From 83d660c76fb1287f2cd9e6b94ddccb7069a6fae5 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Fri, 31 Dec 1999 00:04:07 +0000 Subject: Update. 1999-12-30 Ulrich Drepper * wcsmbs/wcscoll.c: Use multibyte character version. * wcsmbs/wcsxfrm.c: Likewise. * string/strcoll.c: Prepare to be used for the wide character version. * string/strxfrm.c: Likewise. * locale/weightwc.h: New file. --- ChangeLog | 8 ++++ string/strcoll.c | 101 +++++++++++++++++++++++++-------------- string/strxfrm.c | 141 ++++++++++++++++++++++++++++++++++++++++--------------- wcsmbs/wcscoll.c | 24 ++++------ wcsmbs/wcsxfrm.c | 28 +++++------ 5 files changed, 197 insertions(+), 105 deletions(-) diff --git a/ChangeLog b/ChangeLog index 9fd1e69d5f..35b9f26830 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,11 @@ +1999-12-30 Ulrich Drepper + + * wcsmbs/wcscoll.c: Use multibyte character version. + * wcsmbs/wcsxfrm.c: Likewise. + * string/strcoll.c: Prepare to be used for the wide character version. + * string/strxfrm.c: Likewise. + * locale/weightwc.h: New file. + 1999-12-30 Geoff Keating * sysdeps/powerpc/fpu/bits/fenvinline.h (feraiseexcept): Remove diff --git a/string/strcoll.c b/string/strcoll.c index 0f0a45a2d1..32d9124421 100644 --- a/string/strcoll.c +++ b/string/strcoll.c @@ -24,24 +24,36 @@ #include #include -#include "../locale/localeinfo.h" - -#ifdef USE_IN_EXTENDED_LOCALE_MODEL -# define STRCOLL __strcoll_l -#else -# define STRCOLL strcoll +#ifndef STRING_TYPE +# define STRING_TYPE char +# define USTRING_TYPE unsigned char +# ifdef USE_IN_EXTENDED_LOCALE_MODEL +# define STRCOLL __strcoll_l +# else +# define STRCOLL strcoll +# endif +# define STRCMP strcmp +# define STRLEN strlen +# define WEIGHT_H "../locale/weight.h" +# define SUFFIX MB +# define L(arg) arg #endif +#define CONCAT(a,b) CONCAT1(a,b) +#define CONCAT1(a,b) a##b + +#include "../locale/localeinfo.h" + #ifndef USE_IN_EXTENDED_LOCALE_MODEL int STRCOLL (s1, s2) - const char *s1; - const char *s2; + const STRING_TYPE *s1; + const STRING_TYPE *s2; #else int STRCOLL (s1, s2, l) - const char *s1; - const char *s2; + const STRING_TYPE *s1; + const STRING_TYPE *s2; __locale_t l; #endif { @@ -49,19 +61,19 @@ STRCOLL (s1, s2, l) struct locale_data *current = l->__locales[LC_COLLATE]; uint_fast32_t nrules = *((uint32_t *) current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].string); #else - uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); #endif /* We don't assign the following values right away since it might be unnecessary in case there are no rules. */ const unsigned char *rulesets; const int32_t *table; - const unsigned char *weights; - const unsigned char *extra; + const USTRING_TYPE *weights; + const USTRING_TYPE *extra; const int32_t *indirect; uint_fast32_t pass; int result = 0; - const unsigned char *us1; - const unsigned char *us2; + const USTRING_TYPE *us1; + const USTRING_TYPE *us2; size_t s1len; size_t s2len; int32_t *idx1arr; @@ -83,45 +95,62 @@ STRCOLL (s1, s2, l) int position; int seq1len; int seq2len; - int use_malloc = 0; + int use_malloc; +#ifdef WIDE_CHAR_VERSION + size_t size; + size_t layers; + const wint_t *names; +#endif -#include "../locale/weight.h" +#include WEIGHT_H if (nrules == 0) - return strcmp (s1, s2); + return STRCMP (s1, s2); #ifdef USE_IN_EXTENDED_LOCALE_MODEL rulesets = (const unsigned char *) current->values[_NL_ITEM_INDEX (_NL_COLLATE_RULESETS)].string; table = (const int32_t *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_TABLEMB)].string; - weights = (const unsigned char *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_WEIGHTMB)].string; - extra = (const unsigned char *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_EXTRAMB)].string; + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_TABLE,SUFFIX))].string; + weights = (const USTRING_TYPE *) + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_WEIGHT,SUFFIX))].string; + extra = (const USTRING_TYPE *) + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_EXTRA,SUFFIX))].string; indirect = (const int32_t *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_INDIRECTMB)].string; + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; +# ifdef WIDE_CHAR_VERSION + names = (const wint_t *) + current->values[_NL_ITEM_INDEX (_NL_COLLATE_NAMES)].string; + size = current->values[_NL_ITEM_INDEX (_NL_COLLATE_HASH_SIZE)].word; + layers = current->values[_NL_ITEM_INDEX (_NL_COLLATE_HASH_LAYERS)].word; +# endif #else rulesets = (const unsigned char *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_RULESETS); table = (const int32_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); - weights = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB); - extra = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_TABLE,SUFFIX)); + weights = (const USTRING_TYPE *) + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_WEIGHT,SUFFIX)); + extra = (const USTRING_TYPE *) + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_EXTRA,SUFFIX)); indirect = (const int32_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB); + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_INDIRECT,SUFFIX)); +# ifdef WIDE_CHAR_VERSION + names = (const wint_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_NAMES); + size = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_HASH_SIZE); + layers = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_HASH_LAYERS); +# endif #endif + use_malloc = 0; /* We need this a few times. */ - s1len = strlen (s1); - s2len = strlen (s2); + s1len = STRLEN (s1); + s2len = STRLEN (s2); /* We need the elements of the strings as unsigned values since they are used as indeces. */ - us1 = (const unsigned char *) s1; - us2 = (const unsigned char *) s2; + us1 = (const USTRING_TYPE *) s1; + us2 = (const USTRING_TYPE *) s2; /* Perform the first pass over the string and while doing this find and store the weights for each character. Since we want this to @@ -204,7 +233,7 @@ STRCOLL (s1, s2, l) { backw1_stop = idx1max; - while (*us1 != '\0') + while (*us1 != L('\0')) { int32_t tmp = findidx (&us1); rule1arr[idx1max] = tmp >> 24; @@ -263,7 +292,7 @@ STRCOLL (s1, s2, l) { backw2_stop = idx2max; - while (*us2 != '\0') + while (*us2 != L('\0')) { int32_t tmp = findidx (&us2); rule2arr[idx2max] = tmp >> 24; diff --git a/string/strxfrm.c b/string/strxfrm.c index 6fc795f9c8..9fd9526008 100644 --- a/string/strxfrm.c +++ b/string/strxfrm.c @@ -1,6 +1,6 @@ /* Copyright (C) 1995, 1996, 1997, 1998, 1999 Free Software Foundation, Inc. This file is part of the GNU C Library. - Written by Ulrich Drepper , 1995. + Written by Ulrich Drepper , 1995. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as @@ -23,15 +23,29 @@ #include #include -#include "../locale/localeinfo.h" - -#ifdef USE_IN_EXTENDED_LOCALE_MODEL -# define STRXFRM __strxfrm_l -#else -# define STRXFRM strxfrm +#ifndef STRING_TYPE +# define STRING_TYPE char +# define USTRING_TYPE unsigned char +# ifdef USE_IN_EXTENDED_LOCALE_MODEL +# define STRXFRM __strxfrm_l +# else +# define STRXFRM strxfrm +# endif +# define STRCMP strcmp +# define STRLEN strlen +# define STPNCPY __stpncpy +# define WEIGHT_H "../locale/weight.h" +# define SUFFIX MB +# define L(arg) arg #endif +#define CONCAT(a,b) CONCAT1(a,b) +#define CONCAT1(a,b) a##b +#include "../locale/localeinfo.h" + + +#ifndef WIDE_CHAR_VERSION /* These are definitions used by some of the functions for handling UTF-8 encoding below. */ static const uint32_t encoding_mask[] = @@ -79,14 +93,15 @@ utf8_encode (char *buf, int val) return buf - startp; } +#endif #ifndef USE_IN_EXTENDED_LOCALE_MODEL size_t -STRXFRM (char *dest, const char *src, size_t n) +STRXFRM (STRING_TYPE *dest, const STRING_TYPE *src, size_t n) #else size_t -STRXFRM (char *dest, const char *src, size_t n, __locale_t l) +STRXFRM (STRING_TYPE *dest, const STRING_TYPE *src, size_t n, __locale_t l) #endif { #ifdef USE_IN_EXTENDED_LOCALE_MODEL @@ -99,25 +114,30 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) unnecessary in case there are no rules. */ const unsigned char *rulesets; const int32_t *table; - const unsigned char *weights; - const unsigned