From e2ac4098ed507957ba6dac9b408d31054a8e6b6d Mon Sep 17 00:00:00 2001 From: Alexander Barkov Date: Thu, 19 Jul 2018 13:02:14 +0400 Subject: Simplify caseup() and casedn() in charsets After the MDEV-13118 fix there's no code in the server that wants caseup/casedn to change the argument in place for simple charsets. Let's remove this logic and always return the result in a new string for all charsets, both simple and complex. 1. Removing the optimization that *some* character sets used in casedn() and caseup(), which allowed (and required) to change the case in-place, overwriting the string passed as the "src" argument. Now all CHARSET_INFO's work in the same way: non of them change the source string in-place, all of them now convert case from the source string to the destination string, leaving the source string untouched. 2. Adding "const" qualifier to the "char *src" parameter to caseup() and casedn(). 3. Removing duplicate implementations in ctype-mb.c. Now both caseup() and casedn() implementations for all CJK character sets use internally the same function my_casefold_mb() (the former my_casefold_mb_varlen()). 4. Removing the "unused" attribute from parameters of some my_case{up|dn}_xxx() implementations, as the affected parameters are now *used* in the code. Previously these parameters were used only in DBUG_ASSERT(). --- strings/ctype-bin.c | 8 ++-- strings/ctype-euc_kr.c | 4 +- strings/ctype-mb.c | 100 +++++++------------------------------------------ strings/ctype-simple.c | 22 +++++------ strings/ctype-ucs2.c | 78 ++++++++++++++++++++------------------ strings/ctype-ujis.c | 8 ++-- strings/ctype-utf8.c | 20 ++++++---- 7 files changed, 88 insertions(+), 152 deletions(-) (limited to 'strings') diff --git a/strings/ctype-bin.c b/strings/ctype-bin.c index 2e699db0bd3..53a0e850b54 100644 --- a/strings/ctype-bin.c +++ b/strings/ctype-bin.c @@ -220,11 +220,11 @@ static size_t my_case_str_bin(CHARSET_INFO *cs __attribute__((unused)), static size_t my_case_bin(CHARSET_INFO *cs __attribute__((unused)), - char *src __attribute__((unused)), - size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) + const char *src, size_t srclen, + char *dst, size_t dstlen) { + DBUG_ASSERT(srclen <= dstlen); + memcpy(dst, src, srclen); return srclen; } diff --git a/strings/ctype-euc_kr.c b/strings/ctype-euc_kr.c index b7065369258..f35c3f07c79 100644 --- a/strings/ctype-euc_kr.c +++ b/strings/ctype-euc_kr.c @@ -9994,8 +9994,8 @@ static MY_CHARSET_HANDLER my_charset_handler= my_mb_ctype_mb, my_caseup_str_mb, my_casedn_str_mb, - my_caseup_mb_varlen, /* UPPER() can reduce length: Turkish DOTLESS i -> I */ - my_casedn_mb, /* LOWER() does not change length, use simple version*/ + my_caseup_mb, /* UPPER() can reduce length: Turkish DOTLESS i -> I */ + my_casedn_mb, /* LOWER() does not change length */ my_snprintf_8bit, my_long10_to_str_8bit, my_longlong10_to_str_8bit, diff --git a/strings/ctype-mb.c b/strings/ctype-mb.c index 02a9a91ca6a..318e581792a 100644 --- a/strings/ctype-mb.c +++ b/strings/ctype-mb.c @@ -71,81 +71,8 @@ get_case_info_for_ch(const CHARSET_INFO *cs, uint page, uint offs) /* - For character sets which don't change octet length in case conversion. -*/ -size_t my_caseup_mb(CHARSET_INFO * cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) -{ - register uint32 l; - register char *srcend= src + srclen; - register const uchar *map= cs->to_upper; - - DBUG_ASSERT(cs->caseup_multiply == 1); - DBUG_ASSERT(src == dst && srclen == dstlen); - DBUG_ASSERT(cs->mbmaxlen == 2); - - while (src < srcend) - { - if ((l=my_ismbchar(cs, src, srcend))) - { - MY_UNICASE_CHARACTER *ch; - if ((ch= get_case_info_for_ch(cs, (uchar) src[0], (uchar) src[1]))) - { - *src++= ch->toupper >> 8; - *src++= ch->toupper & 