diff options
author | Ulrich Drepper <drepper@redhat.com> | 2004-03-14 21:12:06 +0000 |
---|---|---|
committer | Ulrich Drepper <drepper@redhat.com> | 2004-03-14 21:12:06 +0000 |
commit | ccadf7b5346a3e21c692dfcbfcf38a63433bc36a (patch) | |
tree | a1e826af3345eee581925387513b1b6c76629c31 /string | |
parent | 3bc9b83f49805969b7d139b52ef6b003e5136354 (diff) | |
download | glibc-ccadf7b5346a3e21c692dfcbfcf38a63433bc36a.zip glibc-ccadf7b5346a3e21c692dfcbfcf38a63433bc36a.tar.gz glibc-ccadf7b5346a3e21c692dfcbfcf38a63433bc36a.tar.bz2 |
Update.
2004-03-14 Ulrich Drepper <drepper@redhat.com>
Make the non-_l functions wrappers around the _l functions.
* include/monetary.h: Declare __vstrmon_l.
* include/string.h: Add libc_hidden_proto for __strcoll_l and
__strxfrm_l.
* include/time.h: Define ptime_locale_status. Declare
__strptime_internal.
* include/wchar.h: Add libc_hidden_proto for __wcscoll_l and
__wcsxfrm_l.
* stdlib/strfmon.c: Move the code to strfmon_l.c. Add little
wrapper around __vstrfmon_l.
* stdlib/strfmon_l.c: Add real implementation. Split into new
function __vstrfmon_l to allow calling it from strfmon.
* stdlib/strtod.c: Move real code to strtod_l.c and add wrapper.
* stdlib/strtod_l.c: Add real implementation.
* stdlib/strtof.c: Adjust to changed strtod.c.
* stdlib/strtof_l.c: Include strtod_l.c now.
* stdlib/strtold.c: New file.
* stdlib/strtold_l.c: Removed.
* string/strcoll.c: Move real code to strcoll_l.c: Add wrapper.
* string/strcoll_l.c: Add real implementation.
* string/strxfrm.c: Move real code to strxfrm_l.c: Add wrapper.
* string/strxfrm_l.c: Add real implementation.
* sysdeps/generic/strtol.c: Move real implementation to strtol_l.c.
Add wrappers.
* sysdeps/generic/strtol_l.c: Add real implementation.
* sysdeps/generic/strtold.c: Removed.
* sysdeps/generic/strtold_l.c: New file.
* sysdeps/generic/strtoll_l.c: Include strtol_l.c now. Adjust
#defines.
* sysdeps/generic/strtoul_l.c: Likewise.
* sysdeps/generic/strtoull_l.c: Likewise.
* sysdeps/generic/wcstol_l.c: Likewise.
* sysdeps/generic/wcstoll_l.c: Likewise.
* sysdeps/generic/wcstoul_l.c: Likewise.
* sysdeps/generic/wcstoull_l.c: Likewise.
* sysdeps/ieee754/ldbl-128/strtold.c: Removed.
* sysdeps/ieee754/ldbl-128/strtold_l.c: New file.
* sysdeps/ieee754/ldbl-96/strtold.c: Removed.
* sysdeps/ieee754/ldbl-96/strtold_l.c: New file.
* sysdeps/m68k/strtold.c: Removed.
* sysdeps/m68k/strtold_l.c: New file.
* time/strftime.c: Move real code to strftime_l.c. Add wrapper.
* time/strftime_l.c: Add real implementation.
* time/strptime.c: Move real code to strptime_l.c. Add wrapper.
* time/strptime_l.c: Add real implementation.
* time/wcsftime.c: Simplify since only wrappers are defined in
strftime.c.
* time/wcsftime_l.c: Include strftime_l.c.
* wcsmbs/wcscoll.c: Simplify since the file is not used by wcscoll_l.c
anymore.
* wcsmbs/wcscoll_l.c: Include strcoll_l.c.
* wcsmbs/wcsxfrm.c: Simplify since the file is not used by wcsxfrm_l.c
anymore.
* wcsmbs/wcsxfrm_l.c: Include strxfrm_l.c.
* wcsmbs/wcstod.c: Prepare to include new strtod.c.
* wcsmbs/wcstod_l.c: Include strtod_l.c.
* wcsmbs/wcstof.c: Prepare to include new strtof.c.
* wcsmbs/wcstof_l.c: Include strtof_l.c.
* wcsmbs/wcstold.c: Prepare to include new strtold.c.
* wcsmbs/wcstold_l.c: Include strtold_l.c.
* locale/uselocale.c: Use _NL_CURRENT_LOCALE instead of __libc_tsd_get.
* sysdeps/generic/strcasecmp.c: Optimize a bit. It's better to get
a reference to the current locale and then use the _l functions.
* sysdeps/generic/strncase.c: Likewise.
