, *ctx->ptr)) + pattern, ptr, pattern[0])); + if (ptr >= end + || !char_loc_ignore(*pattern, *ptr)) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_NOT_LITERAL_IGNORE: + TARGET(SRE_OP_NOT_LITERAL_IGNORE): TRACE(("|%p|%p|NOT_LITERAL_IGNORE %d\n", - ctx->pattern, ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end || - sre_lower_ascii(*ctx->ptr) == *ctx->pattern) + pattern, ptr, *pattern)); + if (ptr >= end || + sre_lower_ascii(*ptr) == *pattern) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_NOT_LITERAL_UNI_IGNORE: + TARGET(SRE_OP_NOT_LITERAL_UNI_IGNORE): TRACE(("|%p|%p|NOT_LITERAL_UNI_IGNORE %d\n", - ctx->pattern, ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end || - sre_lower_unicode(*ctx->ptr) == *ctx->pattern) + pattern, ptr, *pattern)); + if (ptr >= end || + sre_lower_unicode(*ptr) == *pattern) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_NOT_LITERAL_LOC_IGNORE: + TARGET(SRE_OP_NOT_LITERAL_LOC_IGNORE): TRACE(("|%p|%p|NOT_LITERAL_LOC_IGNORE %d\n", - ctx->pattern, ctx->ptr, *ctx->pattern)); - if (ctx->ptr >= end - || char_loc_ignore(*ctx->pattern, *ctx->ptr)) + pattern, ptr, *pattern)); + if (ptr >= end + || char_loc_ignore(*pattern, *ptr)) RETURN_FAILURE; - ctx->pattern++; - ctx->ptr++; - break; + pattern++; + ptr++; + DISPATCH; - case SRE_OP_IN_IGNORE: - TRACE(("|%p|%p|IN_IGNORE\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end - || !SRE(charset)(state, ctx->pattern+1, - (SRE_CODE)sre_lower_ascii(*ctx->ptr))) + TARGET(SRE_OP_IN_IGNORE): + TRACE(("|%p|%p|IN_IGNORE\n", pattern, ptr)); + if (ptr >= end + || !SRE(charset)(state, pattern+1, + (SRE_CODE)sre_lower_ascii(*ptr))) RETURN_FAILURE; - ctx->pattern += ctx->pattern[0]; - ctx->ptr++; - break; + pattern += pattern[0]; + ptr++; + DISPATCH; - case SRE_OP_IN_UNI_IGNORE: - TRACE(("|%p|%p|IN_UNI_IGNORE\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end - || !SRE(charset)(state, ctx->pattern+1, - (SRE_CODE)sre_lower_unicode(*ctx->ptr))) + TARGET(SRE_OP_IN_UNI_IGNORE): + TRACE(("|%p|%p|IN_UNI_IGNORE\n", pattern, ptr)); + if (ptr >= end + || !SRE(charset)(state, pattern+1, + (SRE_CODE)sre_lower_unicode(*ptr))) RETURN_FAILURE; - ctx->pattern += ctx->pattern[0]; - ctx->ptr++; - break; + pattern += pattern[0]; + ptr++; + DISPATCH; - case SRE_OP_IN_LOC_IGNORE: - TRACE(("|%p|%p|IN_LOC_IGNORE\n", ctx->pattern, ctx->ptr)); - if (ctx->ptr >= end - || !SRE(charset_loc_ignore)(state, ctx->pattern+1, *ctx->ptr)) + TARGET(SRE_OP_IN_LOC_IGNORE): + TRACE(("|%p|%p|IN_LOC_IGNORE\n", pattern, ptr)); + if (ptr >= end + || !SRE(charset_loc_ignore)(state, pattern+1, *ptr)) RETURN_FAILURE; - ctx->pattern += ctx->pattern[0]; - ctx->ptr++; - break; + pattern += pattern[0]; + ptr++; + DISPATCH; - case SRE_OP_JUMP: - case SRE_OP_INFO: + TARGET(SRE_OP_JUMP): + TARGET(SRE_OP_INFO): /* jump forward */ /* */ - TRACE(("|%p|%p|JUMP %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - ctx->pattern += ctx->pattern[0]; - break; + TRACE(("|%p|%p|JUMP %d\n", pattern, + ptr, pattern[0])); + pattern += pattern[0]; + DISPATCH; - case SRE_OP_BRANCH: + TARGET(SRE_OP_BRANCH): /* alternation */ /* <0=skip> code ... */ - TRACE(("|%p|%p|BRANCH\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|BRANCH\n", pattern, ptr)); LASTMARK_SAVE(); if (state->repeat) MARK_PUSH(ctx->lastmark); - for (; ctx->pattern[0]; ctx->pattern += ctx->pattern[0]) { - if (ctx->pattern[1] == SRE_OP_LITERAL && - (ctx->ptr >= end || - (SRE_CODE) *ctx->ptr != ctx->pattern[2])) + for (; pattern[0]; pattern += pattern[0]) { + if (pattern[1] == SRE_OP_LITERAL && + (ptr >= end || + (SRE_CODE) *ptr != pattern[2])) continue; - if (ctx->pattern[1] == SRE_OP_IN && - (ctx->ptr >= end || - !SRE(charset)(state, ctx->pattern + 3, - (SRE_CODE) *ctx->ptr))) + if (pattern[1] == SRE_OP_IN && + (ptr >= end || + !SRE(charset)(state, pattern + 3, + (SRE_CODE) *ptr))) continue; - state->ptr = ctx->ptr; - DO_JUMP(JUMP_BRANCH, jump_branch, ctx->pattern+1); + state->ptr = ptr; + DO_JUMP(JUMP_BRANCH, jump_branch, pattern+1); if (ret) { if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); @@ -798,7 +838,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP_DISCARD(ctx->lastmark); RETURN_FAILURE; - case SRE_OP_REPEAT_ONE: + TARGET(SRE_OP_REPEAT_ONE): /* match repeated sequence (maximizing regexp) */ /* this operator only works if the repeated item is @@ -808,34 +848,34 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) /* <1=min> <2=max> item tail */ - TRACE(("|%p|%p|REPEAT_ONE %d %d\n", ctx->pattern, ctx->ptr, - ctx->pattern[1], ctx->pattern[2])); + TRACE(("|%p|%p|REPEAT_ONE %d %d\n", pattern, ptr, + pattern[1], pattern[2])); - if ((Py_ssize_t) ctx->pattern[1] > end - ctx->ptr) + if ((Py_ssize_t) pattern[1] > end - ptr) RETURN_FAILURE; /* cannot match */ - state->ptr = ctx->ptr; + state->ptr = ptr; - ret = SRE(count)(state, ctx->pattern+3, ctx->pattern[2]); + ret = SRE(count)(state, pattern+3, pattern[2]); RETURN_ON_ERROR(ret); DATA_LOOKUP_AT(SRE(match_context), ctx, ctx_pos); ctx->count = ret; - ctx->ptr += ctx->count; + ptr += ctx->count; /* when we arrive here, count contains the number of - matches, and ctx->ptr points to the tail of the target + matches, and ptr points to the tail of the target string. check if the rest of the pattern matches, and backtrack if not. */ - if (ctx->count < (Py_ssize_t) ctx->pattern[1]) + if (ctx->count < (Py_ssize_t) pattern[1]) RETURN_FAILURE; - if (ctx->pattern[ctx->pattern[0]] == SRE_OP_SUCCESS && - ctx->ptr == state->end && - !(ctx->toplevel && state->must_advance && ctx->ptr == state->start)) + if (pattern[pattern[0]] == SRE_OP_SUCCESS && + ptr == state->end && + !(ctx->toplevel && state->must_advance && ptr == state->start)) { /* tail is empty. we're finished */ - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_SUCCESS; } @@ -843,21 +883,21 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (state->repeat) MARK_PUSH(ctx->lastmark); - if (ctx->pattern[ctx->pattern[0]] == SRE_OP_LITERAL) { + if (pattern[pattern[0]] == SRE_OP_LITERAL) { /* tail starts with a literal. skip positions where the rest of the pattern cannot possibly match */ - ctx->u.chr = ctx->pattern[ctx->pattern[0]+1]; + ctx->u.chr = pattern[pattern[0]+1]; for (;;) { - while (ctx->count >= (Py_ssize_t) ctx->pattern[1] && - (ctx->ptr >= end || *ctx->ptr != ctx->u.chr)) { - ctx->ptr--; + while (ctx->count >= (Py_ssize_t) pattern[1] && + (ptr >= end || *ptr != ctx->u.chr)) { + ptr--; ctx->count--; } - if (ctx->count < (Py_ssize_t) ctx->pattern[1]) + if (ctx->count < (Py_ssize_t) pattern[1]) break; - state->ptr = ctx->ptr; + state->ptr = ptr; DO_JUMP(JUMP_REPEAT_ONE_1, jump_repeat_one_1, - ctx->pattern+ctx->pattern[0]); + pattern+pattern[0]); if (ret) { if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); @@ -868,17 +908,17 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP_KEEP(ctx->lastmark); LASTMARK_RESTORE(); - ctx->ptr--; + ptr--; ctx->count--; } if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); } else { /* general case */ - while (ctx->count >= (Py_ssize_t) ctx->pattern[1]) { - state->ptr = ctx->ptr; + while (ctx->count >= (Py_ssize_t) pattern[1]) { + state->ptr = ptr; DO_JUMP(JUMP_REPEAT_ONE_2, jump_repeat_one_2, - ctx->pattern+ctx->pattern[0]); + pattern+pattern[0]); if (ret) { if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); @@ -889,7 +929,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP_KEEP(ctx->lastmark); LASTMARK_RESTORE(); - ctx->ptr--; + ptr--; ctx->count--; } if (state->repeat) @@ -897,7 +937,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) } RETURN_FAILURE; - case SRE_OP_MIN_REPEAT_ONE: + TARGET(SRE_OP_MIN_REPEAT_ONE): /* match repeated sequence (minimizing regexp) */ /* this operator only works if the repeated item is @@ -907,36 +947,36 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) /* <1=min> <2=max> item tail */ - TRACE(("|%p|%p|MIN_REPEAT_ONE %d %d\n", ctx->pattern, ctx->ptr, - ctx->pattern[1], ctx->pattern[2])); + TRACE(("|%p|%p|MIN_REPEAT_ONE %d %d\n", pattern, ptr, + pattern[1], pattern[2])); - if ((Py_ssize_t) ctx->pattern[1] > end - ctx->ptr) + if ((Py_ssize_t) pattern[1] > end - ptr) RETURN_FAILURE; /* cannot match */ - state->ptr = ctx->ptr; + state->ptr = ptr; - if (ctx->pattern[1] == 0) + if (pattern[1] == 0) ctx->count = 0; else { /* count using pattern min as the maximum */ - ret = SRE(count)(state, ctx->pattern+3, ctx->pattern[1]); + ret = SRE(count)(state, pattern+3, pattern[1]); RETURN_ON_ERROR(ret); DATA_LOOKUP_AT(SRE(match_context), ctx, ctx_pos); - if (ret < (Py_ssize_t) ctx->pattern[1]) + if (ret < (Py_ssize_t) pattern[1]) /* didn't match minimum number of times */ RETURN_FAILURE; /* advance past minimum matches of repeat */ ctx->count = ret; - ctx->ptr += ctx->count; + ptr += ctx->count; } - if (ctx->pattern[ctx->pattern[0]] == SRE_OP_SUCCESS && + if (pattern[pattern[0]] == SRE_OP_SUCCESS && !(ctx->toplevel && - ((state->match_all && ctx->ptr != state->end) || - (state->must_advance && ctx->ptr == state->start)))) + ((state->match_all && ptr != state->end) || + (state->must_advance && ptr == state->start)))) { /* tail is empty. we're finished */ - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_SUCCESS; } else { @@ -945,11 +985,11 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (state->repeat) MARK_PUSH(ctx->lastmark); - while ((Py_ssize_t)ctx->pattern[2] == SRE_MAXREPEAT - || ctx->count <= (Py_ssize_t)ctx->pattern[2]) { - state->ptr = ctx->ptr; + while ((Py_ssize_t)pattern[2] == SRE_MAXREPEAT + || ctx->count <= (Py_ssize_t)pattern[2]) { + state->ptr = ptr; DO_JUMP(JUMP_MIN_REPEAT_ONE,jump_min_repeat_one, - ctx->pattern+ctx->pattern[0]); + pattern+pattern[0]); if (ret) { if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); @@ -960,14 +1000,14 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP_KEEP(ctx->lastmark); LASTMARK_RESTORE(); - state->ptr = ctx->ptr; - ret = SRE(count)(state, ctx->pattern+3, 1); + state->ptr = ptr; + ret = SRE(count)(state, pattern+3, 1); RETURN_ON_ERROR(ret); DATA_LOOKUP_AT(SRE(match_context), ctx, ctx_pos); if (ret == 0) break; assert(ret == 1); - ctx->ptr++; + ptr++; ctx->count++; } if (state->repeat) @@ -975,7 +1015,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) } RETURN_FAILURE; - case SRE_OP_POSSESSIVE_REPEAT_ONE: + TARGET(SRE_OP_POSSESSIVE_REPEAT_ONE): /* match repeated sequence (maximizing regexp) without backtracking */ @@ -987,67 +1027,67 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) /* <1=min> <2=max> item tail */ - TRACE(("|%p|%p|POSSESSIVE_REPEAT_ONE %d %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[1], ctx->pattern[2])); + TRACE(("|%p|%p|POSSESSIVE_REPEAT_ONE %d %d\n", pattern, + ptr, pattern[1], pattern[2])); - if (ctx->ptr + ctx->pattern[1] > end) { + if (ptr + pattern[1] > end) { RETURN_FAILURE; /* cannot match */ } - state->ptr = ctx->ptr; + state->ptr = ptr; - ret = SRE(count)(state, ctx->pattern + 3, ctx->pattern[2]); + ret = SRE(count)(state, pattern + 3, pattern[2]); RETURN_ON_ERROR(ret); DATA_LOOKUP_AT(SRE(match_context), ctx, ctx_pos); ctx->count = ret; - ctx->ptr += ctx->count; + ptr += ctx->count; /* when we arrive here, count contains the number of - matches, and ctx->ptr points to the tail of the target + matches, and ptr points to the tail of the target string. check if the rest of the pattern matches, and fail if not. */ /* Test for not enough repetitions in match */ - if (ctx->count < (Py_ssize_t) ctx->pattern[1]) { + if (ctx->count < (Py_ssize_t) pattern[1]) { RETURN_FAILURE; } /* Update the pattern to point to the next op code */ - ctx->pattern += ctx->pattern[0]; + pattern += pattern[0]; /* Let the tail be evaluated separately and consider this match successful. */ - if (*ctx->pattern == SRE_OP_SUCCESS && - ctx->ptr == state->end && - !