0xFF; - } - else - src+= l; - } - else - { - *src=(char) map[(uchar) *src]; - src++; - } - } - return srclen; -} - - -size_t my_casedn_mb(CHARSET_INFO * cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) -{ - register uint32 l; - register char *srcend= src + srclen; - register const uchar *map=cs->to_lower; - - DBUG_ASSERT(cs->casedn_multiply == 1); - DBUG_ASSERT(src == dst && srclen == dstlen); - DBUG_ASSERT(cs->mbmaxlen == 2); - - while (src < srcend) - { - if ((l= my_ismbchar(cs, src, srcend))) - { - MY_UNICASE_CHARACTER *ch; - if ((ch= get_case_info_for_ch(cs, (uchar) src[0], (uchar) src[1]))) - { - *src++= ch->tolower >> 8; - *src++= ch->tolower & 0xFF; - } - else - src+= l; - } - else - { - *src= (char) map[(uchar)*src]; - src++; - } - } - return srclen; -} - - -/* - Case folding functions for character set - where case conversion can change string octet length. + Case folding functions for CJK character set. + Case conversion can optionally reduce string octet length. For example, in EUCKR, _euckr 0xA9A5 == "LATIN LETTER DOTLESS I" (Turkish letter) is upper-cased to to @@ -153,13 +80,14 @@ size_t my_casedn_mb(CHARSET_INFO * cs, char *src, size_t srclen, Length is reduced in this example from two bytes to one byte. */ static size_t -my_casefold_mb_varlen(CHARSET_INFO *cs, - char *src, size_t srclen, - char *dst, size_t dstlen __attribute__((unused)), - const uchar *map, - size_t is_upper) +my_casefold_mb(CHARSET_INFO *cs, + const char *src, size_t srclen, + char *dst, size_t dstlen __attribute__((unused)), + const uchar *map, + size_t is_upper) { - char *srcend= src + srclen, *dst0= dst; + const char *srcend= src + srclen; + char *dst0= dst; DBUG_ASSERT(cs->mbmaxlen == 2); @@ -193,22 +121,22 @@ my_casefold_mb_varlen(CHARSET_INFO *cs, size_t -my_casedn_mb_varlen(CHARSET_INFO * cs, char *src, size_t srclen, +my_casedn_mb(CHARSET_INFO * cs, const char *src, size_t srclen, char *dst, size_t dstlen) { DBUG_ASSERT(dstlen >= srclen * cs->casedn_multiply); DBUG_ASSERT(src != dst || cs->casedn_multiply == 1); - return my_casefold_mb_varlen(cs, src, srclen, dst, dstlen, cs->to_lower, 0); + return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_lower, 0); } size_t -my_caseup_mb_varlen(CHARSET_INFO * cs, char *src, size_t srclen, - char *dst, size_t dstlen) +my_caseup_mb(CHARSET_INFO * cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { DBUG_ASSERT(dstlen >= srclen * cs->caseup_multiply); DBUG_ASSERT(src != dst || cs->caseup_multiply == 1); - return my_casefold_mb_varlen(cs, src, srclen, dst, dstlen, cs->to_upper, 1); + return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_upper, 1); } diff --git a/strings/ctype-simple.c b/strings/ctype-simple.c index 7f13cef4474..61b14b84820 100644 --- a/strings/ctype-simple.c +++ b/strings/ctype-simple.c @@ -214,28 +214,26 @@ size_t my_casedn_str_8bit(CHARSET_INFO * cs,char *str) } -size_t my_caseup_8bit(CHARSET_INFO * cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +size_t my_caseup_8bit(CHARSET_INFO * cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { - char *end= src + srclen; + const char *end= src + srclen; register const uchar *map= cs->to_upper; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); for ( ; src != end ; src++) - *src= (char) map[(uchar) *src]; + *dst++= (char) map[(uchar) *src]; return srclen; } -size_t my_casedn_8bit(CHARSET_INFO * cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +size_t my_casedn_8bit(CHARSET_INFO * cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { - char *end= src + srclen; + const char *end= src + srclen; register const uchar *map=cs->to_lower; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); for ( ; src != end ; src++) - *src= (char) map[(uchar) *src]; + *dst++= (char) map[(uchar) *src]; return