diff options
author | Sverker Eriksson <[email protected]> | 2019-02-06 19:10:26 +0100 |
---|---|---|
committer | Sverker Eriksson <[email protected]> | 2019-02-06 19:10:26 +0100 |
commit | 98cfd6016f8b40fc97e03b31177d14318349040f (patch) | |
tree | c0fcdd768071c36bfbcbf186d369d9ca14c47421 /erts/emulator/pcre/pcre_xclass.c | |
parent | e2ca71b6e7172b320b5b171359d53a161383fb19 (diff) | |
parent | 3825199794da28d79b21052a2e69e2335921d55e (diff) | |
download | otp-98cfd6016f8b40fc97e03b31177d14318349040f.tar.gz otp-98cfd6016f8b40fc97e03b31177d14318349040f.tar.bz2 otp-98cfd6016f8b40fc97e03b31177d14318349040f.zip |
Merge tag 'OTP-21.2' into sverker/map-from-ks-vs-bug
Diffstat (limited to 'erts/emulator/pcre/pcre_xclass.c')
-rw-r--r-- | erts/emulator/pcre/pcre_xclass.c | 94 |
1 files changed, 75 insertions, 19 deletions
diff --git a/erts/emulator/pcre/pcre_xclass.c b/erts/emulator/pcre/pcre_xclass.c index cf07451a10..67095fa18b 100644 --- a/erts/emulator/pcre/pcre_xclass.c +++ b/erts/emulator/pcre/pcre_xclass.c @@ -82,6 +82,11 @@ additional data. */ if (c < 256) { + if ((*data & XCL_HASPROP) == 0) + { + if ((*data & XCL_MAP) == 0) return negated; + return (((pcre_uint8 *)(data + 1))[c/8] & (1 << (c&7))) != 0; + } if ((*data & XCL_MAP) != 0 && (((pcre_uint8 *)(data + 1))[c/8] & (1 << (c&7))) != 0) return !negated; /* char found */ @@ -129,55 +134,62 @@ while ((t = *data++) != XCL_END) else /* XCL_PROP & XCL_NOTPROP */ { const ucd_record *prop = GET_UCD(c); + BOOL isprop = t == XCL_PROP; switch(*data) { case PT_ANY: - if (t == XCL_PROP) return !negated; + if (isprop) return !negated; break; case PT_LAMP: if ((prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || - prop->chartype == ucp_Lt) == (t == XCL_PROP)) return !negated; + prop->chartype == ucp_Lt) == isprop) return !negated; break; case PT_GC: - if ((data[1] == PRIV(ucp_gentype)[prop->chartype]) == (t == XCL_PROP)) + if ((data[1] == PRIV(ucp_gentype)[prop->chartype]) == isprop) return !negated; break; case PT_PC: - if ((data[1] == prop->chartype) == (t == XCL_PROP)) return !negated; + if ((data[1] == prop->chartype) == isprop) return !negated; break; case PT_SC: - if ((data[1] == prop->script) == (t == XCL_PROP)) return !negated; + if ((data[1] == prop->script) == isprop) return !negated; break; case PT_ALNUM: if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || - PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (t == XCL_PROP)) + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == isprop) return !negated; break; - case PT_SPACE: /* Perl space */ - if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || - c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR) - == (t == XCL_PROP)) - return !negated; - break; + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + case PT_SPACE: /* Perl space */ case PT_PXSPACE: /* POSIX space */ - if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || - c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || - c == CHAR_FF || c == CHAR_CR) == (t == XCL_PROP)) - return !negated; + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (isprop) return !negated; + break; + + default: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) == isprop) + return !negated; + break; + } break; case PT_WORD: if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || PRIV(ucp_gentype)[prop->chartype] == ucp_N || c == CHAR_UNDERSCORE) - == (t == XCL_PROP)) + == isprop) return !negated; break; @@ -185,16 +197,60 @@ while ((t = *data++) != XCL_END) if (c < 0xa0) { if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || - c == CHAR_GRAVE_ACCENT) == (t == XCL_PROP)) + c == CHAR_GRAVE_ACCENT) == isprop) return !negated; } else { - if ((c < 0xd800 || c > 0xdfff) == (t == XCL_PROP)) + if ((c < 0xd800 || c > 0xdfff) == isprop) return !negated; } break; + /* The following three properties can occur only in an XCLASS, as there + is no \p or \P coding for them. */ + + /* Graphic character. Implement this as not Z (space or separator) and + not C (other), except for Cf (format) with a few exceptions. This seems + to be what Perl does. The exceptional characters are: + + U+061C Arabic Letter Mark + U+180E Mongolian Vowel Separator + U+2066 - U+2069 Various "isolate"s + */ + + case PT_PXGRAPH: + if ((PRIV(ucp_gentype)[prop->chartype] != ucp_Z && + (PRIV(ucp_gentype)[prop->chartype] != ucp_C || + (prop->chartype == ucp_Cf && + c != 0x061c && c != 0x180e && (c < 0x2066 || c > 0x2069)) + )) == isprop) + return !negated; + break; + + /* Printable character: same as graphic, with the addition of Zs, i.e. + not Zl and not Zp, and U+180E. */ + + case PT_PXPRINT: + if ((prop->chartype != ucp_Zl && + prop->chartype != ucp_Zp && + (PRIV(ucp_gentype)[prop->chartype] != ucp_C || + (prop->chartype == ucp_Cf && + c != 0x061c && (c < 0x2066 || c > 0x2069)) + )) == isprop) + return !negated; + break; + + /* Punctuation: all Unicode punctuation, plus ASCII characters that + Unicode treats as symbols rather than punctuation, for Perl + compatibility (these are $+<=>^`|~). */ + + case PT_PXPUNCT: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_P || + (c < 128 && PRIV(ucp_gentype)[prop->chartype] == ucp_S)) == isprop) + return !negated; + break; + /* This should never occur, but compilers may mutter if there is no default. */ |