From c386ddb4d8a1b076d94ebe8b85ca5d0dd124892b Mon Sep 17 00:00:00 2001 From: Manuel Novoa III Date: Fri, 22 Nov 2002 03:05:27 +0000 Subject: Ok... here's the summary: Hopefully locale support will build when cross compiling now. Collation is still not supported, but that's what I'm currently working on. In the next couple of days, I'll probably put up a couple of files for download that will save people the trouble of generating all the glibc locales. Added *wprintf functions, although they currently don't support floating point. That will be fixed when I rewrite _dtostr... or possibly before. Added the wcsto{inttype} functions. Added iconv() and a mini iconv utility. The require locale support and only provide for conversions involving the various unicode encodings { UCS-4*, UCS-2*, UTF-32*, UTF-16*, UTF-8 }, the 8-bit codesets built with the locale data, and the internal WCHAR_T. --- Makefile | 6 + TODO | 24 +- docs/Glibc_vs_uClibc_Differences.txt | 8 +- extra/locale/Makefile | 50 +- extra/locale/README | 2 +- extra/locale/gen_ldc.c | 460 ++++++++ extra/locale/gen_locale.c | 4 +- extra/locale/gen_mmap.c | 6 +- extra/locale/gen_wc8bit.c | 36 +- extra/locale/gen_wctype.c | 76 +- extra/locale/locale_mmap.h | 4 +- include/iconv.h | 54 + ldso/util/Makefile | 2 +- libc/misc/locale/locale.c | 4 +- libc/misc/time/time.c | 7 + libc/misc/wchar/Makefile | 15 +- libc/misc/wchar/wchar.c | 658 ++++++++++- libc/misc/wchar/wstdio.c | 21 +- libc/stdio/Makefile | 10 +- libc/stdio/printf.c | 1382 +++++++++++++++--------- libc/stdio/stdio.c | 58 + libc/stdlib/Makefile | 11 +- libc/stdlib/stdlib.c | 215 +++- libc/sysdeps/linux/common/bits/uClibc_locale.h | 2 +- libc/sysdeps/linux/common/bits/uClibc_stdio.h | 5 + 25 files changed, 2423 insertions(+), 697 deletions(-) create mode 100644 extra/locale/gen_ldc.c create mode 100644 include/iconv.h diff --git a/Makefile b/Makefile index 822b0192f..19dcf90ab 100644 --- a/Makefile +++ b/Makefile @@ -274,6 +274,11 @@ ifeq ($(strip $(HAVE_SHARED)),y) install -m 755 ldso/util/ldconfig.target $(PREFIX)$(TARGET_PREFIX)/sbin/ldconfig; \ fi; endif +ifeq ($(strip $(UCLIBC_HAS_LOCALE)),y) + @$(MAKE) -C libc/misc/wchar iconv.target + install -d $(PREFIX)$(TARGET_PREFIX)/usr/bin; + install -m 755 libc/misc/wchar/iconv.target $(PREFIX)$(TARGET_PREFIX)/usr/bin/iconv +endif finished2: @echo @@ -339,6 +344,7 @@ clean: $(MAKE) -C test clean $(MAKE) -C ldso clean $(MAKE) -C libc/misc/internals clean + $(MAKE) -C libc/misc/wchar clean $(MAKE) -C libc/unistd clean $(MAKE) -C libc/sysdeps/linux/common clean $(MAKE) -C extra clean diff --git a/TODO b/TODO index e93a843c0..e6dfb79b6 100644 --- a/TODO +++ b/TODO @@ -18,19 +18,16 @@ Erik's TODO list for the next release: ----------------------------------------------------------------------------- Manuel's todo: -In the next couple of months, I plan on completing the wide char and locale -support. Here's my current list of things to implement by early November. +I'm currently working on completing the wide char and locale support. +Right now i'm looking at collation. Other tasks on my list include: 1) Little things that need fixing: ---------------------------------- a) Use locale decimal point in *printf() and strto*d() -- slightly complicated by the fact that at least one locale uses a wchar radix that does not map to a single byte in UTF-8. - b) Use locale digit grouping in *printf(). - c) Implement %lc (%C) and %ls (%S) handling for printf. - d) Deal with mb format string issues in printf, scanf, - strftime (at least). - e) Implement wcwidth() and wcswidth() -- use Markus Kuhn's versions. + b) Use locale digit grouping in *printf() flosting point. + c) Deal with mb format string issues in scanf and strftime (at least). 