Diffstat (limited to 'string')
-rw-r--r-- | string/strcoll.c | 528 | ||||
-rw-r--r-- | string/strcoll_l.c | 514 | ||||
-rw-r--r-- | string/strxfrm.c | 444 | ||||
-rw-r--r-- | string/strxfrm_l.c | 429 |
4 files changed, 948 insertions, 967 deletions
diff --git a/string/strcoll.c b/string/strcoll.c index 2607201..8a73cae 100644 --- a/string/strcoll.c +++ b/string/strcoll.c @@ -1,4 +1,4 @@ -/* Copyright (C) 1995,96,97,98,99,2000,2001,2002 Free Software Foundation, Inc. +/* Copyright (C) 1995-2002, 2004 Free Software Foundation, Inc. This file is part of the GNU C Library. Written by Ulrich Drepper <drepper@cygnus.com>, 1995. @@ -17,540 +17,24 @@ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. */ -#include <assert.h> -#include <langinfo.h> -#include <locale.h> -#include <stddef.h> -#include <stdint.h> -#include <stdlib.h> #include <string.h> #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 +# define STRCOLL strcoll +# define STRCOLL_L __strcoll_l #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 STRING_TYPE *s1; const STRING_TYPE *s2; -#else -int -STRCOLL (s1, s2, l) - const STRING_TYPE *s1; - const STRING_TYPE *s2; - __locale_t l; -#endif { -#ifdef USE_IN_EXTENDED_LOCALE_MODEL - struct locale_data *current = l->__locales[LC_COLLATE]; - uint_fast32_t nrules = current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].word; -#else - 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 USTRING_TYPE *weights; - const USTRING_TYPE *extra; - const int32_t *indirect; - uint_fast32_t pass; - int result = 0; - const USTRING_TYPE *us1; - const USTRING_TYPE *us2; - size_t s1len; - size_t s2len; - int32_t *idx1arr; - int32_t *idx2arr; - unsigned char *rule1arr; - unsigned char *rule2arr; - size_t idx1max; - size_t idx2max; - size_t idx1cnt; - size_t idx2cnt; - size_t idx1now; - size_t idx2now; - size_t backw1_stop; - size_t backw2_stop; - size_t backw1; - size_t backw2; - int val1; - int val2; - int position; - int seq1len; - int seq2len; - int use_malloc; - -#include WEIGHT_H - - if (nrules == 0) - 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 (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 (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; -#else - rulesets = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_RULESETS); - table = (const int32_t *) - _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, CONCAT(_NL_COLLATE_INDIRECT,SUFFIX)); -#endif - use_malloc = 0; - - assert (((uintptr_t) table) % __alignof__ (table[0]) == 0); - assert (((uintptr_t) weights) % __alignof__ (weights[0]) == 0); - assert (((uintptr_t) extra) % __alignof__ (extra[0]) == 0); - assert (((uintptr_t) indirect) % __alignof__ (indirect[0]) == 0); - - /* We need this a few times. */ - s1len = STRLEN (s1); - s2len = STRLEN (s2); - - /* Catch empty strings. */ - if (__builtin_expect (s1len == 0, 0) || __builtin_expect (s2len == 0, 0)) - return (s1len != 0) - (s2len != 0); - - /* We need the elements of the strings as unsigned values since they - are used as indeces. */ - 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 - be as fast as possible we are using `alloca' to store the temporary - values. But since there is no limit on the length of the string - we have to use `malloc' if the string is too long. We should be - very conservative here. - - Please note that the localedef programs makes sure that `position' - is not used at the first level. */ - if (! __libc_use_alloca (s1len + s2len)) - { - idx1arr = (int32_t *) malloc ((s1len + s2len) * (sizeof (int32_t) + 1)); - idx2arr = &idx1arr[s1len]; - rule1arr = (unsigned char *) &idx2arr[s2len]; - rule2arr = &rule1arr[s1len]; - - if (idx1arr == NULL) - /* No memory. Well, go with the stack then. - - XXX Once this implementation is stable we will handle this - differently. Instead of precomputing the indeces we will - do this in time. This means, though, that this happens for - every pass again. */ - goto try_stack; - use_malloc = 1; - } - else - { - try_stack: - idx1arr = (int32_t *) alloca (s1len * sizeof (int32_t)); - idx2arr = (int32_t *) alloca (s2len * sizeof (int32_t)); - rule1arr = (unsigned char *) alloca (s1len); - rule2arr = (unsigned char *) alloca (s2len); - } - - idx1cnt = 0; - idx2cnt = 0; - idx1max = 0; - idx2max = 0; - idx1now = 0; - idx2now = 0; - backw1_stop = ~0ul; - backw2_stop = ~0ul; - backw1 = ~0ul; - backw2 = ~0ul; - seq1len = 0; - seq2len = 0; - position = rulesets[0] & sort_position; - while (1) - { - val1 = 0; - val2 = 0; - - /* Get the next non-IGNOREd element for string `s1'. */ - if (seq1len == 0) - do - { - ++val1; - - if (backw1_stop != ~0ul) - { - /* The is something pushed. */ - if (backw1 == backw1_stop) - { - /* The last pushed character was handled. Continue - with forward characters. */ - if (idx1cnt < idx1max) - idx1now = idx1cnt; - else - /* Nothing anymore. The backward sequence ended with - the last sequence in the string. Note that seq1len - is still zero. */ - break; - } - else - idx1now = --backw1; - } - else - { - backw1_stop = idx1max; - - while (*us1 != L('\0')) - { - int32_t tmp = findidx (&us1); - rule1arr[idx1max] = tmp >> 24; - idx1arr[idx1max] = tmp & 0xffffff; - idx1cnt = idx1max++; - - if ((rulesets[rule1arr[idx1cnt] * nrules] - & sort_backward) == 0) - /* No more backward characters to push. */ - break; - ++idx1cnt; - } - - if (backw1_stop >= idx1cnt) - { - /* No sequence at all or just one. */ - if (idx1cnt == idx1max || backw1_stop > idx1cnt) - /* Note that seq1len is still zero. */ - break; - - backw1_stop = ~0ul; - idx1now = idx1cnt; - } - else - /* We pushed backward sequences. */ - idx1now = backw1 = idx1cnt - 1; - } - } - while ((seq1len = weights[idx1arr[idx1now]++]) == 