From f1813b562b7d4aebfde07f0991126e2de7a55d73 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Sat, 6 Jan 2001 20:21:33 +0000 Subject: Update. 2001-01-06 Ulrich Drepper * iconv/skeleton.c: Reset outbuf for next round of the loop. Reported by Owen Taylor . * iconv/Makefile (tests): Add tst-iconv3. * iconv/tst-iconv3.c: New file. * iconvdata/ibm930.c: Fix handling of state. Optimize a bit. * iconvdata/ibm933.c: Likewise. * iconvdata/ibm935.c: Likewise. * iconvdata/ibm937.c: Likewise. * iconvdata/ibm939.c: Likewise. * iconvdata/ibm930.h: Adjust single byte table for optimization. * iconvdata/ibm933.h: Likewise. * iconvdata/ibm935.h: Likewise. * iconvdata/ibm939.h: Likewise. * iconvdata/testdata/IBM930: Add misssing SI. * iconvdata/testdata/IBM933: Likewise. * iconvdata/testdata/IBM935: Likewise. * iconvdata/testdata/IBM937: Likewise. * iconvdata/testdata/IBM939: Likewise. * configure.in: Check for old add-ons that shouldn't be used with current glibc anymore. --- iconvdata/ibm935.c | 131 +++++++++++++++++++++++++++++++++-------------------- 1 file changed, 82 insertions(+), 49 deletions(-) (limited to 'iconvdata/ibm935.c') diff --git a/iconvdata/ibm935.c b/iconvdata/ibm935.c index f63398d9ea..04b8e9dd38 100644 --- a/iconvdata/ibm935.c +++ b/iconvdata/ibm935.c @@ -1,5 +1,5 @@ /* Conversion from and to IBM935 - Copyright (C) 2000 Free Software Foundation, Inc. + Copyright (C) 2000, 2001 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Masahide Washizawa , 2000. @@ -32,42 +32,76 @@ #define CHARSET_NAME "IBM935//" #define FROM_LOOP from_ibm935 #define TO_LOOP to_ibm935 +#define MIN_NEEDED_FROM 1 +#define MAX_NEEDED_FROM 2 +#define MIN_NEEDED_TO 4 +#define MAX_NEEDED_TO 4 +#define PREPARE_LOOP \ + int save_curcs; \ + int *curcsp = &data->__statep->__count; +#define EXTRA_LOOP_ARGS , curcsp /* Definitions of initialization and destructor function. */ #define DEFINE_INIT 1 #define DEFINE_FINI 1 -#define MIN_NEEDED_FROM 1 -#define MIN_NEEDED_TO 4 + +/* Since this is a stateful encoding we have to provide code which resets + the output state to the initial state. This has to be done during the + flushing. */ +#define EMIT_SHIFT_TO_INIT \ + if ((data->__statep->__count & ~7) != sb) \ + { \ + if (FROM_DIRECTION) \ + data->__statep->__count &= 7; \ + else \ + { \ + unsigned char *outbuf = data->__outbuf; \ + \ + /* We are not in the initial state. To switch back we have \ + to emit `SI'. */ \ + if (__builtin_expect (outbuf >= data->__outbufend, 0)) \ + /* We don't have enough room in the output buffer. */ \ + status = __GCONV_FULL_OUTPUT; \ + else \ + { \ + /* Write out the shift sequence. */ \ + *outbuf++ = SI; \ + data->__outbuf = outbuf; \ + data->__statep->__count &= 7; \ + } \ + } \ + } + + +/* Since we might have to reset input pointer we must be able to save + and retore the state. */ +#define SAVE_RESET_STATE(Save) \ + if (Save) \ + save_curcs = *curcsp; \ + else \ + *curcsp = save_curcs + /* Current codeset type. */ enum { - init = 0, - sb, - db + sb = 0, + db = 64 }; /* First, define the conversion function from IBM-935 to UCS4. */ #define MIN_NEEDED_INPUT MIN_NEEDED_FROM +#define MAX_NEEDED_INPUT MAX_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO -#define INIT_PARAMS int curcs = init; #define LOOPFCT FROM_LOOP #define BODY \ { \ uint32_t ch = *inptr; \ uint32_t res; \ - const struct gap *rp1 = __ibm935sb_to_ucs4_idx; \ - const struct gap *rp2 = __ibm935db_to_ucs4_idx; \ \ if (__builtin_expect(ch, 0) == SO) \ { \ - if (__builtin_expect (inptr + 1 >= inend, 0)) \ - { \ - result = __GCONV_INCOMPLETE_INPUT; \ - break; \ - } \ - \ /* Shift OUT, change to DBCS converter. */ \ if (curcs == db) \ { \ @@ -76,16 +110,10 @@ enum } \ curcs = db; \ ++inptr; \ - ch = *inptr; \ + continue; \ } \ else if (__builtin_expect (ch, 0) == SI) \ { \ - if (__builtin_expect (inptr + 1 >= inend, 0)) \ - { \ - result = __GCONV_INCOMPLETE_INPUT; \ - break; \ - } \ - \ /* Shift IN, change to SBCS converter. */ \ if (curcs == sb) \ { \ @@ -94,19 +122,14 @@ enum } \ curcs = sb; \ ++inptr; \ - ch = *inptr; \ + continue; \ } \ \ - if (curcs == sb || curcs == init) \ + if (curcs == sb) \ { \ /* Use the IBM935 table for single byte. */ \ - while (ch > rp1->end) \ - ++rp1; \ - \ - if (__builtin_expect (rp1 == NULL, 0) \ - || __builtin_expect (ch < rp1->start, 0) \ - || (res = __ibm935sb_to_ucs4[ch + rp1->idx], \ - __builtin_expect (res, L'\1') == L'\0' && ch != '\0')) \ + res = __ibm935sb_to_ucs4[ch]; \ + if (__builtin_expect (res, L'\1') == L'\0' && ch != '\0') \ { \ /* This is an illegal character. */ \ if (! ignore_errors_p ()) \ @@ -115,18 +138,20 @@ enum break; \ } \ ++*irreversible; \ - ++inptr; \ - continue; \ } \ else \ { \ put32 (outptr, res); \ outptr += 4; \ - inptr++; \ } \ + ++inptr; \ } \ - else if (curcs == db) \ + else \ { \ + const struct gap *rp2 = __ibm935db_to_ucs4_idx; \ + \ + assert (curcs == db); \ + \ /* Use the IBM935 table for double byte. */ \ if (__builtin_expect (inptr + 1 >= inend, 0)) \ { \ @@ -152,24 +177,25 @@ enum break; \ } \ ++*irreversible; \ - inptr += 2; \ - continue; \ } \ else \ { \ put32 (outptr, res); \ outptr += 4; \ - inptr += 2; \ } \ + inptr += 2; \ } \ } #define LOOP_NEED_FLAGS +#define EXTRA_LOOP_DECLS , int *curcsp +#define INIT_PARAMS int curcs = *curcsp & ~7 +#define UPDATE_PARAMS *curcsp = curcs #include /* Next, define the other direction. */ #define MIN_NEEDED_INPUT MIN_NEEDED_TO #define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM -#define INIT_PARAMS int curcs = init; +#define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM #define LOOPFCT TO_LOOP #define BODY \ { \ @@ -180,16 +206,21 @@ enum \ if (__builtin_expect (ch, 0) >= 0xffff) \ { \ - rp1 = NULL; \ - rp2 = NULL; \ + if (! ignore_errors_p ()) \ + { \ + result = __GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + ++*irreversible; \ + inptr += 4; \ + continue; \ } \ - else \ - while (ch > rp1->end) \ - ++rp1; \ + \ + while (ch > rp1->end) \ + ++rp1; \ \ /* Use the UCS4 table for single byte. */ \ - if (__builtin_expect (rp1 == NULL, 0) \ - || __builtin_expect (ch < rp1->start, 0) \ + if (__builtin_expect (ch < rp1->start, 0) \ || (cp = __ucs4_to_ibm935sb[ch + rp1->idx], \ __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \ { \ @@ -197,8 +228,7 @@ enum while (ch > rp2->end) \ ++rp2; \ \ - if (__builtin_expect (rp2 == NULL, 0) \ - || __builtin_expect (ch < rp2->start, 0) \ + if (__builtin_expect (ch < rp2->start, 0) \ || (cp = __ucs4_to_ibm935db[ch + rp2->idx], \ __builtin_expect (cp[0], L'\1')==L'\0' && ch != '\0')) \ { \ @@ -212,7 +242,7 @@ enum } \ else \ { \ - if (curcs == init || curcs == sb) \ + if (curcs == sb) \ { \ if (__builtin_expect (outptr+1 > outend, 0)) \ { \ @@ -257,6 +287,9 @@ enum inptr += 4; \ } #define LOOP_NEED_FLAGS +#define EXTRA_LOOP_DECLS , int *curcsp +#define INIT_PARAMS int curcs = *curcsp & ~7 +#define UPDATE_PARAMS *curcsp = curcs #include /* Now define the toplevel functions. */ -- cgit v1.2.3