2) Implement wide char numeric conversion functions -- wcsto*(). ---------------------------------------------------------------- @@ -38,15 +35,6 @@ support. Here's my current list of things to implement by early November. strto*() functions, unless locale specific alternate digit support is desired. - 3) Implement wide formatted output -- *wprintf(): - ------------------------------------------------- - The stream output variants should be easily created by appropriately - modifying the char *printf() functions. - - The string buffer variants may take some time to implement because - they need to store wchars in a buffer. It will be necessary to hack - the core stdio code slightly to support this. - 4) Reimplement scanf for narrow and wide streams. ------------------------------------------------- The current char version of scanf() needs some cleanup. Also, @@ -54,9 +42,6 @@ support. Here's my current list of things to implement by early November. will require reworking the implementation of matching char sets (enclosed in []). - Also, as with *wprintf() above, the string buffer variants of wscanf() - may require some additional work to the core stdio code. - 5) Additional str{f|p}time issues. ---------------------------------- Support locale specific alternate digits. @@ -71,6 +56,7 @@ support. Here's my current list of things to implement by early November. b) Additional clean up of ctype and wctype. c) Implement collation support for char and wchar -- *xfrm() and *coll(). d) transliteration of unsupported wchars in 8-bit locales (like glibc). + e) Support ISO/IEC 14652 draft locale extensions (LC_PAPER, etc). Other stuff: diff --git a/docs/Glibc_vs_uClibc_Differences.txt b/docs/Glibc_vs_uClibc_Differences.txt index e0e1e6af4..8ebca10a9 100644 --- a/docs/Glibc_vs_uClibc_Differences.txt +++ b/docs/Glibc_vs_uClibc_Differences.txt @@ -100,17 +100,13 @@ wide char support representation of wchar's is assumed to be 31 bit unicode values in native endian representation. Also, the underlying char encoding is assumed to match ASCII in the range 0-0x7f. -2) In the C locale, uClibc's mb<->wc conversion functions map 0x80-CHAR_MAX - onto their wide/narrow equivalents. glibc's conversion functions treat - them as illegal. locale support -------------- 1) The target for support is SUSv3 locale functionality. While nl_langinfo has been extended, similar to glibc, it only returns values for related locale entries. -2) The locale code is not cross-compiler friendly. This should be fixed soon. -3) Currently, collation support is being implemented. +2) Currently, collation support is being implemented. stdio ----- @@ -136,6 +132,8 @@ stdio 8) uClibc's setvbuf is more restrictive about when it can be called than glibc's is. The standards specify that setvbuf must occur before any other operations take place on the stream. +9) Right now, %m is not handled properly by printf when the format uses positional + args. More to follow as I think of it... diff --git a/extra/locale/Makefile b/extra/locale/Makefile index bdad5cea7..3f6e7476e 100644 --- a/extra/locale/Makefile +++ b/extra/locale/Makefile @@ -26,13 +26,15 @@ codesets.txt: fi; gen_wc8bit: gen_wc8bit.c - gcc -Os -Wall $(CFLAGS_wc8bit) gen_wc8bit.c -o gen_wc8bit + $(NATIVE_CC) $(NATIVE_CFLAGS) $(NATIVE_LDFLAGS) $(CFLAGS_wc8bit) $^ -o $@ +#gen_wc8bit.c -o gen_wc8bit c8tables.h: gen_wc8bit codesets.txt ./gen_wc8bit `cat codesets.txt` gen_wctype: gen_wctype.c - gcc -Os -Wall gen_wctype.c -o gen_wctype + $(NATIVE_CC) $(NATIVE_CFLAGS) $(NATIVE_LDFLAGS) $^ -o $@ +#gen_wctype.c -o gen_wctype # Warning! Beware tr_TR toupper/tolower exceptions! # Warning! Ignore the width table. It will go away. @@ -40,7 +42,8 @@ wctables.h: gen_wctype ./gen_wctype en_US gen_locale: gen_locale.c c8tables.h - gcc -Os -Wall gen_locale.c -o gen_locale + $(NATIVE_CC) $(NATIVE_CFLAGS) $(NATIVE_LDFLAGS) gen_locale.c -o $@ +#gen_locale.c -o gen_locale # TODO: if no wide char support, we should auto-disable all UTF-8 locales locale_tables.h: gen_locale locales.txt @@ -64,30 +67,37 @@ locales.txt: false; \ fi; -ifneq ($(TARGET_ARCH),$(NATIVE_ARCH)) +# ifneq ($(TARGET_ARCH),$(NATIVE_ARCH)) -locale.mmap: gen_mmap - echo Your locale.mmap file is either missing or out of date. - echo The developmental