(ctx->toplevel && state->must_advance && ctx->ptr == state->start)) + if (*pattern == SRE_OP_SUCCESS && + ptr == state->end && + !(ctx->toplevel && state->must_advance && ptr == state->start)) { /* tail is empty. we're finished */ - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_SUCCESS; } /* Attempt to match the rest of the string */ - break; + DISPATCH; - case SRE_OP_REPEAT: + TARGET(SRE_OP_REPEAT): /* create repeat context. all the hard work is done by the UNTIL operator (MAX_UNTIL, MIN_UNTIL) */ /* <1=min> <2=max> <3=repeat_index> item tail */ - TRACE(("|%p|%p|REPEAT %d %d %d\n", ctx->pattern, ctx->ptr, - ctx->pattern[1], ctx->pattern[2], ctx->pattern[3])); + TRACE(("|%p|%p|REPEAT %d %d %d\n", pattern, ptr, + pattern[1], pattern[2], pattern[3])); /* install repeat context */ - ctx->u.rep = &state->repeats_array[ctx->pattern[3]]; + ctx->u.rep = &state->repeats_array[pattern[3]]; ctx->u.rep->count = -1; - ctx->u.rep->pattern = ctx->pattern; + ctx->u.rep->pattern = pattern; ctx->u.rep->prev = state->repeat; ctx->u.rep->last_ptr = NULL; state->repeat = ctx->u.rep; - state->ptr = ctx->ptr; - DO_JUMP(JUMP_REPEAT, jump_repeat, ctx->pattern+ctx->pattern[0]); + state->ptr = ptr; + DO_JUMP(JUMP_REPEAT, jump_repeat, pattern+pattern[0]); state->repeat = ctx->u.rep->prev; if (ret) { @@ -1056,7 +1096,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) } RETURN_FAILURE; - case SRE_OP_MAX_UNTIL: + TARGET(SRE_OP_MAX_UNTIL): /* maximizing repeat */ /* <1=min> <2=max> <3=repeat_index> item tail */ @@ -1068,12 +1108,12 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!ctx->u.rep) RETURN_ERROR(SRE_ERROR_STATE); - state->ptr = ctx->ptr; + state->ptr = ptr; ctx->count = ctx->u.rep->count+1; - TRACE(("|%p|%p|MAX_UNTIL %zd\n", ctx->pattern, - ctx->ptr, ctx->count)); + TRACE(("|%p|%p|MAX_UNTIL %zd\n", pattern, + ptr, ctx->count)); if (ctx->count < (Py_ssize_t) ctx->u.rep->pattern[1]) { /* not enough matches */ @@ -1085,7 +1125,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) RETURN_SUCCESS; } ctx->u.rep->count = ctx->count-1; - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; } @@ -1111,20 +1151,20 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP(ctx->lastmark); LASTMARK_RESTORE(); ctx->u.rep->count = ctx->count-1; - state->ptr = ctx->ptr; + state->ptr = ptr; } /* cannot match more repeated items here. make sure the tail matches */ state->repeat = ctx->u.rep->prev; - DO_JUMP(JUMP_MAX_UNTIL_3, jump_max_until_3, ctx->pattern); + DO_JUMP(JUMP_MAX_UNTIL_3, jump_max_until_3, pattern); state->repeat = ctx->u.rep; // restore repeat before return RETURN_ON_SUCCESS(ret); - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; - case SRE_OP_MIN_UNTIL: + TARGET(SRE_OP_MIN_UNTIL): /* minimizing repeat */ /* <1=min> <2=max> <3=repeat_index> item tail */ @@ -1133,12 +1173,12 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!ctx->u.rep) RETURN_ERROR(SRE_ERROR_STATE); - state->ptr = ctx->ptr; + state->ptr = ptr; ctx->count = ctx->u.rep->count+1; - TRACE(("|%p|%p|MIN_UNTIL %zd %p\n", ctx->pattern, - ctx->ptr, ctx->count, ctx->u.rep->pattern)); + TRACE(("|%p|%p|MIN_UNTIL %zd %p\n", pattern, + ptr, ctx->count, ctx->u.rep->pattern)); if (ctx->count < (Py_ssize_t) ctx->u.rep->pattern[1]) { /* not enough matches */ @@ -1150,7 +1190,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) RETURN_SUCCESS; } ctx->u.rep->count = ctx->count-1; - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; } @@ -1161,7 +1201,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (state->repeat) MARK_PUSH(ctx->lastmark); - DO_JUMP(JUMP_MIN_UNTIL_2, jump_min_until_2, ctx->pattern); + DO_JUMP(JUMP_MIN_UNTIL_2, jump_min_until_2, pattern); SRE_REPEAT *repeat_of_tail = state->repeat; state->repeat = ctx->u.rep; // restore repeat before return @@ -1175,7 +1215,7 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP(ctx->lastmark); LASTMARK_RESTORE(); - state->ptr = ctx->ptr; + state->ptr = ptr; if ((ctx->count >= (Py_ssize_t) ctx->u.rep->pattern[2] && ctx->u.rep->pattern[2] != SRE_MAXREPEAT) || @@ -1194,34 +1234,34 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) RETURN_SUCCESS; } ctx->u.rep->count = ctx->count-1; - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; - case SRE_OP_POSSESSIVE_REPEAT: + TARGET(SRE_OP_POSSESSIVE_REPEAT): /* create possessive repeat contexts. */ /* <1=min> <2=max> pattern tail */ - TRACE(("|%p|%p|POSSESSIVE_REPEAT %d %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[1], ctx->pattern[2])); + TRACE(("|%p|%p|POSSESSIVE_REPEAT %d %d\n", pattern, + ptr, pattern[1], pattern[2])); /* Set the global Input pointer to this context's Input pointer */ - state->ptr = ctx->ptr; + state->ptr = ptr; /* Initialize Count to 0 */ ctx->count = 0; /* Check for minimum required matches. */ - while (ctx->count < (Py_ssize_t)ctx->pattern[1]) { + while (ctx->count < (Py_ssize_t)pattern[1]) { /* not enough matches */ DO_JUMP(JUMP_POSS_REPEAT_1, jump_poss_repeat_1, - &ctx->pattern[3]); + &pattern[3]); if (ret) { RETURN_ON_ERROR(ret); ctx->count++; } else { - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; } } @@ -1229,13 +1269,13 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) /* Clear the context's Input stream pointer so that it doesn't match the global state so that the while loop can be entered. */ - ctx->ptr = NULL; + ptr = NULL; /* Keep trying to parse the sub-pattern until the end is reached, creating a new context each time. */ - while ((ctx->count < (Py_ssize_t)ctx->pattern[2] || - (Py_ssize_t)ctx->pattern[2] == SRE_MAXREPEAT) && - state->ptr != ctx->ptr) { + while ((ctx->count < (Py_ssize_t)pattern[2] || + (Py_ssize_t)pattern[2] == SRE_MAXREPEAT) && + state->ptr != ptr) { /* Save the Capture Group Marker state into the current Context and back up the current highest number Capture Group marker. */ @@ -1257,12 +1297,12 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) maximum number of matches are counted, and because of this, we could immediately stop at that point and consider this match successful. */ - ctx->ptr = state->ptr; + ptr = state->ptr; /* We have not reached the maximin matches, so try to match once more. */ DO_JUMP(JUMP_POSS_REPEAT_2, jump_poss_repeat_2, - &ctx->pattern[3]); + &pattern[3]); /* Check to see if the last attempted match succeeded. */ @@ -1293,47 +1333,47 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) /* Evaluate Tail */ /* Jump to end of pattern indicated by skip, and then skip the SUCCESS op code that follows it. */ - ctx->pattern += ctx->pattern[0] + 1; - ctx->ptr = state->ptr; - break; + pattern += pattern[0] + 1; + ptr = state->ptr; + DISPATCH; - case SRE_OP_ATOMIC_GROUP: + TARGET(SRE_OP_ATOMIC_GROUP): /* Atomic Group Sub Pattern */ /* pattern tail */ - TRACE(("|%p|%p|ATOMIC_GROUP\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|ATOMIC_GROUP\n", pattern, ptr)); /* Set the global Input pointer to this context's Input pointer */ - state->ptr = ctx->ptr; + state->ptr = ptr; /* Evaluate the Atomic Group in a new context, terminating when the end of the group, represented by a SUCCESS op code, is reached. */ /* Group Pattern begins at an offset of 1 code. */ DO_JUMP(JUMP_ATOMIC_GROUP, jump_atomic_group, - &ctx->pattern[1]); + &pattern[1]); /* Test Exit Condition */ RETURN_ON_ERROR(ret); if (ret == 0) { /* Atomic Group failed to Match. */ - state->ptr = ctx->ptr; + state->ptr = ptr; RETURN_FAILURE; } /* Evaluate Tail */ /* Jump to end of pattern indicated by skip, and then skip the SUCCESS op code that follows it. */ - ctx->pattern += ctx->pattern[0]; - ctx->ptr = state->ptr; - break; + pattern += pattern[0]; + ptr = state->ptr; + DISPATCH; - case SRE_OP_GROUPREF: + TARGET(SRE_OP_GROUPREF): /* match backreference */ - TRACE(("|%p|%p|GROUPREF %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - i = ctx->pattern[0]; + TRACE(("|%p|%p|GROUPREF %d\n", pattern, + ptr, pattern[0])); + i = pattern[0]; { Py_ssize_t groupref = i+i; if (groupref >= state->lastmark) { @@ -1344,21 +1384,21 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!p || !e || e < p) RETURN_FAILURE; while (p < e) { - if (ctx->ptr >= end || *ctx->ptr != *p) + if (ptr >= end || *ptr != *p) RETURN_FAILURE; p++; - ctx->ptr++; + ptr++; } } } - ctx->pattern++; - break; + pattern++; + DISPATCH; - case SRE_OP_GROUPREF_IGNORE: + TARGET(SRE_OP_GROUPREF_IGNORE): /* match backreference */ - TRACE(("|%p|%p|GROUPREF_IGNORE %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - i = ctx->pattern[0]; + TRACE(("|%p|%p|GROUPREF_IGNORE %d\n", pattern, + ptr, pattern[0])); + i = pattern[0]; { Py_ssize_t groupref = i+i; if (groupref >= state->lastmark) { @@ -1369,22 +1409,22 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!p || !e || e < p) RETURN_FAILURE; while (p < e) { - if (ctx->ptr >= end || - sre_lower_ascii(*ctx->ptr) != sre_lower_ascii(*p)) + if (ptr >= end || + sre_lower_ascii(*ptr) != sre_lower_ascii(*p)) RETURN_FAILURE; p++; - ctx->ptr++; + ptr++; } } } - ctx->pattern++; - break; + pattern++; + DISPATCH; - case SRE_OP_GROUPREF_UNI_IGNORE: + TARGET(SRE_OP_GROUPREF_UNI_IGNORE): /* match backreference */ - TRACE(("|%p|%p|GROUPREF_UNI_IGNORE %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - i = ctx->pattern[0]; + TRACE(("|%p|%p|GROUPREF_UNI_IGNORE %d\n", pattern, + ptr, pattern[0])); + i = pattern[0]; { Py_ssize_t groupref = i+i; if (groupref >= state->lastmark) { @@ -1395,22 +1435,22 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!p || !e || e < p) RETURN_FAILURE; while (p < e) { - if (ctx->ptr >= end || - sre_lower_unicode(*ctx->ptr) != sre_lower_unicode(*p)) + if (ptr >= end || + sre_lower_unicode(*ptr) != sre_lower_unicode(*p)) RETURN_FAILURE; p++; - ctx->ptr++; + ptr++; } } } - ctx->pattern++; - break; + pattern++; + DISPATCH; - case SRE_OP_GROUPREF_LOC_IGNORE: + TARGET(SRE_OP_GROUPREF_LOC_IGNORE): /* match backreference */ - TRACE(("|%p|%p|GROUPREF_LOC_IGNORE %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); - i = ctx->pattern[0]; + TRACE(("|%p|%p|GROUPREF_LOC_IGNORE %d\n", pattern, + ptr, pattern[0])); + i = pattern[0]; { Py_ssize_t groupref = i+i; if (groupref >= state->lastmark) { @@ -1421,64 +1461,64 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) if (!p || !e || e < p) RETURN_FAILURE; while (p < e) { - if (ctx->ptr >= end || - sre_lower_locale(*ctx->ptr) != sre_lower_locale(*p)) + if (ptr >= end || + sre_lower_locale(*ptr) != sre_lower_locale(*p)) RETURN_FAILURE; p++; - ctx->ptr++; + ptr++; } } } - ctx->pattern++; - break; + pattern++; + DISPATCH; - case SRE_OP_GROUPREF_EXISTS: - TRACE(("|%p|%p|GROUPREF_EXISTS %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[0])); + TARGET(SRE_OP_GROUPREF_EXISTS): + TRACE(("|%p|%p|GROUPREF_EXISTS %d\n", pattern, + ptr, pattern[0])); /* codeyes codeno ... */ - i = ctx->pattern[0]; + i = pattern[0]; { Py_ssize_t groupref = i+i; if (groupref >= state->lastmark) { - ctx->pattern += ctx->pattern[1]; - break; + pattern += pattern[1]; + DISPATCH; } else { SRE_CHAR* p = (SRE_CHAR*) state->mark[groupref]; SRE_CHAR* e = (SRE_CHAR*) state->mark[groupref+1]; if (!p || !e || e < p) { - ctx->pattern += ctx->pattern[1]; - break; + pattern += pattern[1]; + DISPATCH; } } } - ctx->pattern += 2; - break; + pattern += 2; + DISPATCH; - case SRE_OP_ASSERT: + TARGET(SRE_OP_ASSERT): /* assert subpattern */ /* */ - TRACE(("|%p|%p|ASSERT %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[1])); - if (ctx->ptr - (SRE_CHAR *)state->beginning < (Py_ssize_t)ctx->pattern[1]) + TRACE(("|%p|%p|ASSERT %d\n", pattern, + ptr, pattern[1])); + if (ptr - (SRE_CHAR *)state->beginning < (Py_ssize_t)pattern[1]) RETURN_FAILURE; - state->ptr = ctx->ptr - ctx->pattern[1]; - DO_JUMP0(JUMP_ASSERT, jump_assert, ctx->pattern+2); + state->ptr = ptr - pattern[1]; + DO_JUMP0(JUMP_ASSERT, jump_assert, pattern+2); RETURN_ON_FAILURE(ret); - ctx->pattern += ctx->pattern[0]; - break; + pattern += pattern[0]; + DISPATCH; - case SRE_OP_ASSERT_NOT: + TARGET(SRE_OP_ASSERT_NOT): /* assert not subpattern */ /* */ - TRACE(("|%p|%p|ASSERT_NOT %d\n", ctx->pattern, - ctx->ptr, ctx->pattern[1])); - if (ctx->ptr - (SRE_CHAR *)state->beginning >= (Py_ssize_t)ctx->pattern[1]) { - state->ptr = ctx->ptr - ctx->pattern[1]; + TRACE(("|%p|%p|ASSERT_NOT %d\n", pattern, + ptr, pattern[1])); + if (ptr - (SRE_CHAR *)state->beginning >= (Py_ssize_t)pattern[1]) { + state->ptr = ptr - pattern[1]; LASTMARK_SAVE(); if (state->repeat) MARK_PUSH(ctx->lastmark); - DO_JUMP0(JUMP_ASSERT_NOT, jump_assert_not, ctx->pattern+2); + DO_JUMP0(JUMP_ASSERT_NOT, jump_assert_not, pattern+2); if (ret) { if (state->repeat) MARK_POP_DISCARD(ctx->lastmark); @@ -1489,19 +1529,29 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) MARK_POP(ctx->lastmark); LASTMARK_RESTORE(); } - ctx->pattern += ctx->pattern[0]; - break; + pattern += pattern[0]; + DISPATCH; - case SRE_OP_FAILURE: + TARGET(SRE_OP_FAILURE): /* immediate failure */ - TRACE(("|%p|%p|FAILURE\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|FAILURE\n", pattern, ptr)); RETURN_FAILURE; +#if !USE_COMPUTED_GOTOS default: - TRACE(("|%p|%p|UNKNOWN %d\n", ctx->pattern, ctx->ptr, - ctx->pattern[-1])); +#endif + // Also any unused opcodes: + TARGET(SRE_OP_RANGE_UNI_IGNORE): + TARGET(SRE_OP_SUBPATTERN): + TARGET(SRE_OP_RANGE): + TARGET(SRE_OP_NEGATE): + TARGET(SRE_OP_BIGCHARSET): + TARGET(SRE_OP_CHARSET): + TARGET(SRE_OP_CALL): + TRACE(("|%p|%p|UNKNOWN %d\n", pattern, ptr, + pattern[-1])); RETURN_ERROR(SRE_ERROR_ILLEGAL); - } + } exit: @@ -1514,56 +1564,56 @@ SRE(match)(SRE_STATE* state, const SRE_CODE* pattern, int toplevel) switch (jump) { case JUMP_MAX_UNTIL_2: - TRACE(("|%p|%p|JUMP_MAX_UNTIL_2\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MAX_UNTIL_2\n", pattern, ptr)); goto jump_max_until_2; case JUMP_MAX_UNTIL_3: - TRACE(("|%p|%p|JUMP_MAX_UNTIL_3\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MAX_UNTIL_3\n", pattern, ptr)); goto jump_max_until_3; case JUMP_MIN_UNTIL_2: - TRACE(("|%p|%p|JUMP_MIN_UNTIL_2\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MIN_UNTIL_2\n", pattern, ptr)); goto jump_min_until_2; case JUMP_MIN_UNTIL_3: - TRACE(("|%p|%p|JUMP_MIN_UNTIL_3\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MIN_UNTIL_3\n", pattern, ptr)); goto jump_min_until_3; case JUMP_BRANCH: - TRACE(("|%p|%p|JUMP_BRANCH\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_BRANCH\n", pattern, ptr)); goto jump_branch; case JUMP_MAX_UNTIL_1: - TRACE(("|%p|%p|JUMP_MAX_UNTIL_1\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MAX_UNTIL_1\n", pattern, ptr)); goto jump_max_until_1; case JUMP_MIN_UNTIL_1: - TRACE(("|%p|%p|JUMP_MIN_UNTIL_1\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MIN_UNTIL_1\n", pattern, ptr)); goto jump_min_until_1; case JUMP_POSS_REPEAT_1: - TRACE(("|%p|%p|JUMP_POSS_REPEAT_1\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_POSS_REPEAT_1\n", pattern, ptr)); goto jump_poss_repeat_1; case JUMP_POSS_REPEAT_2: - TRACE(("|%p|%p|JUMP_POSS_REPEAT_2\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_POSS_REPEAT_2\n", pattern, ptr)); goto jump_poss_repeat_2; case JUMP_REPEAT: - TRACE(("|%p|%p|JUMP_REPEAT\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_REPEAT\n", pattern, ptr)); goto jump_repeat; case JUMP_REPEAT_ONE_1: - TRACE(("|%p|%p|JUMP_REPEAT_ONE_1\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_REPEAT_ONE_1\n", pattern, ptr)); goto jump_repeat_one_1; case JUMP_REPEAT_ONE_2: - TRACE(("|%p|%p|JUMP_REPEAT_ONE_2\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_REPEAT_ONE_2\n", pattern, ptr)); goto jump_repeat_one_2; case JUMP_MIN_REPEAT_ONE: - TRACE(("|%p|%p|JUMP_MIN_REPEAT_ONE\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_MIN_REPEAT_ONE\n", pattern, ptr)); goto jump_min_repeat_one; case JUMP_ATOMIC_GROUP: - TRACE(("|%p|%p|JUMP_ATOMIC_GROUP\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_ATOMIC_GROUP\n", pattern, ptr)); goto jump_atomic_group; case JUMP_ASSERT: - TRACE(("|%p|%p|JUMP_ASSERT\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_ASSERT\n", pattern, ptr)); goto jump_assert; case JUMP_ASSERT_NOT: - TRACE(("|%p|%p|JUMP_ASSERT_NOT\n", ctx->pattern, ctx->ptr)); + TRACE(("|%p|%p|JUMP_ASSERT_NOT\n", pattern, ptr)); goto jump_assert_not; case JUMP_NONE: - TRACE(("|%p|%p|RETURN %zd\n", ctx->pattern, - ctx->ptr, ret)); + TRACE(("|%p|%p|RETURN %zd\n", pattern, + ptr, ret)); break; } diff --git a/Modules/_sre/sre_targets.h b/Modules/_sre/sre_targets.h new file mode 100644 index 00000000000..389e7d77e94 --- /dev/null +++ b/Modules/_sre/sre_targets.h @@ -0,0 +1,59 @@ +/* + * Secret Labs' Regular Expression Engine + * + * regular expression matching engine + * + * Auto-generated by Tools/scripts/generate_sre_constants.py from + * Lib/re/_constants.py. + * + * Copyright (c) 1997-2001 by Secret Labs AB. All rights reserved. + * + * See the sre.c file for information on usage and redistribution. + */ + +static void *sre_targets[44] = { + &&TARGET_SRE_OP_FAILURE, + &&TARGET_SRE_OP_SUCCESS, + &&TARGET_SRE_OP_ANY, + &&TARGET_SRE_OP_ANY_ALL, + &&TARGET_SRE_OP_ASSERT, + &&TARGET_SRE_OP_ASSERT_NOT, + &&TARGET_SRE_OP_AT, + &&TARGET_SRE_OP_BRANCH, + &&TARGET_SRE_OP_CALL, + &&TARGET_SRE_OP_CATEGORY, + &&TARGET_SRE_OP_CHARSET, + &&TARGET_SRE_OP_BIGCHARSET, + &&TARGET_SRE_OP_GROUPREF, + &&TARGET_SRE_OP_GROUPREF_EXISTS, + &&TARGET_SRE_OP_IN, + &&TARGET_SRE_OP_INFO, + &&TARGET_SRE_OP_JUMP, + &&TARGET_SRE_OP_LITERAL, + &&TARGET_SRE_OP_MARK, + &&TARGET_SRE_OP_MAX_UNTIL, + &&TARGET_SRE_OP_MIN_UNTIL, + &&TARGET_SRE_OP_NOT_LITERAL, + &&TARGET_SRE_OP_NEGATE, + &&TARGET_SRE_OP_RANGE, + &&TARGET_SRE_OP_REPEAT, + &&TARGET_SRE_OP_REPEAT_ONE, + &&TARGET_SRE_OP_SUBPATTERN, + &&TARGET_SRE_OP_MIN_REPEAT_ONE, + &&TARGET_SRE_OP_ATOMIC_GROUP, + &&TARGET_SRE_OP_POSSESSIVE_REPEAT, + &&TARGET_SRE_OP_POSSESSIVE_REPEAT_ONE, + &&TARGET_SRE_OP_GROUPREF_IGNORE, + &&TARGET_SRE_OP_IN_IGNORE, + &&TARGET_SRE_OP_LITERAL_IGNORE, + &&TARGET_SRE_OP_NOT_LITERAL_IGNORE, + &&TARGET_SRE_OP_GROUPREF_LOC_IGNORE, + &&TARGET_SRE_OP_IN_LOC_IGNORE, + &&TARGET_SRE_OP_LITERAL_LOC_IGNORE, + &&TARGET_SRE_OP_NOT_LITERAL_LOC_IGNORE, + &&TARGET_SRE_OP_GROUPREF_UNI_IGNORE, + &&TARGET_SRE_OP_IN_UNI_IGNORE, + &&TARGET_SRE_OP_LITERAL_UNI_IGNORE, + &&TARGET_SRE_OP_NOT_LITERAL_UNI_IGNORE, + &&TARGET_SRE_OP_RANGE_UNI_IGNORE, +}; diff --git a/Tools/scripts/generate_sre_constants.py b/Tools/scripts/generate_sre_constants.py index b8f0df9cc23..72715076d29 100755 --- a/Tools/scripts/generate_sre_constants.py +++ b/Tools/scripts/generate_sre_constants.py @@ -29,7 +29,11 @@ def update_file(file, content): """ -def main(infile='Lib/re/_constants.py', outfile='Modules/_sre/sre_constants.h'): +def main( + infile="Lib/re/_constants.py", + outfile_constants="Modules/_sre/sre_constants.h", + outfile_targets="Modules/_sre/sre_targets.h", +): ns = {} with open(infile) as fp: code = fp.read() @@ -46,6 +50,11 @@ def dump2(d, prefix): for value, name in sorted(items): yield "#define %s %d\n" % (name, value) + def dump_gotos(d, prefix): + for i, item in enumerate(sorted(d)): + assert i == item + yield f" &&{prefix}_{item},\n" + content = [sre_constants_header] content.append("#define SRE_MAGIC %d\n" % ns["MAGIC"]) content.extend(dump(ns["OPCODES"], "SRE_OP")) @@ -54,7 +63,14 @@ def dump2(d, prefix): content.extend(dump2(ns, "SRE_FLAG_")) content.extend(dump2(ns, "SRE_INFO_")) - update_file(outfile, ''.join(content)) + update_file(outfile_constants, ''.join(content)) + + content = [sre_constants_header] + content.append(f"static void *sre_targets[{len(ns['OPCODES'])}] = {{\n") + content.extend(dump_gotos(ns["OPCODES"], "TARGET_SRE_OP")) + content.append("};\n") + + update_file(outfile_targets, ''.join(content)) if __name__ == '__main__':