From: Ben Pfaff Date: Fri, 19 Nov 2010 05:30:24 +0000 (-0800) Subject: identifier: Add some functions for Unicode syntax. X-Git-Tag: v0.7.7~127 X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=325168fb62e51d5154e932abec91ba40dd081f3f;p=pspp-builds.git identifier: Add some functions for Unicode syntax. --- diff --git a/Smake b/Smake index 6e6b7023..014208b2 100644 --- a/Smake +++ b/Smake @@ -70,6 +70,8 @@ GNULIB_MODULES = \ trunc \ unilbrk/ulc-width-linebreaks \ unistd \ + unictype/property-id-continue \ + unictype/property-id-start \ unistr/u8-cpy \ unistr/u8-mbtouc \ unistr/u8-strlen \ diff --git a/src/data/identifier.c b/src/data/identifier.c index 7bd2261e..fa9e3208 100644 --- a/src/data/identifier.c +++ b/src/data/identifier.c @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 1997-9, 2000, 2005, 2009 Free Software Foundation, Inc. + Copyright (C) 1997-9, 2000, 2005, 2009, 2010 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -20,34 +20,83 @@ */ #include -#include "identifier.h" +#include "data/identifier.h" #include #include -#include +#include + +#include "libpspp/assertion.h" + +#include "gl/c-ctype.h" /* Recognizing identifiers. */ -/* Returns true if C may be the first character in an +static bool +is_ascii_id1 (unsigned char c) +{ + return c_isalpha (c) || c == '@' || c == '#' || c == '$'; +} + +static bool +is_ascii_idn (unsigned char c) +{ + return is_ascii_id1 (c) || isdigit (c) || c == '.' || c == '_'; +} + +/* Returns true if C may be the first byte in an identifier in the current + locale. + + (PSPP is transitioning to using Unicode internally for syntax, so please + use lex_uc_is_id1() instead, if possible.) */ +bool +lex_is_id1 (char c) +{ + return is_ascii_id1 (c) || (unsigned char) c >= 128; +} + +/* Returns true if C may be a byte in an identifier other than the first. + + (PSPP is transitioning to using Unicode internally for syntax, so please + use lex_uc_is_idn() instead, if possible.) */ +bool +lex_is_idn (char c) +{ + return is_ascii_idn (c) || (unsigned char) c >= 128; +} + +/* Returns true if Unicode code point UC may be the first character in an identifier in the current locale. */ bool -lex_is_id1 (char c_) +lex_uc_is_id1 (ucs4_t uc) { - unsigned char c = c_; - return isalpha (c) || c == '@' || c == '#' || c == '$' || c >= 128; + return is_ascii_id1 (uc) || (uc >= 0x80 && uc_is_property_id_start (uc)); } +/* Returns true if Unicode code point UC may be a character in an identifier + other than the first. */ +bool +lex_uc_is_idn (ucs4_t uc) +{ + return (is_ascii_id1 (uc) || isdigit (uc) || uc == '.' || uc == '_' + || (uc >= 0x80 && uc_is_property_id_continue (uc))); +} -/* Returns true if C may be a character in an identifier other - than the first. */ +/* Returns true if Unicode code point UC is a space that separates tokens. */ bool -lex_is_idn (char c_) +lex_uc_is_space (ucs4_t uc) { - unsigned char c = c_; - return lex_is_id1 (c) || isdigit (c) || c == '.' || c == '_'; + /* These are all of the Unicode characters in category Zs, Zl, or Zp. */ + return (uc == ' ' || (uc <= 0x000d && uc >= 0x0009) + || (uc >= 0x80 + && (uc == 0xa0 || uc == 0x85 || uc == 0x1680 || uc == 0x180e + || (uc >= 0x2000 && uc <= 0x200a) + || uc == 0x2028 || uc == 0x2029 || uc == 0x202f + || uc == 0x205f || uc == 0x3000))); } + /* Returns the length of the longest prefix of STRING that forms a valid identifier. Returns zero if STRING does not begin with a valid identifier. */ @@ -71,7 +120,10 @@ lex_id_get_length (struct substring string) Keywords match if one of the following is true: KEYWORD and TOKEN are identical, or TOKEN is at least 3 characters long - and those characters are identical to KEYWORD. */ + and those characters are identical to KEYWORD. (Letters that + differ only in case are considered identical.) + + KEYWORD must be ASCII, but TOKEN may be ASCII or UTF-8. */ bool lex_id_match (struct substring keyword, struct substring token) { @@ -79,7 +131,9 @@ lex_id_match (struct substring keyword, struct substring token) } /* Returns true if TOKEN is a case-insensitive match for at least - the first N characters of KEYWORD. */ + the first N characters of KEYWORD. + + KEYWORD must be ASCII, but TOKEN may be ASCII or UTF-8. */ bool lex_id_match_n (struct substring keyword, struct substring token, size_t n) { diff --git a/src/data/identifier.h b/src/data/identifier.h index 352ef2ea..85b99edb 100644 --- a/src/data/identifier.h +++ b/src/data/identifier.h @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 1997-9, 2000 Free Software Foundation, Inc. + Copyright (C) 1997-9, 2000, 2010 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -19,8 +19,8 @@ #include #include -#include -#include +#include +#include "libpspp/str.h" /* Token types. */ enum @@ -56,6 +56,9 @@ bool lex_is_keyword (int token); /* Recognizing identifiers. */ bool lex_is_id1 (char); bool lex_is_idn (char); +bool lex_uc_is_id1 (ucs4_t); +bool lex_uc_is_idn (ucs4_t); +bool lex_uc_is_space (ucs4_t); size_t lex_id_get_length (struct substring); /* Comparing identifiers. */