0); - - /* And the same for string `s2'. */ - if (seq2len == 0) - do - { - ++val2; - - if (backw2_stop != ~0ul) - { - /* The is something pushed. */ - if (backw2 == backw2_stop) - { - /* The last pushed character was handled. Continue - with forward characters. */ - if (idx2cnt < idx2max) - idx2now = idx2cnt; - else - /* Nothing anymore. The backward sequence ended with - the last sequence in the string. Note that seq2len - is still zero. */ - break; - } - else - idx2now = --backw2; - } - else - { - backw2_stop = idx2max; - - while (*us2 != L('\0')) - { - int32_t tmp = findidx (&us2); - rule2arr[idx2max] = tmp >> 24; - idx2arr[idx2max] = tmp & 0xffffff; - idx2cnt = idx2max++; - - if ((rulesets[rule2arr[idx2cnt] * nrules] - & sort_backward) == 0) - /* No more backward characters to push. */ - break; - ++idx2cnt; - } - - if (backw2_stop >= idx2cnt) - { - /* No sequence at all or just one. */ - if (idx2cnt == idx2max || backw2_stop > idx2cnt) - /* Note that seq1len is still zero. */ - break; - - backw2_stop = ~0ul; - idx2now = idx2cnt; - } - else - /* We pushed backward sequences. */ - idx2now = backw2 = idx2cnt - 1; - } - } - while ((seq2len = weights[idx2arr[idx2now]++]) == 0); - - /* See whether any or both strings are empty. */ - if (seq1len == 0 || seq2len == 0) - { - if (seq1len == seq2len) - /* Both ended. So far so good, both strings are equal at the - first level. */ - break; - - /* This means one string is shorter than the other. Find out - which one and return an appropriate value. */ - result = seq1len == 0 ? -1 : 1; - goto free_and_return; - } - - /* Test for position if necessary. */ - if (position && val1 != val2) - { - result = val1 - val2; - goto free_and_return; - } - - /* Compare the two sequences. */ - do - { - if (weights[idx1arr[idx1now]] != weights[idx2arr[idx2now]]) - { - /* The sequences differ. */ - result = weights[idx1arr[idx1now]] - weights[idx2arr[idx2now]]; - goto free_and_return; - } - - /* Increment the offsets. */ - ++idx1arr[idx1now]; - ++idx2arr[idx2now]; - - --seq1len; - --seq2len; - } - while (seq1len > 0 && seq2len > 0); - - if (position && seq1len != seq2len) - { - result = seq1len - seq2len; - goto free_and_return; - } - } - - /* Now the remaining passes over the weights. We now use the - indeces we found before. */ - for (pass = 1; pass < nrules; ++pass) - { - /* We assume that if a rule has defined `position' in one section - this is true for all of them. */ - idx1cnt = 0; - idx2cnt = 0; - backw1_stop = ~0ul; - backw2_stop = ~0ul; - backw1 = ~0ul; - backw2 = ~0ul; - position = rulesets[rule1arr[0] * nrules + pass] & sort_position; - - while (1) - { - val1 = 0; - val2 = 0; - - /* Get the next non-IGNOREd element for string `s1'. */ - if (seq1len == 0) - do - { - ++val1; - - if (backw1_stop != ~0ul) - { - /* The is something pushed. */ - if (backw1 == backw1_stop) - { - /* The last pushed character was handled. Continue - with forward characters. */ - if (idx1cnt < idx1max) - idx1now = idx1cnt; - else - { - /* Nothing anymore. The backward sequence - ended with the last sequence in the string. */ - idx1now = ~0ul; - break; - } - } - else - idx1now = --backw1; - } - else - { - backw1_stop = idx1cnt; - - while (idx1cnt < idx1max) - { - if ((rulesets[rule1arr[idx1cnt] * nrules + pass] - & sort_backward) == 0) - /* No more backward characters to push. */ - break; - ++idx1cnt; - } - - if (backw1_stop == idx1cnt) - { - /* No sequence at all or just one. */ - if (idx1cnt == idx1max) - /* Note that seq1len is still zero. */ - break; - - backw1_stop = ~0ul; - idx1now = idx1cnt++; - } - else - /* We pushed backward sequences. */ - idx1now = backw1 = idx1cnt - 1; - } - } - while ((seq1len = weights[idx1arr[idx1now]++]) == 0); - - /* And the same for string `s2'. */ - if (seq2len == 0) - do - { - ++val2; - - if (backw2_stop != ~0ul) - { - /* The is something pushed. */ - if (backw2 == backw2_stop) - { - /* The last pushed character was handled. Continue - with forward characters. */ - if (idx2cnt < idx2max) - idx2now = idx2cnt; - else - { - /* Nothing anymore. The backward sequence - ended with the last sequence in the string. */ - idx2now = ~0ul; - break; - } - } - else - idx2now = --backw2; - } - else - { - backw2_stop = idx2cnt; - - while (idx2cnt < idx2max) - { - if ((rulesets[rule2arr[idx2cnt] * nrules + pass] - & sort_backward) == 0) - /* No more backward characters to push. */ - break; - ++idx2cnt; - } - - if (backw2_stop == idx2cnt) - { - /* No sequence at all or just one. */ - if (idx2cnt == idx2max) - /* Note that seq2len is still zero. */ - break; - - backw2_stop = ~0ul; - idx2now = idx2cnt++; - } - else - /* We pushed backward sequences. */ - idx2now = backw2 = idx2cnt - 1; - } - } - while ((seq2len = weights[idx2arr[idx2now]++]) == 0); - - /* See whether any or both strings are empty. */ - if (seq1len == 0 || seq2len == 0) - { - if (seq1len == seq2len) - /* Both ended. So far so good, both strings are equal - at this level. */ - break; - - /* This means one string is shorter than the other. Find out - which one and return an appropriate value. */ - result = seq1len == 0 ? -1 : 1; - goto free_and_return; - } - - /* Test for position if necessary. */ - if (position && val1 != val2) - { - result = val1 - val2; - goto free_and_return; - } - - /* Compare the two sequences. */ - do - { - if (weights[idx1arr[idx1now]] != weights[idx2arr[idx2now]]) - { - /* The sequences differ. */ - result = (weights[idx1arr[idx1now]] - - weights[idx2arr[idx2now]]); - goto free_and_return; - } - - /* Increment the offsets. */ - ++idx1arr[idx1now]; - ++idx2arr[idx2now]; - - --seq1len; - --seq2len; - } - while (seq1len > 0 && seq2len > 0); - - if (position && seq1len != seq2len) - { - result = seq1len - seq2len; - goto free_and_return; - } - } - } - - /* Free the memory if needed. */ - free_and_return: - if (use_malloc) - free (idx1arr); - - return result; + return STRCOLL_L (s1, s2, _NL_CURRENT_LOCALE); } -#if !defined WIDE_CHAR_VERSION && !defined USE_IN_EXTENDED_LOCALE_MODEL +#if !defined WIDE_CHAR_VERSION libc_hidden_def (strcoll) #endif diff --git a/string/strcoll_l.c b/string/strcoll_l.c index 6611589..c46921d 100644 --- a/string/strcoll_l.c +++ b/string/strcoll_l.c @@ -1,4 +1,4 @@ -/* Copyright (C) 1995,96,97,2002 Free Software Foundation, Inc. +/* Copyright (C) 1995,96,97,2002, 2004 Free Software Foundation, Inc. This file is part of the GNU C Library. Written by Ulrich Drepper <drepper@gnu.org>, 1995. @@ -17,7 +17,515 @@ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. */ -#define USE_IN_EXTENDED_LOCALE_MODEL 1 -#include <strcoll.c> +#include <assert.h> +#include <langinfo.h> +#include <locale.h> +#include <stddef.h> +#include <stdint.h> +#include <stdlib.h> +#include <string.h> + +#ifndef STRING_TYPE +# define STRING_TYPE char +# define USTRING_TYPE unsigned char +# define STRCOLL __strcoll_l +# 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" + +int +STRCOLL (s1, s2, l) + const STRING_TYPE *s1; + const STRING_TYPE *s2; + __locale_t l; +{ + struct locale_data *current = l->__locales[LC_COLLATE]; + uint_fast32_t nrules = current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].word; + /* 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 USTRING_TYPE *weights; + const USTRING_TYPE *extra; + const int32_t *indirect; + uint_fast32_t pass; + int result = 0; + const USTRING_TYPE *us1; + const USTRING_TYPE *us2; + size_t s1len; + size_t s2len; + int32_t *idx1arr; + int32_t *idx2arr; + unsigned char *rule1arr; + unsigned char *rule2arr; + size_t idx1max; + size_t idx2max; + size_t idx1cnt; + size_t idx2cnt; + size_t idx1now; + size_t idx2now; + size_t backw1_stop; + size_t backw2_stop; + size_t backw1; + size_t backw2; + int val1; + int val2; + int position; + int seq1len; + int seq2len; + int use_malloc; + +#include WEIGHT_H + + if (nrules == 0) + return STRCMP (s1, s2); + + rulesets = (const unsigned char *) + current->values[_NL_ITEM_INDEX (_NL_COLLATE_RULESETS)].string; + table = (const int32_t *) + 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 (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; + use_malloc = 0; + + assert (((uintptr_t) table) % __alignof__ (table[0]) == 0); + assert (((uintptr_t) weights) % __alignof__ (weights[0]) == 0); + assert (((uintptr_t) extra) % __alignof__ (extra[0]) == 0); + assert (((uintptr_t) indirect) % __alignof__ (indirect[0]) == 0); + + /* We need this a few times. */ + s1len = STRLEN (s1); + s2len = STRLEN (s2); + + /* Catch empty strings. */ + if (__builtin_expect (s1len == 0, 0) || __builtin_expect (s2len == 0, 0)) + return (s1len != 0) - (s2len != 0); + + /* We need the elements of the strings as unsigned values since they + are used as indeces. */ + 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 + be as fast as possible we are using `alloca' to store the temporary + values. But since there is no limit on the length of the string + we have to use `malloc' if the string is too long. We should be + very conservative here. + + Please note that the localedef programs makes sure that `position' + is not used at the first level. */ + if (! __libc_use_alloca (s1len + s2len)) + { + idx1arr = (int32_t *) malloc ((s1len + s2len) * (sizeof (int32_t) + 1)); + idx2arr = &idx1arr[s1len]; + rule1arr = (unsigned char *) &idx2arr[s2len]; + rule2arr = &rule1arr[s1len]; + + if (idx1arr == NULL) + /* No memory. Well, go with the stack then. + + XXX Once this implementation is stable we will handle this + differently. Instead of precomputing the indeces we will + do this in time. This means, though, that this happens for + every pass again. */ + goto try_stack; + use_malloc = 1; + } + else + { + try_stack: + idx1arr = (int32_t *) alloca (s1len * sizeof (int32_t)); + idx2arr = (int32_t *) alloca (s2len * sizeof (int32_t)); + rule1arr = (unsigned char *) alloca (s1len); + rule2arr = (unsigned char *) alloca (s2len); + } + + idx1cnt = 0; + idx2cnt = 0; + idx1max = 0; + idx2max = 0; + idx1now = 0; + idx2now = 0; + backw1_stop = ~0ul; + backw2_stop = ~0ul; + backw1 = ~0ul; + backw2 = ~0ul; + seq1len = 0; + seq2len = 0; + position = rulesets[0] & sort_position; + while (1) + { + val1 = 0; + val2 = 0; + + /* Get the next non-IGNOREd element for string `s1'. */ + if (seq1len == 0) + do + { + ++val1; + + if (backw1_stop != ~0ul) + { + /* The is something pushed. */ + if (backw1 == backw1_stop) + { + /* The last pushed character was handled. Continue + with forward characters. */ + if (idx1cnt < idx1max) + idx1now = idx1cnt; + else + /* Nothing anymore. The backward sequence ended with + the last sequence in the string. Note that seq1len + is still zero. */ + break; + } + else + idx1now = --backw1; + } + else + { + backw1_stop = idx1max; + + while (*us1 != L('\0')) + { + int32_t tmp = findidx (&us1); + rule1arr[idx1max] = tmp >> 24; + idx1arr[idx1max] = tmp & 0xffffff; + idx1cnt = idx1max++; + + if ((rulesets[rule1arr[idx1cnt] * nrules] + & sort_backward) == 0) + /* No more backward characters to push. */ + break; + ++idx1cnt; + } + + if (backw1_stop >= idx1cnt) + { + /* No sequence at all or just one. */ + if (idx1cnt == idx1max || backw1_stop > idx1cnt) + /* Note that seq1len is still zero. */ + break; + + backw1_stop = ~0ul; + idx1now = idx1cnt; + } + else + /* We pushed backward sequences. */ + idx1now = backw1 = idx1cnt - 1; + } + } + while ((seq1len = weights[idx1arr[idx1now]++]) == 