char *extra; + const USTRING_TYPE *weights; + const USTRING_TYPE *extra; const int32_t *indirect; uint_fast32_t pass; size_t needed; - const unsigned char *usrc; - size_t srclen = strlen (src); + const USTRING_TYPE *usrc; + size_t srclen = STRLEN (src); int32_t *idxarr; unsigned char *rulearr; size_t idxmax; size_t idxcnt; - int use_malloc = 0; + int use_malloc; +#ifdef WIDE_CHAR_VERSION + size_t size; + size_t layers; + const wint_t *names; +#endif -#include "../locale/weight.h" +#include WEIGHT_H if (nrules == 0) { if (n != 0) - __stpncpy (dest, src, n); + STPNCPY (dest, src, n); return srclen; } @@ -126,37 +146,49 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) rulesets = (const unsigned char *) current->values[_NL_ITEM_INDEX (_NL_COLLATE_RULESETS)].string; table = (const int32_t *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_TABLEMB)].string; - weights = (const unsigned char *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_WEIGHTMB)].string; - extra = (const unsigned char *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_EXTRAMB)].string; + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_TABLE,SUFFIX))].string; + weights = (const USTRING_TYPE *) + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_WEIGHT,SUFFIX))].string; + extra = (const USTRING_TYPE *) + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_EXTRA,SUFFIX))].string; indirect = (const int32_t *) - current->values[_NL_ITEM_INDEX (_NL_COLLATE_INDIRECTMB)].string; + current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; +# ifdef WIDE_CHAR_VERSION + names = (const wint_t *) + current->values[_NL_ITEM_INDEX (_NL_COLLATE_NAMES)].string; + size = current->values[_NL_ITEM_INDEX (_NL_COLLATE_HASH_SIZE)].word; + layers = current->values[_NL_ITEM_INDEX (_NL_COLLATE_HASH_LAYERS)].word; +# endif #else rulesets = (const unsigned char *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_RULESETS); table = (const int32_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); - weights = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB); - extra = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_TABLE,SUFFIX)); + weights = (const USTRING_TYPE *) + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_WEIGHT,SUFFIX)); + extra = (const USTRING_TYPE *) + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_EXTRA,SUFFIX)); indirect = (const int32_t *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB); + _NL_CURRENT (LC_COLLATE, CONCAT(_NL_COLLATE_INDIRECT,SUFFIX)); +# ifdef WIDE_CHAR_VERSION + names = (const wint_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_NAMES); + size = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_HASH_SIZE); + layers = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_HASH_LAYERS); +# endif #endif + use_malloc = 0; /* Handle an empty string as a special case. */ if (srclen == 0) { if (n != 0) - *dest = '\0'; + *dest = L('\0'); return 1; } /* We need the elements of the string as unsigned values since they are used as indeces. */ - usrc = (const unsigned char *) src; + usrc = (const USTRING_TYPE *) src; /* Perform the first pass over the string and while doing this find and store the weights for each character. Since we want this to @@ -195,7 +227,7 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) ++idxmax; } - while (*usrc != '\0'); + while (*usrc != L('\0')); /* Now the passes over the weights. We now use the indeces we found before. */ @@ -287,8 +319,10 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) else { int val = 1; +#ifndef WIDE_CHAR_VERSION char buf[7]; size_t buflen; +#endif size_t i; for (idxcnt = 0; idxcnt < idxmax; ++idxcnt) @@ -307,6 +341,16 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) len = weights[idxarr[backw]++]; if (len != 0) { +#ifdef WIDE_CHAR_VERSION + if (needed + 1 + len < n) + { + dest[needed] = val; + for (i = 0; i < len; ++i) + dest[needed + 1 + i] = + weights[idxarr[backw] + i]; + } + needed += 1 + len; +#else buflen = utf8_encode (buf, val); if (needed + buflen + len < n) { @@ -316,8 +360,9 