diff --git a/ChangeLog b/ChangeLog index f0eab7b..ffb0880 100644 --- a/ChangeLog +++ b/ChangeLog @@ -92,6 +92,9 @@ pattern, the optimizing "must be present for a match" character check was not being flagged as caseless, causing some matches that should have succeeded to fail. +23. Fixed a unicode properrty matching issue in JIT. The character was not +fully read in caseless matching. + Version 10.39 29-October-2021 ----------------------------- diff --git a/src/pcre2_jit_compile.c b/src/pcre2_jit_compile.c index e7dd26c..94f6a58 100644 --- a/src/pcre2_jit_compile.c +++ b/src/pcre2_jit_compile.c @@ -7489,7 +7489,7 @@ while (*cc != XCL_END) { SLJIT_ASSERT(*cc == XCL_PROP || *cc == XCL_NOTPROP); cc++; - if (*cc == PT_CLIST) + if (*cc == PT_CLIST && *cc == XCL_PROP) { other_cases = PRIV(ucd_caseless_sets) + cc[1]; while (*other_cases != NOTACHAR) diff --git a/src/pcre2_jit_test.c b/src/pcre2_jit_test.c index 3b57ce2..8450f0b 100644 --- a/src/pcre2_jit_test.c +++ b/src/pcre2_jit_test.c @@ -415,6 +415,7 @@ static struct regression_test_case regression_test_cases[] = { { MUP, 0, 0, 0 | F_NOMATCH, "[^\\p{Hangul}\\p{Z}]", " " }, { MUP, 0, 0, 0, "[\\p{Lu}\\P{Latin}]+", "c\xEA\xA4\xAE,A,b" }, { MUP, 0, 0, 0, "[\\x{a92e}\\p{Lu}\\P{Latin}]+", "c\xEA\xA4\xAE,A,b" }, + { CMUP, 0, 0, 0, "[^S]\\B", "\xe2\x80\x8a" }, /* Possible empty brackets. */ { MU, A, 0, 0, "(?:|ab||bc|a)+d", "abcxabcabd" },