diff options
| author | Andrew Gierth <rhodiumtoad@postgresql.org> | 2018-09-12 19:31:06 +0100 |
|---|---|---|
| committer | Andrew Gierth <rhodiumtoad@postgresql.org> | 2018-09-12 19:31:06 +0100 |
| commit | b7f6bcbffcc0b41d783c0c9c61766428159969ff (patch) | |
| tree | 97f6adb12342296dd1d7dba55586141914da4bb1 /src/backend/utils/adt/regexp.c | |
| parent | ba37349cff781120f61b2778257f594f0d10253c (diff) | |
| download | postgresql-b7f6bcbffcc0b41d783c0c9c61766428159969ff.tar.gz postgresql-b7f6bcbffcc0b41d783c0c9c61766428159969ff.zip | |
Repair bug in regexp split performance improvements.
Commit c8ea87e4b introduced a temporary conversion buffer for
substrings extracted during regexp splits. Unfortunately the code that
sized it was failing to ignore the effects of ignored degenerate
regexp matches, so for regexp_split_* calls it could under-size the
buffer in such cases.
Fix, and add some regression test cases (though those will only catch
the bug if run in a multibyte encoding).
Backpatch to 9.3 as the faulty code was.
Thanks to the PostGIS project, Regina Obe and Paul Ramsey for the
report (via IRC) and assistance in analysis. Patch by me.
Diffstat (limited to 'src/backend/utils/adt/regexp.c')
| -rw-r--r-- | src/backend/utils/adt/regexp.c | 16 |
1 files changed, 10 insertions, 6 deletions
diff --git a/src/backend/utils/adt/regexp.c b/src/backend/utils/adt/regexp.c index d8b69212342..171fcc8a448 100644 --- a/src/backend/utils/adt/regexp.c +++ b/src/backend/utils/adt/regexp.c @@ -982,6 +982,7 @@ setup_regexp_matches(text *orig_str, text *pattern, pg_re_flags *re_flags, int array_len; int array_idx; int prev_match_end; + int prev_valid_match_end; int start_search; int maxlen = 0; /* largest fetch length in characters */ @@ -1024,6 +1025,7 @@ setup_regexp_matches(text *orig_str, text *pattern, pg_re_flags *re_flags, /* search for the pattern, perhaps repeatedly */ prev_match_end = 0; + prev_valid_match_end = 0; start_search = 0; while (RE_wchar_execute(cpattern, wide_str, wide_len, start_search, pmatch_len, pmatch)) @@ -1076,13 +1078,15 @@ setup_regexp_matches(text *orig_str, text *pattern, pg_re_flags *re_flags, matchctx->nmatches++; /* - * check length of unmatched portion between end of previous match - * and start of current one + * check length of unmatched portion between end of previous valid + * (nondegenerate, or degenerate but not ignored) match and start + * of current one */ if (fetching_unmatched && pmatch[0].rm_so >= 0 && - (pmatch[0].rm_so - prev_match_end) > maxlen) - maxlen = (pmatch[0].rm_so - prev_match_end); + (pmatch[0].rm_so - prev_valid_match_end) > maxlen) + maxlen = (pmatch[0].rm_so - prev_valid_match_end); + prev_valid_match_end = pmatch[0].rm_eo; } prev_match_end = pmatch[0].rm_eo; @@ -1108,8 +1112,8 @@ setup_regexp_matches(text *orig_str, text *pattern, pg_re_flags *re_flags, * input string */ if (fetching_unmatched && - (wide_len - prev_match_end) > maxlen) - maxlen = (wide_len - prev_match_end); + (wide_len - prev_valid_match_end) > maxlen) + maxlen = (wide_len - prev_valid_match_end); /* * Keep a note of the end position of the string for the benefit of |