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) dest[needed + buflen + i] = weights[idxarr[backw] + i]; } - idxarr[backw] += len; needed += buflen + len; +#endif + idxarr[backw] += len; val = 1; } else @@ -331,6 +376,16 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) len = weights[idxarr[idxcnt]++]; if (len != 0) { +#ifdef WIDE_CHAR_VERSION + if (needed + 1+ len < n) + { + dest[needed] = val; + for (i = 0; i < len; ++i) + dest[needed + 1 + i] = + weights[idxarr[idxcnt] + i]; + } + needed += 1 + len; +#else buflen = utf8_encode (buf, val); if (needed + buflen + len < n) { @@ -340,8 +395,9 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) dest[needed + buflen + i] = weights[idxarr[idxcnt] + i]; } - idxarr[idxcnt] += len; needed += buflen + len; +#endif + idxarr[idxcnt] += len; val = 1; } else @@ -370,6 +426,16 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) size_t len = weights[idxarr[--backw]++]; if (len != 0) { +#ifdef WIDE_CHAR_VERSION + if (needed + 1 + len < n) + { + dest[needed] = val; + for (i = 0; i < len; ++i) + dest[needed + 1 + i] = + weights[idxarr[backw] + i]; + } + needed += 1 + len; +#else buflen = utf8_encode (buf, val); if (needed + buflen + len < n) { @@ -379,8 +445,9 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) dest[needed + buflen + i] = weights[idxarr[backw] + i]; } - idxarr[backw] += len; needed += buflen + len; +#endif + idxarr[backw] += len; val = 1; } else @@ -392,7 +459,7 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) /* Finally store the byte to separate the passes or terminate the string. */ if (needed < n) - dest[needed] = pass + 1 < nrules ? '\1' : '\0'; + dest[needed] = pass + 1 < nrules ? L('\1') : L('\0'); ++needed; } @@ -400,11 +467,11 @@ STRXFRM (char *dest, const char *src, size_t n, __locale_t l) a `position' rule at the end and if no non-ignored character is found the last \1 byte is immediately followed by a \0 byte signalling this. We can avoid the \1 byte(s). */ - if (needed <= n && needed > 2 && dest[needed - 2] == '\1') + if (needed <= n && needed > 2 && dest[needed - 2] == L('\1')) { /* Remove the \1 byte. */ --needed; - dest[needed - 1] = '\0'; + dest[needed - 1] = L('\0'); } /* Free the memory if needed. */ diff --git a/wcsmbs/wcscoll.c b/wcsmbs/wcscoll.c index b5b6f5cb45..1dfe93918b 100644 --- a/wcsmbs/wcscoll.c +++ b/wcsmbs/wcscoll.c @@ -19,26 +19,18 @@ #include +#define STRING_TYPE wchar_t +#define USTRING_TYPE wint_t #ifdef USE_IN_EXTENDED_LOCALE_MODEL # define STRCOLL __wcscoll_l #else # define STRCOLL wcscoll #endif #define STRCMP wcscmp +#define STRLEN wcslen +#define WEIGHT_H "../locale/weightwc.h" +#define SUFFIX WC +#define L(arg) L##arg +#define WIDE_CHAR_VERSION 1 - -#ifndef USE_IN_EXTENDED_LOCALE_MODEL -int -STRCOLL (s1, s2) - const wchar_t *s1; - const wchar_t *s2; -#else -int -STRCOLL (s1, s2, l) - const wchar_t *s1; - const wchar_t *s2; - __locale_t l; -#endif -{ - return STRCMP (s1, s2); -} +#include "../string/strcoll.c" diff --git a/wcsmbs/wcsxfrm.c b/wcsmbs/wcsxfrm.c index 99a359399e..c515657244 100644 --- a/wcsmbs/wcsxfrm.c +++ b/wcsmbs/wcsxfrm.c @@ -1,6 +1,6 @@ /* Copyright (C) 1996, 1997, 1998, 1999 Free Software Foundation, Inc. This file is part of the GNU C Library. - Contributed by Ulrich Drepper , 1996. + Contributed by Ulrich Drepper , 1996. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as @@ -19,23 +19,19 @@ #include +#define STRING_TYPE wchar_t +#define USTRING_TYPE wint_t #ifdef USE_IN_EXTENDED_LOCALE_MODEL # define STRXFRM __wcsxfrm_l #else # define STRXFRM wcsxfrm #endif - - -#ifndef USE_IN_EXTENDED_LOCALE_MODEL -size_t -STRXFRM (wchar_t *dest, const wchar_t *src, size_t n) -#else -size_t -STRXFRM (wchar_t *dest, const wchar_t *src, size_t n, __locale_t l) -#endif -{ - if (n != 0) - __wcpncpy (dest, src, n); - - return __wcslen (src); -} +#define STRCMP wcscmp +#define STRLEN wcslen +#define STPNCPY __wcpncpy +#define WEIGHT_H "../locale/weightwc.h" +#define SUFFIX WC +#define L(arg) L##arg +#define WIDE_CHAR_VERSION 1 + +#include "../string/strxfrm.c" -- cgit 1.4.1