0); + + /* And the same for string `s2'. */ + if (seq2len == 0) + do + { + ++val2; + + if (backw2_stop != ~0ul) + { + /* The is something pushed. */ + if (backw2 == backw2_stop) + { + /* The last pushed character was handled. Continue + with forward characters. */ + if (idx2cnt < idx2max) + idx2now = idx2cnt; + else + /* Nothing anymore. The backward sequence ended with + the last sequence in the string. Note that seq2len + is still zero. */ + break; + } + else + idx2now = --backw2; + } + else + { + backw2_stop = idx2max; + + while (*us2 != L('\0')) + { + int32_t tmp = findidx (&us2); + rule2arr[idx2max] = tmp >> 24; + idx2arr[idx2max] = tmp & 0xffffff; + idx2cnt = idx2max++; + + if ((rulesets[rule2arr[idx2cnt] * nrules] + & sort_backward) == 0) + /* No more backward characters to push. */ + break; + ++idx2cnt; + } + + if (backw2_stop >= idx2cnt) + { + /* No sequence at all or just one. */ + if (idx2cnt == idx2max || backw2_stop > idx2cnt) + /* Note that seq1len is still zero. */ + break; + + backw2_stop = ~0ul; + idx2now = idx2cnt; + } + else + /* We pushed backward sequences. */ + idx2now = backw2 = idx2cnt - 1; + } + } + while ((seq2len = weights[idx2arr[idx2now]++]) == 0); + + /* See whether any or both strings are empty. */ + if (seq1len == 0 || seq2len == 0) + { + if (seq1len == seq2len) + /* Both ended. So far so good, both strings are equal at the + first level. */ + break; + + /* This means one string is shorter than the other. Find out + which one and return an appropriate value. */ + result = seq1len == 0 ? -1 : 1; + goto free_and_return; + } + + /* Test for position if necessary. */ + if (position && val1 != val2) + { + result = val1 - val2; + goto free_and_return; + } + + /* Compare the two sequences. */ + do + { + if (weights[idx1arr[idx1now]] != weights[idx2arr[idx2now]]) + { + /* The sequences differ. */ + result = weights[idx1arr[idx1now]] - weights[idx2arr[idx2now]]; + goto free_and_return; + } + + /* Increment the offsets. */ + ++idx1arr[idx1now]; + ++idx2arr[idx2now]; + + --seq1len; + --seq2len; + } + while (seq1len > 0 && seq2len > 0); + + if (position && seq1len != seq2len) + { + result = seq1len - seq2len; + goto free_and_return; + } + } + + /* Now the remaining passes over the weights. We now use the + indeces we found before. */ + for (pass = 1; pass < nrules; ++pass) + { + /* We assume that if a rule has defined `position' in one section + this is true for all of them. */ + idx1cnt = 0; + idx2cnt = 0; + backw1_stop = ~0ul; + backw2_stop = ~0ul; + backw1 = ~0ul; + backw2 = ~0ul; + position = rulesets[rule1arr[0] * nrules + pass] & sort_position; + + while (1) + { + val1 = 0; + val2 = 0; + + /* Get the next non-IGNOREd element for string `s1'. */ + if (seq1len == 0) + do + { + ++val1; + + if (backw1_stop != ~0ul) + { + /* The is something pushed. */ + if (backw1 == backw1_stop) + { + /* The last pushed character was handled. Continue + with forward characters. */ + if (idx1cnt < idx1max) + idx1now = idx1cnt; + else + { + /* Nothing anymore. The backward sequence + ended with the last sequence in the string. */ + idx1now = ~0ul; + break; + } + } + else + idx1now = --backw1; + } + else + { + backw1_stop = idx1cnt; + + while (idx1cnt < idx1max) + { + if ((rulesets[rule1arr[idx1cnt] * nrules + pass] + & sort_backward) == 0) + /* No more backward characters to push. */ + break; + ++idx1cnt; + } + + if (backw1_stop == idx1cnt) + { + /* No sequence at all or just one. */ + if (idx1cnt == idx1max) + /* Note that seq1len is still zero. */ + break; + + backw1_stop = ~0ul; + idx1now = idx1cnt++; + } + else + /* We pushed backward sequences. */ + idx1now = backw1 = idx1cnt - 1; + } + } + while ((seq1len = weights[idx1arr[idx1now]++]) == 0); + + /* And the same for string `s2'. */ + if (seq2len == 0) + do + { + ++val2; + + if (backw2_stop != ~0ul) + { + /* The is something pushed. */ + if (backw2 == backw2_stop) + { + /* The last pushed character was handled. Continue + with forward characters. */ + if (idx2cnt < idx2max) + idx2now = idx2cnt; + else + { + /* Nothing anymore. The backward sequence + ended with the last sequence in the string. */ + idx2now = ~0ul; + break; + } + } + else + idx2now = --backw2; + } + else + { + backw2_stop = idx2cnt; + + while (idx2cnt < idx2max) + { + if ((rulesets[rule2arr[idx2cnt] * nrules + pass] + & sort_backward) == 0) + /* No more backward characters to push. */ + break; + ++idx2cnt; + } + + if (backw2_stop == idx2cnt) + { + /* No sequence at all or just one. */ + if (idx2cnt == idx2max) + /* Note that seq2len is still zero. */ + break; + + backw2_stop = ~0ul; + idx2now = idx2cnt++; + } + else + /* We pushed backward sequences. */ + idx2now = backw2 = idx2cnt - 1; + } + } + while ((seq2len = weights[idx2arr[idx2now]++]) == 0); + + /* See whether any or both strings are empty. */ + if (seq1len == 0 || seq2len == 0) + { + if (seq1len == seq2len) + /* Both ended. So far so good, both strings are equal + at this level. */ + break; + + /* This means one string is shorter than the other. Find out + which one and return an appropriate value. */ + result = seq1len == 0 ? -1 : 1; + goto free_and_return; + } + + /* Test for position if necessary. */ + if (position && val1 != val2) + { + result = val1 - val2; + goto free_and_return; + } + + /* Compare the two sequences. */ + do + { + if (weights[idx1arr[idx1now]] != weights[idx2arr[idx2now]]) + { + /* The sequences differ. */ + result = (weights[idx1arr[idx1now]] + - weights[idx2arr[idx2now]]); + goto free_and_return; + } + + /* Increment the offsets. */ + ++idx1arr[idx1now]; + ++idx2arr[idx2now]; + + --seq1len; + --seq2len; + } + while (seq1len > 0 && seq2len > 0); + + if (position && seq1len != seq2len) + { + result = seq1len - seq2len; + goto free_and_return; + } + } + } + + /* Free the memory if needed. */ + free_and_return: + if (use_malloc) + free (idx1arr); + + return result; +} +libc_hidden_def (STRCOLL) + +#ifndef WIDE_CHAR_VERSION weak_alias (__strcoll_l, strcoll_l) +#endif diff --git a/string/strxfrm.c b/string/strxfrm.c index 549d68c..840f270 100644 --- a/string/strxfrm.c +++ b/string/strxfrm.c @@ -1,4 +1,4 @@ -/* Copyright (C) 1995-1999,2000,2001,2002,2003 Free Software Foundation, Inc. +/* Copyright (C) 1995-1999,2003, 2004 Free Software Foundation, Inc. This file is part of the GNU C Library. Written by Ulrich Drepper <drepper@cygnus.com>, 1995. @@ -17,451 +17,17 @@ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. */ -#include <assert.h> -#include <langinfo.h> -#include <locale.h> -#include <stddef.h> -#include <stdint.h> -#include <stdlib.h> #include <string.h> -#include <sys/param.h> +#include <locale/localeinfo.h> #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 +# define STRXFRM strxfrm +# define STRXFRM_L __strxfrm_l #endif -#define CONCAT(a,b) CONCAT1(a,b) -#define CONCAT1(a,b) a##b - -#include "../locale/localeinfo.h" - - -#ifndef WIDE_CHAR_VERSION - -/* We need UTF-8 encoding of numbers. */ -static int -utf8_encode (char *buf, int val) -{ - int retval; - - if (val < 0x80) - { - *buf++ = (char) val; - retval = 1; - } - else - { - int step; - - for (step = 2; step < 6; ++step) - if ((val & (~(uint32_t)0 << (5 * step + 1))) == 0) - break; - retval = step; - - *buf = (unsigned char) (~0xff >> step); - --step; - do - { - buf[step] = 0x80 | (val & 0x3f); - val >>= 6; - } - while (--step > 0); - *buf |= val; - } - - return retval; -} -#endif - - -#ifndef USE_IN_EXTENDED_LOCALE_MODEL size_t STRXFRM (STRING_TYPE *dest, const STRING_TYPE *src, size_t n) -#else -size_t -STRXFRM (STRING_TYPE *dest, const STRING_TYPE *src, size_t n, __locale_t l) -#endif { -#ifdef USE_IN_EXTENDED_LOCALE_MODEL - struct locale_data *current = l->__locales[LC_COLLATE]; - uint_fast32_t nrules = current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].word; -#else - uint32_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 USTRING_TYPE *weights; - const USTRING_TYPE *extra; - const int32_t *indirect; - uint_fast32_t pass; - size_t needed; - const USTRING_TYPE *usrc; - size_t srclen = STRLEN (src); - int32_t *idxarr; - unsigned char *rulearr; - size_t idxmax; - size_t idxcnt; - int use_malloc; - -#include WEIGHT_H - - if (nrules == 0) - { - if (n != 0) - STPNCPY (dest, src, MIN (srclen + 1, n)); - - return srclen; - } - -#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 (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 (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; -#else - rulesets = (const unsigned char *) - _NL_CURRENT (LC_COLLATE, _NL_COLLATE_RULESETS); - table = (const int32_t *) - _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, CONCAT(_NL_COLLATE_INDIRECT,SUFFIX)); -#endif - use_malloc = 0; - - assert (((uintptr_t) table) % __alignof__ (table[0]) == 0); - assert (((uintptr_t) weights) % __alignof__ (weights[0]) == 0); - assert (((uintptr_t) extra) % __alignof__ (extra[0]) == 0); - assert (((uintptr_t) indirect) % __alignof__ (indirect[0]) == 0); - - /* Handle an empty string as a special case. */ - if (srclen == 0) - { - if (n != 0) - *dest = L('\0'); - return 0; - } - - /* We need the elements of the string as unsigned values since they - are used as indeces. */ - 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 - be as fast as possible we are using `alloca' to store the temporary - values. But since there is no limit on the length of the string - we have to use `malloc' if the string is too long. We should be - very conservative here. */ - if (! __libc_use_alloca (srclen)) - { - idxarr = (int32_t *) malloc ((srclen + 1) * (sizeof (int32_t) + 1)); - rulearr = (unsigned char *) &idxarr[srclen]; - - if (idxarr == NULL) - /* No memory. Well, go with the stack then. - - XXX Once this implementation is stable we will handle this - differently. Instead of precomputing the indeces we will - do this in time. This means, though, that this happens for - every pass again. */ - goto try_stack; - use_malloc = 1; - } - else - { - try_stack: - idxarr = (int32_t *) alloca (srclen * sizeof (int32_t)); - rulearr = (unsigned char *) alloca (srclen + 1); - } - - idxmax = 0; - do - { - int32_t tmp = findidx (&usrc); - rulearr[idxmax] = tmp >> 24; - idxarr[idxmax] = tmp & 0xffffff; - - ++idxmax; - } - while (*usrc != L('\0')); - - /* This element is only read, the value never used but to determine - another value which then is ignored. */ - rulearr[idxmax] = '\0'; - - /* Now the passes over the weights. We now use the indeces we found - before. */ - needed = 0; - for (pass = 0; pass < nrules; ++pass) - { - size_t backw_stop = ~0ul; - int rule = rulesets[rulearr[0] * nrules + pass]; - /* We assume that if a rule has defined `position' in one section - this is true for all of them. */ - int position = rule & sort_position; - - if (position == 0) - { - for (idxcnt = 0; idxcnt < idxmax; ++idxcnt) - { - if ((rule & sort_forward) != 0) - { - size_t len; - - if (backw_stop != ~0ul) - { - /* Handle the pushed elements now. */ - size_t backw; - - for (backw = idxcnt - 1; backw >= backw_stop; --backw) - { - len = weights[idxarr[backw]++]; - - if (needed + len < n) - while (len-- > 0) - dest[needed++] = weights[idxarr[backw]++]; - else - { - /* No more characters fit into the buffer. */ - needed += len; - idxarr[backw] += len; - } - } - - backw_stop = ~0ul; - } - - /* Now handle the forward element. */ - len = weights[idxarr[idxcnt]++]; - if (needed + len < n) - while (len-- > 0) - dest[needed++] = weights[idxarr[idxcnt]++]; - else - { - /* No more characters fit into the buffer. */ - needed += len; - idxarr[idxcnt] += len; - } - } - else - { - /* Remember where the backwards series started. */ - if (backw_stop == ~0ul) - backw_stop = idxcnt; - } - - rule = rulesets[rulearr[idxcnt + 1] * nrules + pass]; - } - - - if (backw_stop != ~0ul) - { - /* Handle the pushed elements now. */ - size_t backw; - - backw = idxcnt; - while (backw > backw_stop) - { - size_t len = weights[idxarr[--backw]++]; - - if (needed + len < n) - while (len-- > 0) - dest[needed++] = weights[idxarr[backw]++]; - else - { - /* No more characters fit into the buffer. */ - needed += len; - idxarr[backw] += len; - } - } - } - } - else - { - int val = 1; -#ifndef WIDE_CHAR_VERSION - char buf[7]; - size_t buflen; -#endif - size_t i; - - for (idxcnt = 0; idxcnt < idxmax; ++idxcnt) - { - if ((rule & sort_forward) != 0) - { - size_t len; - - if (backw_stop != ~0ul) - { - /* Handle the pushed elements now. */ - size_t backw; - - for (backw = idxcnt - 1; backw >= backw_stop; --backw) - { - 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) - { - for (i = 0; i < buflen; ++i) - dest[needed + i] = buf[i]; - for (i = 0; i < len; ++i) - dest[needed + buflen + i] = - weights[idxarr[backw] + i]; - } - needed += buflen + len; -#endif - idxarr[backw] += len; - val = 1; - } - else - ++val; - } - - backw_stop = ~0ul; - } - - /* Now handle the forward element. */ - 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) - { - for (i = 0; i < buflen; ++i) - dest[needed + i] = buf[i]; - for (i = 0; i < len; ++i) - dest[needed + buflen + i] = - weights[idxarr[idxcnt] + i]; - } - needed += buflen + len; -#endif - idxarr[idxcnt] += len; - val = 1; - } - else - /* Note that we don't have to increment `idxarr[idxcnt]' - since the length is zero. */ - ++val; - } - else - { - /* Remember where the backwards series started. */ - if (backw_stop == ~0ul) - backw_stop = idxcnt; - } - - rule = rulesets[rulearr[idxcnt + 1] * nrules + pass]; - } - - if (backw_stop != ~0ul) - { - /* Handle the pushed elements now. */ - size_t backw; - - backw = idxmax - 1; - while (backw > backw_stop) - { - 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) - { - for (i = 0; i < buflen; ++i) - dest[needed + i] = buf[i]; - for (i = 0; i < len; ++i) - dest[needed + buflen + i] = - weights[idxarr[backw] + i]; - } - needed += buflen + len; -#endif - idxarr[backw] += len; - val = 1; - } - else - ++val; - } - } - } - - /* Finally store the byte to separate the passes or terminate - the string. */ - if (needed < n) - dest[needed] = pass + 1 < nrules ? L('\1') : L('\0'); - ++needed; - } - - /* This is a little optimization: many collation specifications have - 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] == L('\1')) - { - /* Remove the \1 byte. */ - --needed; - dest[needed - 1] = L('\0'); - } - - /* Free the memory if needed. */ - if (use_malloc) - free (idxarr); - - /* Return the number of bytes/words we need, but don't count the NUL - byte/word at the end. */ - return needed - 1; + return STRXFRM_L (dest, src, n, _NL_CURRENT_LOCALE); } diff --git a/string/strxfrm_l.c b/string/strxfrm_l.c index 264ab9b..44b6051 100644 --- a/string/strxfrm_l.c +++ b/string/strxfrm_l.c @@ -1,4 +1,4 @@ -/* Copyright (C) 1995,96,97,2002 Free Software Foundation, Inc. +/* Copyright (C) 1995,96,97,2002, 2004 Free Software Foundation, Inc. This file is part of the GNU C Library. Written by Ulrich Drepper <drepper@gnu.org>, 1995. @@ -17,7 +17,430 @@ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. */ -#define USE_IN_EXTENDED_LOCALE_MODEL 1 -#include <strxfrm.c> +#include <assert.h> +#include <langinfo.h> +#include <locale.h> +#include <stddef.h> +#include <stdint.h> +#include <stdlib.h> +#include <string.h> +#include <sys/param.h> +#ifndef STRING_TYPE +# define STRING_TYPE char +# define USTRING_TYPE unsigned char +# define STRXFRM __strxfrm_l +# 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 + +/* We need UTF-8 encoding of numbers. */ +static int +utf8_encode (char *buf, int val) +{ + int retval; + + if (val < 0x80) + { + *buf++ = (char) val; + retval = 1; + } + else + { + int step; + + for (step = 2; step < 6; ++step) + if ((val & (~(uint32_t)0 << (5 * step + 1))) == 0) + break; + retval = step; + + *buf = (unsigned char) (~0xff >> step); + --step; + do + { + buf[step] = 0x80 | (val & 0x3f); + val >>= 6; + } + while (--step > 0); + *buf |= val; + } + + return retval; +} +#endif + + +size_t +STRXFRM (STRING_TYPE *dest, const STRING_TYPE *src, size_t n, __locale_t l) +{ + struct locale_data *current = l->__locales[LC_COLLATE]; + uint_fast32_t nrules = current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].word; + /* 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 USTRING_TYPE *weights; + const USTRING_TYPE *extra; + const int32_t *indirect; + uint_fast32_t pass; + size_t needed; + const USTRING_TYPE *usrc; + size_t srclen = STRLEN (src); + int32_t *idxarr; + unsigned char *rulearr; + size_t idxmax; + size_t idxcnt; + int use_malloc; + +#include WEIGHT_H + + if (nrules == 0) + { + if (n != 0) + STPNCPY (dest, src, MIN (srclen + 1, n)); + + return srclen; + } + + rulesets = (const unsigned char *) + current->values[_NL_ITEM_INDEX (_NL_COLLATE_RULESETS)].string; + table = (const int32_t *) + 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 (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string; + use_malloc = 0; + + assert (((uintptr_t) table) % __alignof__ (table[0]) == 0); + assert (((uintptr_t) weights) % __alignof__ (weights[0]) == 0); + assert (((uintptr_t) extra) % __alignof__ (extra[0]) == 0); + assert (((uintptr_t) indirect) % __alignof__ (indirect[0]) == 0); + + /* Handle an empty string as a special case. */ + if (srclen == 0) + { + if (n != 0) + *dest = L('\0'); + return 0; + } + + /* We need the elements of the string as unsigned values since they + are used as indeces. */ + 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 + be as fast as possible we are using `alloca' to store the temporary + values. But since there is no limit on the length of the string + we have to use `malloc' if the string is too long. We should be + very conservative here. */ + if (! __libc_use_alloca (srclen)) + { + idxarr = (int32_t *) malloc ((srclen + 1) * (sizeof (int32_t) + 1)); + rulearr = (unsigned char *) &idxarr[srclen]; + + if (idxarr == NULL) + /* No memory. Well, go with the stack then. + + XXX Once this implementation is stable we will handle this + differently. Instead of precomputing the indeces we will + do this in time. This means, though, that this happens for + every pass again. */ + goto try_stack; + use_malloc = 1; + } + else + { + try_stack: + idxarr = (int32_t *) alloca (srclen * sizeof (int32_t)); + rulearr = (unsigned char *) alloca (srclen + 1); + } + + idxmax = 0; + do + { + int32_t tmp = findidx (&usrc); + rulearr[idxmax] = tmp >> 24; + idxarr[idxmax] = tmp & 0xffffff; + + ++idxmax; + } + while (*usrc != L('\0')); + + /* This element is only read, the value never used but to determine + another value which then is ignored. */ + rulearr[idxmax] = '\0'; + + /* Now the passes over the weights. We now use the indeces we found + before. */ + needed = 0; + for (pass = 0; pass < nrules; ++pass) + { + size_t backw_stop = ~0ul; + int rule = rulesets[rulearr[0] * nrules + pass]; + /* We assume that if a rule has defined `position' in one section + this is true for all of them. */ + int position = rule & sort_position; + + if (position == 0) + { + for (idxcnt = 0; idxcnt < idxmax; ++idxcnt) + { + if ((rule & sort_forward) != 0) + { + size_t len; + + if (backw_stop != ~0ul) + { + /* Handle the pushed elements now. */ + size_t backw; + + for (backw = idxcnt - 1; backw >= backw_stop; --backw) + { + len = weights[idxarr[backw]++]; + + if (needed + len < n) + while (len-- > 0) + dest[needed++] = weights[idxarr[backw]++]; + else + { + /* No more characters fit into the buffer. */ + needed += len; + idxarr[backw] += len; + } + } + + backw_stop = ~0ul; + } + + /* Now handle the forward element. */ + len = weights[idxarr[idxcnt]++]; + if (needed + len < n) + while (len-- > 0) + dest[needed++] = weights[idxarr[idxcnt]++]; + else + { + /* No more characters fit into the buffer. */ + needed += len; + idxarr[idxcnt] += len; + } + } + else + { + /* Remember where the backwards series started. */ + if (backw_stop == ~0ul) + backw_stop = idxcnt; + } + + rule = rulesets[rulearr[idxcnt + 1] * nrules + pass]; + } + + + if (backw_stop != ~0ul) + { + /* Handle the pushed elements now. */ + size_t backw; + + backw = idxcnt; + while (backw > backw_stop) + { + size_t len = weights[idxarr[--backw]++]; + + if (needed + len < n) + while (len-- > 0) + dest[needed++] = weights[idxarr[backw]++]; + else + { + /* No more characters fit into the buffer. */ + needed += len; + idxarr[backw] += len; + } + } + } + } + else + { + int val = 1; +#ifndef WIDE_CHAR_VERSION + char buf[7]; + size_t buflen; +#endif + size_t i; + + for (idxcnt = 0; idxcnt < idxmax; ++idxcnt) + { + if ((rule & sort_forward) != 0) + { + size_t len; + + if (backw_stop != ~0ul) + { + /* Handle the pushed elements now. */ + size_t backw; + + for (backw = idxcnt - 1; backw >= backw_stop; --backw) + { + 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) + { + for (i = 0; i < buflen; ++i) + dest[needed + i] = buf[i]; + for (i = 0; i < len; ++i) + dest[needed + buflen + i] = + weights[idxarr[backw] + i]; + } + needed += buflen + len; +#endif + idxarr[backw] += len; + val = 1; + } + else + ++val; + } + + backw_stop = ~0ul; + } + + /* Now handle the forward element. */ + 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) + { + for (i = 0; i < buflen; ++i) + dest[needed + i] = buf[i]; + for (i = 0; i < len; ++i) + dest[needed + buflen + i] = + weights[idxarr[idxcnt] + i]; + } + needed += buflen + len; +#endif + idxarr[idxcnt] += len; + val = 1; + } + else + /* Note that we don't have to increment `idxarr[idxcnt]' + since the length is zero. */ + ++val; + } + else + { + /* Remember where the backwards series started. */ + if (backw_stop == ~0ul) + backw_stop = idxcnt; + } + + rule = rulesets[rulearr[idxcnt + 1] * nrules + pass]; + } + + if (backw_stop != ~0ul) + { + /* Handle the pushed elements now. */ + size_t backw; + + backw = idxmax - 1; + while (backw > backw_stop) + { + 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) + { + for (i = 0; i < buflen; ++i) + dest[needed + i] = buf[i]; + for (i = 0; i < len; ++i) + dest[needed + buflen + i] = + weights[idxarr[backw] + i]; + } + needed += buflen + len; +#endif + idxarr[backw] += len; + val = 1; + } + else + ++val; + } + } + } + + /* Finally store the byte to separate the passes or terminate + the string. */ + if (needed < n) + dest[needed] = pass + 1 < nrules ? L('\1') : L('\0'); + ++needed; + } + + /* This is a little optimization: many collation specifications have + 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] == L('\1')) + { + /* Remove the \1 byte. */ + --needed; + dest[needed - 1] = L('\0'); + } + + /* Free the memory if needed. */ + if (use_malloc) + free (idxarr); + + /* Return the number of bytes/words we need, but don't count the NUL + byte/word at the end. */ + return needed - 1; +} +libc_hidden_def (STRXFRM) + +#ifndef WIDE_CHAR_VERSION weak_alias (__strxfrm_l, strxfrm_l) +#endif |