srclen; } diff --git a/strings/ctype-ucs2.c b/strings/ctype-ucs2.c index 318c248f742..724e69872de 100644 --- a/strings/ctype-ucs2.c +++ b/strings/ctype-ucs2.c @@ -1196,25 +1196,26 @@ my_tosort_utf16(MY_UNICASE_INFO *uni_plane, my_wc_t *wc) static size_t -my_caseup_utf16(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +my_caseup_utf16(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; my_charset_conv_mb_wc mb_wc= cs->cset->mb_wc; my_charset_conv_wc_mb wc_mb= cs->cset->wc_mb; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((src < srcend) && (res= mb_wc(cs, &wc, (uchar *) src, (uchar *) srcend)) > 0) { my_toupper_utf16(uni_plane, &wc); - if (res != wc_mb(cs, wc, (uchar *) src, (uchar *) srcend)) + if (res != wc_mb(cs, wc, (uchar *) dst, (uchar *) dstend)) break; src+= res; + dst+= res; } return srclen; } @@ -1243,25 +1244,26 @@ my_hash_sort_utf16(CHARSET_INFO *cs, const uchar *s, size_t slen, static size_t -my_casedn_utf16(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +my_casedn_utf16(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; my_charset_conv_mb_wc mb_wc= cs->cset->mb_wc; my_charset_conv_wc_mb wc_mb= cs->cset->wc_mb; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((src < srcend) && (res= mb_wc(cs, &wc, (uchar *) src, (uchar *) srcend)) > 0) { my_tolower_utf16(uni_plane, &wc); - if (res != wc_mb(cs, wc, (uchar *) src, (uchar *) srcend)) + if (res != wc_mb(cs, wc, (uchar *) dst, (uchar *) dstend)) break; src+= res; + dst+= res; } return srclen; } @@ -1987,23 +1989,24 @@ my_tosort_utf32(MY_UNICASE_INFO *uni_plane, my_wc_t *wc) static size_t -my_caseup_utf32(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +my_caseup_utf32(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((src < srcend) && (res= my_utf32_uni(cs, &wc, (uchar *)src, (uchar*) srcend)) > 0) { my_toupper_utf32(uni_plane, &wc); - if (res != my_uni_utf32(cs, wc, (uchar*) src, (uchar*) srcend)) + if (res != my_uni_utf32(cs, wc, (uchar*) dst, (uchar*) dstend)) break; src+= res; + dst+= res; } return srclen; } @@ -2038,22 +2041,23 @@ my_hash_sort_utf32(CHARSET_INFO *cs, const uchar *s, size_t slen, static size_t -my_casedn_utf32(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +my_casedn_utf32(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((res= my_utf32_uni(cs, &wc, (uchar*) src, (uchar*) srcend)) > 0) { my_tolower_utf32(uni_plane,&wc); - if (res != my_uni_utf32(cs, wc, (uchar*) src, (uchar*) srcend)) + if (res != my_uni_utf32(cs, wc, (uchar*) dst, (uchar*) dstend)) break; src+= res; + dst+= res; } return srclen; } @@ -2950,23 +2954,24 @@ my_tosort_ucs2(MY_UNICASE_INFO *uni_plane, my_wc_t *wc) *wc= page[*wc & 0xFF].sort; } -static size_t my_caseup_ucs2(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +static size_t my_caseup_ucs2(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((src < srcend) && (res= my_ucs2_uni(cs, &wc, (uchar *)src, (uchar*) srcend)) > 0) { my_toupper_ucs2(uni_plane, &wc); - if (res != my_uni_ucs2(cs, wc, (uchar*) src, (uchar*) srcend)) + if (res != my_uni_ucs2(cs, wc, (uchar*) dst, (uchar*) dstend)) break; src+= res; + dst+= res; } return srclen; } @@ -2995,23 +3000,24 @@ static void my_hash_sort_ucs2(CHARSET_INFO *cs, const uchar *s, size_t slen, } -static size_t my_casedn_ucs2(CHARSET_INFO *cs, char *src, size_t srclen, - char *dst __attribute__((unused)), - size_t dstlen __attribute__((unused))) +static size_t my_casedn_ucs2(CHARSET_INFO *cs, const char *src, size_t srclen, + char *dst, size_t dstlen) { my_wc_t wc; int res; - char *srcend= src + srclen; + const char *srcend= src + srclen; + char *dstend= dst + dstlen; MY_UNICASE_INFO *uni_plane= cs->caseinfo; - DBUG_ASSERT(src == dst && srclen == dstlen); + DBUG_ASSERT(srclen <= dstlen); while ((src < srcend) && (res= my_ucs2_uni(cs, &wc, (uchar*) src, (uchar*) srcend)) > 0) { my_tolower_ucs2(uni_plane, &wc); - if (res != my_uni_ucs2(cs, wc, (uchar*) src, (uchar*) srcend)) + if (res != my_uni_ucs2(cs, wc, (uchar*) dst, (uchar*) dstend)) break; src+= res; + dst+= res; } return srclen; } diff --git a/strings/ctype-ujis.c b/strings/ctype-ujis.c index e7dbefe6c1d..25f8881d819 100644 --- a/strings/ctype-ujis.c +++ b/strings/ctype-ujis.c @@ -67179,12 +67179,12 @@ get_case_info_for_ch(CHARSET_INFO *cs, uint plane, uint page, uint offs) */ static size_t my_casefold_ujis(CHARSET_INFO *cs, - char *src, size_t srclen, + const char *src, size_t srclen, char *dst, size_t dstlen __attribute__((unused)), const uchar * const map, size_t is_upper) { - char *srcend= src + srclen, *dst0= dst; + const char *srcend= src + srclen, *dst0= dst; while (src < srcend) { @@ -67226,7 +67226,7 @@ my_casefold_ujis(CHARSET_INFO *cs, LOWER() */ size_t -my_casedn_ujis(CHARSET_INFO * cs, char *src, size_t srclen, +my_casedn_ujis(CHARSET_INFO * cs, const char *src, size_t srclen, char *dst, size_t dstlen) { DBUG_ASSERT(dstlen >= srclen * cs->casedn_multiply); @@ -67239,7 +67239,7 @@ my_casedn_ujis(CHARSET_INFO * cs, char *src, size_t srclen, UPPER() */ size_t -my_caseup_ujis(CHARSET_INFO * cs, char *src, size_t srclen, +my_caseup_ujis(CHARSET_INFO * cs, const char *src, size_t srclen, char *dst, size_t dstlen) { DBUG_ASSERT(dstlen >= srclen * cs->caseup_multiply); diff --git a/strings/ctype-utf8.c b/strings/ctype-utf8.c index cd4e82a107c..7d7e61ce511 100644 --- a/strings/ctype-utf8.c +++ b/strings/ctype-utf8.c @@ -5065,12 +5065,13 @@ my_tosort_utf8mb3(MY_UNICASE_INFO *uni_plane, my_wc_t *wc) *wc= page[*wc & 0xFF].sort; } -static size_t my_caseup_utf8(CHARSET_INFO *cs, char *src, size_t srclen, +static size_t my_caseup_utf8(CHARSET_INFO *cs, const char *src, size_t srclen, char *dst, size_t dstlen) { my_wc_t wc; int srcres, dstres; - char *srcend= src + srclen, *dstend= dst + dstlen, *dst0= dst; + const char *srcend= src + srclen; + char *dstend= dst + dstlen, *dst0= dst; MY_UNICASE_INFO *uni_plane= cs->caseinfo; DBUG_ASSERT(src != dst || cs->caseup_multiply == 1); @@ -5136,12 +5137,13 @@ static size_t my_caseup_str_utf8(CHARSET_INFO *cs, char *src) } -static size_t my_casedn_utf8(CHARSET_INFO *cs, char *src, size_t srclen, +static size_t my_casedn_utf8(CHARSET_INFO *cs, const char *src, size_t srclen, char *dst, size_t dstlen) { my_wc_t wc; int srcres, dstres; - char *srcend= src + srclen, *dstend= dst + dstlen, *dst0= dst; + const char *srcend= src + srclen; + char *dstend= dst + dstlen, *dst0= dst; MY_UNICASE_INFO *uni_plane= cs->caseinfo; DBUG_ASSERT(src != dst || cs->casedn_multiply == 1); @@ -7579,12 +7581,13 @@ my_toupper_utf8mb4(MY_UNICASE_INFO *uni_plane, my_wc_t *wc) static size_t -my_caseup_utf8mb4(CHARSET_INFO *cs, char *src, size_t srclen, +my_caseup_utf8mb4(CHARSET_INFO *cs, const char *src, size_t srclen, char *dst, size_t dstlen) { my_wc_t wc; int srcres, dstres; - char *srcend= src + srclen, *dstend= dst + dstlen, *dst0= dst; + const char *srcend= src + srclen; + char *dstend= dst + dstlen, *dst0= dst; MY_UNICASE_INFO *uni_plane= cs->caseinfo; DBUG_ASSERT(src != dst || cs->caseup_multiply == 1); @@ -7666,12 +7669,13 @@ my_caseup_str_utf8mb4(CHARSET_INFO *cs, char *src) static size_t my_casedn_utf8mb4(CHARSET_INFO *cs, - char *src, size_t srclen, + const char *src, size_t srclen, char *dst, size_t dstlen) { my_wc_t wc; int srcres, dstres; - char *srcend= src + srclen, *dstend= dst + dstlen, *dst0= dst; + const char *srcend= src + srclen; + char *dstend= dst + dstlen, *dst0= dst; MY_UNICASE_INFO *uni_plane= cs->caseinfo; DBUG_ASSERT(src != dst || cs->casedn_multiply == 1); -- cgit v1.2.1