code can only generate one for the - echo target arch == native arch case. Sorry. -else +# locale.mmap: gen_mmap +# echo Your locale.mmap file is either missing or out of date. +# echo The developmental code can only generate one for the +# echo target arch == native arch case. Sorry. +# else -gen_mmap: gen_mmap.c c8tables.h wctables.h locale_tables.h - gcc $(CFLAGS_mmap) -Os -Wall -static gen_mmap.c -o gen_mmap +# gen_mmap: gen_mmap.c c8tables.h wctables.h locale_tables.h +# gcc $(CFLAGS_mmap) -Os -Wall -static gen_mmap.c -o gen_mmap -locale.mmap: gen_mmap - ./gen_mmap +# locale.mmap: gen_mmap +# ./gen_mmap -endif +# endif + +# lmmtolso: lmmtolso.c +# gcc -Os -Wall lmmtolso.c -o lmmtolso + +# locale_data.c: lmmtolso locale.mmap +# ./lmmtolso -lmmtolso: lmmtolso.c - gcc -Os -Wall lmmtolso.c -o lmmtolso +gen_ldc: gen_ldc.c c8tables.h wctables.h locale_tables.h + $(NATIVE_CC) $(NATIVE_CFLAGS) $(NATIVE_LDFLAGS) $(CFLAGS_mmap) gen_ldc.c -o $@ +#gen_ldc.c -o gen_ldc -locale_data.c: lmmtolso locale.mmap - ./lmmtolso +locale_data.c: gen_ldc + ./gen_ldc locale_data.o: locale_data.c lt_defines.h - $(CC) -c locale_data.c + $(CC) $(CFLAGS_mmap) -c locale_data.c $(STRIPTOOL) -x -R .note -R .comment locale_data.o uClibc_locale_data.h: c8tables.h wctables.h lt_defines.h locale_mmap.h @@ -100,7 +110,7 @@ links-target: locale_data.o uClibc_locale_data.h clean: rm -f *.[oa] *~ core - rm -f gen_wc8bit gen_wctype gen_locale + rm -f gen_wc8bit gen_wctype gen_locale gen_ldc rm -f c8tables.h wctables.h locale_tables.h lt_defines.h rm -f gen_mmap locale.mmap lmmtolso rm -f locale_data.c locale_data.o uClibc_locale_data.h diff --git a/extra/locale/README b/extra/locale/README index 63723f97f..244b92538 100644 --- a/extra/locale/README +++ b/extra/locale/README @@ -44,7 +44,7 @@ Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! Warning!!! -1) In the toplevel dir, 'make headers uClibc_config' . +1) In the toplevel dir, 'make headers'. 2) Create a codesets.txt file in this dir listing the codesets you want to support. The easiest way to do this is to edit the output of diff --git a/extra/locale/gen_ldc.c b/extra/locale/gen_ldc.c new file mode 100644 index 000000000..17e6c6875 --- /dev/null +++ b/extra/locale/gen_ldc.c @@ -0,0 +1,460 @@ +#include +#include +#include +#include +#include +#include + +#ifndef __WCHAR_ENABLED +#warning WHOA!!! __WCHAR_ENABLED is not defined! defining it now... +#define __WCHAR_ENABLED +#endif + +#define WANT_DATA +#include "c8tables.h" +#ifndef __CTYPE_HAS_8_BIT_LOCALES +#warning __CTYPE_HAS_8_BIT_LOCALES is not defined... +/* #define __CTYPE_HAS_8_BIT_LOCALES */ +#endif + +/* #define Cctype_TBL_LEN 328 */ +/* #define Cuplow_TBL_LEN 400 */ +/* #define Cc2wc_TBL_LEN 1448 */ +/* #define Cwc2c_TBL_LEN 3744 */ + +#define WANT_WCctype_data +#define WANT_WCuplow_data +#define WANT_WCuplow_diff_data +/* #define WANT_WCcomb_data */ +/* #define WANT_WCwidth_data */ +#include "wctables.h" +#undef WANT_WCctype_data +#undef WANT_WCuplow_data +#undef WANT_WCuplow_diff_data +/* #undef WANT_WCcomb_data */ +/* #undef WANT_WCwidth_data */ + + #define WCctype_TBL_LEN (WCctype_II_LEN + WCctype_TI_LEN + WCctype_UT_LEN) + #define WCuplow_TBL_LEN (WCuplow_II_LEN + WCuplow_TI_LEN + WCuplow_UT_LEN) + #define WCuplow_diff_TBL_LEN (2 * WCuplow_diffs) +/* #define WCcomb_TBL_LEN (WCcomb_II_LEN + WCcomb_TI_LEN + WCcomb_UT_LEN) */ + +#include "locale_tables.h" + +#include "locale_mmap.h" + +/* #undef __PASTE2 */ +/* #define __PASTE2(A,B) A ## B */ +/* #undef __PASTE3 */ +/* #define __PASTE3(A,B,C) A ## B ## C */ + + +/* #define MAGIC_SIZE 64 */ + +/* #define COMMON_MMAP(X) \ */ +/* unsigned char __PASTE3(lc_,X,_data)[__PASTE3(__lc_,X,_data_LEN)]; */ + +/* #define COMMON_MMIDX(X) \ */ +/* unsigned char __PASTE3(lc_,X,_rows)[__PASTE3(__lc_,X,_rows_LEN)]; \ */ +/* uint16_t __PASTE3(lc_,X,_item_offsets)[__PASTE3(__lc_,X,_item_offsets_LEN)]; \ */ +/* uint16_t __PASTE3(lc_,X,_item_idx)[__PASTE3(__lc_,X,_item_idx_LEN)]; \ */ + +/* ---------------------------------------------------------------------- */ + +#define COMMON_OFFSETS(X) \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_rows)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_offsets)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_idx)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_data)) \ + + +static const size_t common_tbl_offsets[CATEGORIES*4] = { + 0, 0, 0, 0, /* ctype */ + COMMON_OFFSETS(numeric), + COMMON_OFFSETS(monetary), + COMMON_OFFSETS(time), + 0, 0, 0, 0, /* collate */ + COMMON_OFFSETS(messages), +}; + + +void out_uc(FILE *f, const unsigned char *p, size_t n, char *comment) +{ + size_t i; + + fprintf(f, "{\t/* %s */", comment); + for (i = 0 ; i < n ; i++) { + if (!(i & 7)) { + fprintf(f, "\n\t"); + } + if (p[i]) { + fprintf(f, "%#04x, ", p[i]); + } else { + fprintf(f, "%#4x, ", p[i]); + } + } + fprintf(f, "\n},\n"); +} + +void out_u16(FILE *f, const uint16_t *p, size_t n, char *comment) +{ + size_t i; + + fprintf(f, "{\t/* %s */", comment); + for (i = 0 ; i < n ; i++) { + if (!(i & 7)) { + fprintf(f, "\n\t"); + } + if (p[i]) { + fprintf(f, "%#06x, ", p[i]); + } else { + fprintf(f, "%#6x, ", p[i]); + } + } + fprintf(f, "\n},\n"); +} + +void out_i16(FILE *f, const int16_t *p, size_t n, char *comment) +{ + size_t i; + + fprintf(f, "{\t/* %s */", comment); + for (i = 0 ; i < n ; i++) { + if (!(i & 7)) { + fprintf(f, "\n\t"); + } + fprintf(f, "%6d, ", p[i]); + } + fprintf(f, "\n},\n"); +} + +void out_size_t(FILE *f, const size_t *p, size_t n, char *comment) +{ + size_t i; + + fprintf(f, "{\t/* %s */", comment); + for (i = 0 ; i < n ; i++) { + if (!(i & 3)) { + fprintf(f, "\n\t"); + } + if (p[i]) { + fprintf(f, "%#010zx, ", p[i]); + } else { + fprintf(f, "%#10zx, ", p[i]); + } + } + fprintf(f, "\n},\n"); +} + + +int main(void) +{ + FILE *lso; /* static object */ + int i; + unsigned char magic[MAGIC_SIZE]; + + memset(magic, 0, MAGIC_SIZE); + + if (!(lso = fopen("locale_data.c", "w"))) { + printf("can't open locale_data.c!\n"); + return EXIT_FAILURE; + } + + fprintf(lso, + "#include \n" + "#include \n" +/* "#define __CTYPE_HAS_8_BIT_LOCALES\n" */ + "#ifndef __WCHAR_ENABLED\n" + "#error __WCHAR_ENABLED not defined\n" + "#endif\n" + "#include \"c8tables.h\"\n" + "#include \"wctables.h\"\n" + "#include \"lt_defines.h\"\n" + "#include \"locale_mmap.h\"\n\n" + "static const __locale_mmap_t locale_mmap = {\n\n" + ); + out_uc(lso, magic, MAGIC_SIZE, "magic"); +#ifdef __CTYPE_HAS_8_BIT_LOCALES + out_uc(lso, Cctype_data, Cctype_TBL_LEN, "tbl8ctype"); + out_uc(lso, Cuplow_data, Cuplow_TBL_LEN, "tbl8uplow"); +#ifdef __WCHAR_ENABLED + out_u16(lso, Cc2wc_data, Cc2wc_TBL_LEN, "tbl8c2wc"); + out_uc(lso, Cwc2c_data, Cwc2c_TBL_LEN, "tbl8wc2c"); + /* translit */ +#endif /* __WCHAR_ENABLED */ +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ +#ifdef __WCHAR_ENABLED + out_uc(lso, WCctype_data, WCctype_TBL_LEN, "tblwctype"); + out_uc(lso, WCuplow_data, WCuplow_TBL_LEN, "tblwuplow"); + out_i16(lso, WCuplow_diff_data, WCuplow_diff_TBL_LEN, "tblwuplow_diff"); +/* const unsigned char tblwcomb[WCcomb_TBL_LEN]; */ + /* width?? */ +#endif /* __WCHAR_ENABLED */ + out_uc(lso, __lc_numeric_data, __lc_numeric_data_LEN, "lc_numeric_data"); + out_uc(lso, __lc_monetary_data, __lc_monetary_data_LEN, "lc_monetary_data"); + out_uc(lso, __lc_time_data, __lc_time_data_LEN, "lc_time_data"); + /* TODO -- collate*/ + out_uc(lso, __lc_messages_data, __lc_messages_data_LEN, "lc_messages_data"); + +#ifdef __CTYPE_HAS_8_BIT_LOCALES + fprintf(lso, "{ /* codeset_8_bit array */\n"); + for (i = 0 ; i < NUM_CODESETS ; i++) { + fprintf(lso, "{ /* codeset_8_bit[%d] */\n", i); + out_uc(lso, codeset_8_bit[i].idx8ctype, Cctype_IDX_LEN, "idx8ctype"); + out_uc(lso, codeset_8_bit[i].idx8uplow, Cuplow_IDX_LEN, "idx8uplow"); + out_uc(lso, codeset_8_bit[i].idx8c2wc, Cc2wc_IDX_LEN, "idx8c2wc"); + out_uc(lso, codeset_8_bit[i].idx8wc2c, Cwc2c_II_LEN, "idx8wc2c"); + fprintf(lso, "},\n"); + } + fprintf(lso, "},\n"); +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ + + out_uc(lso, __lc_numeric_rows, __lc_numeric_rows_LEN, "lc_numeric_rows"); + out_u16(lso, __lc_numeric_item_offsets, __lc_numeric_item_offsets_LEN, "lc_numeric_item_offsets"); + out_u16(lso, __lc_numeric_item_idx, __lc_numeric_item_idx_LEN, "lc_numeric_item_idx"); + + out_uc(lso, __lc_monetary_rows, __lc_monetary_rows_LEN, "lc_monetary_rows"); + out_u16(lso, __lc_monetary_item_offsets, __lc_monetary_item_offsets_LEN, "lc_monetary_item_offsets"); + out_u16(lso, __lc_monetary_item_idx, __lc_monetary_item_idx_LEN, "lc_monetary_item_idx"); + + out_uc(lso, __lc_time_rows, __lc_time_rows_LEN, "lc_time_rows"); + out_u16(lso, __lc_time_item_offsets, __lc_time_item_offsets_LEN, "lc_time_item_offsets"); + out_u16(lso, __lc_time_item_idx, __lc_time_item_idx_LEN, "lc_time_item_idx"); + + /* TODO -- collate*/ + + out_uc(lso, __lc_messages_rows, __lc_messages_rows_LEN, "lc_messages_rows"); + out_u16(lso, __lc_messages_item_offsets, __lc_messages_item_offsets_LEN, "lc_messages_item_offsets"); + out_u16(lso, __lc_messages_item_idx, __lc_messages_item_idx_LEN, "lc_messages_item_idx"); + + { + unsigned char co_buf[CATEGORIES] = { + 1, __lc_numeric_item_offsets_LEN, __lc_monetary_item_offsets_LEN, + __lc_time_item_offsets_LEN, 0, __lc_messages_item_offsets_LEN + }; + out_uc(lso, co_buf, CATEGORIES, "lc_common_item_offsets_LEN"); + } + + out_size_t(lso, common_tbl_offsets, CATEGORIES * 4, "lc_common_tbl_offsets"); + /* offsets from start of locale_mmap_t */ + /* rows, item_offsets, item_idx, data */ + +#ifdef NUM_LOCALES + out_uc(lso, __locales, NUM_LOCALES * WIDTH_LOCALES, "locales"); + out_uc(lso, __locale_names5, 5 * NUM_LOCALE_NAMES, "locale_names5"); +#ifdef LOCALE_AT_MODIFIERS_LENGTH + out_uc(lso, __locale_at_modifiers, LOCALE_AT_MODIFIERS_LENGTH, "locale_at_modifiers"); +#else +#error LOCALE_AT_MODIFIERS_LENGTH not defined! +#endif /* LOCALE_AT_MODIFIERS_LENGTH */ +#endif /* NUM_LOCALES */ + + out_uc(lso, lc_names, lc_names_LEN, "lc_names"); +#ifdef __CTYPE_HAS_8_BIT_LOCALES + out_uc(lso, (const unsigned char*) CODESET_LIST, sizeof(CODESET_LIST), "codeset_list"); +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ + + fprintf(lso, + "\n};\n\n" + "const __locale_mmap_t *__locale_mmap = &locale_mmap;\n\n" + ); + + if (ferror(lso) || fclose(lso)) { + printf("error writing!\n"); + return EXIT_FAILURE; + } + + return EXIT_SUCCESS; +} + + + +/* ---------------------------------------------------------------------- */ + +#if 0 +#define WRITE_COMMON_MMAP(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE3(lc_,X,_data)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(__lc_,X,_data_LEN) ; i++) { \ + putc(__PASTE3(__lc_,X,_data)[i], fp); \ + } + +#define WRITE_COMMON_MMIDX(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE3(lc_,X,_rows)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(__lc_,X,_rows_LEN) ; i++) { \ + putc(__PASTE3(__lc_,X,_rows)[i], fp); \ + } \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_offsets)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(__lc_,X,_item_offsets_LEN) ; i++) { \ + putc( ((unsigned char *) &(__PASTE3(__lc_,X,_item_offsets)[i]))[0], fp); \ + putc( ((unsigned char *) &(__PASTE3(__lc_,X,_item_offsets)[i]))[1], fp); \ + } \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_idx)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(__lc_,X,_item_idx_LEN) ; i++) { \ + putc( ((unsigned char *) &(__PASTE3(__lc_,X,_item_idx)[i]))[0], fp); \ + putc( ((unsigned char *) &(__PASTE3(__lc_,X,_item_idx)[i]))[1], fp); \ + } + +#define WRITE_WC_DATA(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE2(tblw,X)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(WC,X,_TBL_LEN) ; i++) { \ + putc(__PASTE3(WC,X,_data)[i], fp); \ + } + +#define WRITE_WC_I16_DATA(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE2(tblw,X)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(WC,X,_TBL_LEN) ; i++) { \ + putc( ((unsigned char *) &(__PASTE3(WC,X,_data)[i]))[0], fp); \ + putc( ((unsigned char *) &(__PASTE3(WC,X,_data)[i]))[1], fp); \ + } + +#define WRITE_C_DATA(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE2(tbl8,X)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(C,X,_TBL_LEN) ; i++) { \ + putc(__PASTE3(C,X,_data)[i], fp); \ + } + +#define WRITE_C_U16_DATA(X) \ + fseek(fp, (long) offsetof(__locale_mmap_t, __PASTE2(tbl8,X)), SEEK_SET); \ + for (i=0 ; i < __PASTE3(C,X,_TBL_LEN) ; i++) { \ + putc( ((unsigned char *) &(__PASTE3(C,X,_data)[i]))[0], fp); \ + putc( ((unsigned char *) &(__PASTE3(C,X,_data)[i]))[1], fp); \ + } + +/**********************************************************************/ + +#define COMMON_OFFSETS(X) \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_rows)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_offsets)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_item_idx)), \ + offsetof(__locale_mmap_t, __PASTE3(lc_,X,_data)) \ + + +static const size_t common_tbl_offsets[CATEGORIES*4] = { + 0, 0, 0, 0, /* ctype */ + COMMON_OFFSETS(numeric), + COMMON_OFFSETS(monetary), + COMMON_OFFSETS(time), + 0, 0, 0, 0, /* collate */ + COMMON_OFFSETS(messages), +}; + + + + + +int main(void) +{ + FILE *fp; + size_t i; + unsigned char *p; + + if (!(fp = fopen("locale.mmap", "w"))) { + printf("error - can't open locale.mmap for writing!"); + return EXIT_FAILURE; + } + + for (i=0 ; i < sizeof(__locale_mmap_t) ; i++) { + putc(0, fp); /* Zero out the file. */ + } + +#ifdef __CTYPE_HAS_8_BIT_LOCALES + WRITE_C_DATA(ctype); + WRITE_C_DATA(uplow); +#ifdef __WCHAR_ENABLED + WRITE_C_U16_DATA(c2wc); + WRITE_C_DATA(wc2c); + /* translit */ +#endif /* __WCHAR_ENABLED */ +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ + +#ifdef __WCHAR_ENABLED + WRITE_WC_DATA(ctype); + WRITE_WC_DATA(uplow); + WRITE_WC_I16_DATA(uplow_diff); +/* WRITE_WC_DATA(comb); */ + /* width?? */ +#endif /* __WCHAR_ENABLED */ + + WRITE_COMMON_MMAP(numeric); + WRITE_COMMON_MMAP(monetary); + WRITE_COMMON_MMAP(time); + /* TODO -- collate*/ + WRITE_COMMON_MMAP(messages); + +#ifdef __CTYPE_HAS_8_BIT_LOCALES + fseek(fp, (long) offsetof(__locale_mmap_t, codeset_8_bit), SEEK_SET); \ + p = (unsigned char *) codeset_8_bit; + for (i=0 ; i < sizeof(codeset_8_bit) ; i++) { + putc(p[i], fp); + } +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ + + WRITE_COMMON_MMIDX(numeric); + WRITE_COMMON_MMIDX(monetary); + WRITE_COMMON_MMIDX(time); + /* TODO -- collate*/ + WRITE_COMMON_MMIDX(messages); + + fseek(fp, (long) offsetof(__locale_mmap_t, lc_common_item_offsets_LEN), SEEK_SET); + putc(1, fp); /* ctype -- (codeset) handled specially */ + putc(__lc_numeric_item_offsets_LEN, fp); + putc(__lc_monetary_item_offsets_LEN, fp); + putc(__lc_time_item_offsets_LEN, fp); + putc(0, fp); /* collate */ + putc(__lc_messages_item_offsets_LEN, fp); + + fseek(fp, (long) offsetof(__locale_mmap_t, lc_common_tbl_offsets), SEEK_SET); + for (i=0 ; i < sizeof(common_tbl_offsets) ; i++) { + putc(((unsigned char *)common_tbl_offsets)[i], fp); + } + +#ifdef NUM_LOCALES + fseek(fp, (long) offsetof(__locale_mmap_t, locales), SEEK_SET); + for (i=0 ; i < (NUM_LOCALES * WIDTH_LOCALES) ; i++) { + putc(__locales[i], fp); + } + + fseek(fp, (long) offsetof(__locale_mmap_t, locale_names5), SEEK_SET); + for (i=0 ; i < 5 * NUM_LOCALE_NAMES ; i++) { + putc(__locale_names5[i], fp); + } + +#ifdef LOCALE_AT_MODIFIERS_LENGTH + fseek(fp, (long) offsetof(__locale_mmap_t, locale_at_modifiers), SEEK_SET); + for (i=0 ; i < LOCALE_AT_MODIFIERS_LENGTH ; i++) { + putc(__locale_at_modifiers[i], fp); + } +#endif /* LOCALE_AT_MODIFIERS_LENGTH */ +#endif /* NUM_LOCALES */ + + fseek(fp, (long) offsetof(__locale_mmap_t, lc_names), SEEK_SET); + for (i=0 ; i < lc_names_LEN ; i++) { + putc(lc_names[i], fp); + } + +#ifdef __CTYPE_HAS_8_BIT_LOCALES + fseek(fp, (long) offsetof(__locale_mmap_t, codeset_list), SEEK_SET); + for (i=0 ; i < sizeof(CODESET_LIST) ; i++) { + putc((unsigned char)(CODESET_LIST[i]), fp); + } +#endif /* __CTYPE_HAS_8_BIT_LOCALES */ + + + if (ferror(fp) || (fclose(fp) == EOF)) { + printf("error - stream in error state or fclose failed!"); + return EXIT_FAILURE; + } + + printf("sizeof(__locale_mmap_t) = %zd\n", sizeof(__locale_mmap_t)); + + return EXIT_SUCCESS; +} +#endif + +/* TODO: + * collate data (8-bit weighted single char only) + * @ mappings! + * codeset list? yes, since we'll want to be able to inspect them... + * that means putting some header stuff in magic + * fix ctype LEN defines in gen_c8tables + */ diff --git a/extra/locale/gen_locale.c b/extra/locale/gen_locale.c index 047b843dc..268c3d61b 100644 --- a/extra/locale/gen_locale.c +++ b/extra/locale/gen_locale.c @@ -10,7 +10,6 @@ #include #include -#define __CTYPE_HAS_8_BIT_LOCALES #include "c8tables.h" @@ -101,6 +100,9 @@ static void do_locale_names(void) fprintf(ofp, "static const unsigned char __locales[%d];\n", (3 + CATEGORIES)); fprintf(ofp, "static const unsigned char __locale_names5[5];\n"); } else { + if (default_utf8) { + fprintf(ofp, "#define __CTYPE_HAS_UTF_8_LOCALES\t\t\t1\n"); + } fprintf(ofp, "#define CATEGORIES\t\t\t%d\n", CATEGORIES); fprintf(ofp, "#define WIDTH_LOCALES\t\t\t%d\n", 3+CATEGORIES); fprintf(ofp, "#define NUM_LOCALES\t\t\t%d\n", num_locales); diff --git a/extra/locale/gen_mmap.c b/extra/locale/gen_mmap.c index ae5d4dff6..f9bce022c 100644 --- a/extra/locale/gen_mmap.c +++ b/extra/locale/gen_mmap.c @@ -16,13 +16,13 @@ #define WANT_WCctype_data #define WANT_WCuplow_data #define WANT_WCuplow_diff_data -#define WANT_WCcomb_data +/* #define WANT_WCcomb_data */ /* #define WANT_WCwidth_data */ #include "wctables.h" #undef WANT_WCctype_data #undef WANT_WCuplow_data #undef WANT_WCuplow_diff_data -#undef WANT_WCcomb_data +/* #undef WANT_WCcomb_data */ /* #undef WANT_WCwidth_data */ /* #define WCctype_TBL_LEN (WCctype_II_LEN + WCctype_TI_LEN + WCctype_UT_LEN) */ @@ -149,7 +149,7 @@ int main(void) WRITE_WC_DATA(ctype); WRITE_WC_DATA(uplow); WRITE_WC_I16_DATA(uplow_diff); - WRITE_WC_DATA(comb); +/* WRITE_WC_DATA(comb); */ /* width?? */ #endif /* __WCHAR_ENABLED */ diff --git a/extra/locale/gen_wc8bit.c b/extra/locale/gen_wc8bit.c index 785c3f4c0..d76c15a51 100644 --- a/extra/locale/gen_wc8bit.c +++ b/extra/locale/gen_wc8bit.c @@ -7,6 +7,14 @@ #include #include +#ifndef __UCLIBC__ +#ifndef _WCTYPE_H +#define _WCTYPE_H +#endif +#include "../../libc/sysdeps/linux/common/bits/uClibc_ctype.h" +#endif + + /* #define CTYPE_PACKED */ #define UPLOW_IDX_SHIFT 3 /* best if 2 unpacked or 3 packed */ @@ -63,34 +71,6 @@ typedef struct { unsigned char c2wc_idx[C2WC_IDX_LEN]; } charset_data; -/* Taking advantage of the C99 mutual-exclusion guarantees for the various - * (w)ctype classes, including the descriptions of printing and control - * (w)chars, we can place each in one of the following mutually-exlusive - * subsets. Since there are less than 16, we can store the data for - * each (w)chars in a nibble. In contrast, glibc uses an unsigned int - * per (w)char, with one bit flag for each is* type. While this allows - * a simple '&' operation to determine the type vs. a range test and a - * little special handling for the "blank" and "xdigit" types in my - * approach, it also uses 8 times the space for the tables on the typical - * 32-bit archs we supported.*/ -enum { - __CTYPE_unclassified = 0, - __CTYPE_alpha_nonupper_nonlower, - __CTYPE_alpha_lower, - __CTYPE_alpha_upper_lower, - __CTYPE_alpha_upper, - __CTYPE_digit, - __CTYPE_punct, - __CTYPE_graph, - __CTYPE_print_space_nonblank, - __CTYPE_print_space_blank, - __CTYPE_space_nonblank_noncntrl, - __CTYPE_space_blank_noncntrl, - __CTYPE_cntrl_space_nonblank, - __CTYPE_cntrl_space_blank, - __CTYPE_cntrl_nonspace, -}; - int main(int argc, char **argv) { FILE *fp; diff --git a/extra/locale/gen_wctype.c b/extra/locale/gen_wctype.c index 9508a3bbf..1c8c10c43 100644 --- a/extra/locale/gen_wctype.c +++ b/extra/locale/gen_wctype.c @@ -10,6 +10,12 @@ #include #include +#ifndef __UCLIBC__ +#ifndef _WCTYPE_H +#define _WCTYPE_H +#endif +#include "../../libc/sysdeps/linux/common/bits/uClibc_ctype.h" +#endif /* 0x9 : space blank */ /* 0xa : space */ @@ -61,6 +67,7 @@ #define RANGE 0xffffUL /* Restrict for 16-bit wchar_t... */ #endif +#if 0 /* Classification codes. */ static const char *typename[] = { @@ -81,7 +88,9 @@ static const char *typename[] = { "C_cntrl_nonspace", "empty_slot" }; +#endif +#if 0 /* Taking advantage of the C99 mutual-exclusion guarantees for the various * (w)ctype classes, including the descriptions of printing and control * (w)chars, we can place each in one of the following mutually-exlusive @@ -109,21 +118,8 @@ enum { __CTYPE_cntrl_space_blank, __CTYPE_cntrl_nonspace, }; +#endif -/* Some macros that test for various (w)ctype classes when passed one of the - * designator values enumerated above. */ -#define __CTYPE_isalnum(D) ((unsigned int)(D-1) <= (__CTYPE_digit-1)) -#define __CTYPE_isalpha(D) ((unsigned int)(D-1) <= (__CTYPE_alpha_upper-1)) -#define __CTYPE_isblank(D) \ - ((((unsigned int)(D - __CTYPE_print_space_nonblank)) <= 5) && (D & 1)) -#define __CTYPE_iscntrl(D) (((unsigned int)(D - __CTYPE_cntrl_space_nonblank)) <= 2) -#define __CTYPE_isdigit(D) (D == __CTYPE_digit) -#define __CTYPE_isgraph(D) ((unsigned int)(D-1) <= (__CTYPE_graph-1)) -#define __CTYPE_islower(D) (((unsigned int)(D - __CTYPE_alpha_lower)) <= 1) -#define __CTYPE_isprint(D) ((unsigned int)(D-1) <= (__CTYPE_print_space_blank-1)) -#define __CTYPE_ispunct(D) (D == __CTYPE_punct) -#define __CTYPE_isspace(D) (((unsigned int)(D - __CTYPE_print_space_nonblank)) <= 5) -#define __CTYPE_isupper(D) (((unsigned int)(D - __CTYPE_alpha_upper_lower)) <= 1) #define __CTYPE_isxdigit(D,X) \ (__CTYPE_isdigit(D) || (((unsigned int)(((X)|0x20) - 'a')) <= 5)) @@ -232,6 +228,7 @@ int main(int argc, char **argv) table_data ultable; table_data combtable; table_data widthtable; + long int last_comb = 0; unsigned char wct[(RANGE/2)+1]; /* wctype table (nibble per wchar) */ unsigned char ult[RANGE+1]; /* upper/lower table */ @@ -241,8 +238,32 @@ int main(int argc, char **argv) wctype_t is_comb, is_comb3; long int typecount[16]; + const char *typename[16]; + static const char empty_slot[] = "empty_slot"; int built = 0; +#define INIT_TYPENAME(X) typename[__CTYPE_##X] = "C_" #X + + for (i=0 ; i < 16 ; i++) { + typename[i] = empty_slot; + } + + INIT_TYPENAME(unclassified); + INIT_TYPENAME(alpha_nonupper_nonlower); + INIT_TYPENAME(alpha_lower); + INIT_TYPENAME(alpha_upper_lower); + INIT_TYPENAME(alpha_upper); + INIT_TYPENAME(digit); + INIT_TYPENAME(punct); + INIT_TYPENAME(graph); + INIT_TYPENAME(print_space_nonblank); + INIT_TYPENAME(print_space_blank); + INIT_TYPENAME(space_nonblank_noncntrl); + INIT_TYPENAME(space_blank_noncntrl); + INIT_TYPENAME(cntrl_space_nonblank); + INIT_TYPENAME(cntrl_space_blank); + INIT_TYPENAME(cntrl_nonspace); + setvbuf(stdout, NULL, _IONBF, 0); while (--argc) { @@ -393,14 +414,27 @@ int main(int argc, char **argv) /* } */ } #endif - +#if 0 + if (iswctype(c,is_comb) || iswctype(c,is_comb3)) { + if (!last_comb) { + printf("%#8x - ", c); + last_comb = c; + } else if (last_comb + 1 < c) { + printf("%#8x\n%#8x - ", last_comb, c); + last_comb = c; + } else { + last_comb = c; + } + } +#endif } #endif combt[c/4] |= ((((!!iswctype(c,is_comb)) << 1) | !!iswctype(c,is_comb3)) << ((c & 3) << 1)); /* comb3t[c/8] |= ((!!iswctype(c,is_comb3)) << (c & 7)); */ - widtht[c/4] |= (wcwidth(c) << ((c & 3) << 1)); + +/* widtht[c/4] |= (wcwidth(c) << ((c & 3) << 1)); */ if (c & 1) { /* Use the high nibble for odd numbered wchars. */ d <<= 4; @@ -483,6 +517,7 @@ int main(int argc, char **argv) newopt(ult, RANGE+1, n, &ultable); +#if 0 printf("optimizing comb table..\n"); smallest = SIZE_MAX; combtable.ii = NULL; @@ -503,8 +538,10 @@ int main(int argc, char **argv) smallest = SIZE_MAX; newopt(combt, sizeof(combt), n, &combtable); combtable.ti_shift += 4; /* correct for 4 entries per */ +#endif +#if 0 printf("optimizing width table..\n"); smallest = SIZE_MAX; widthtable.ii = NULL; @@ -525,7 +562,7 @@ int main(int argc, char **argv) smallest = SIZE_MAX; newopt(widtht, sizeof(widtht), n, &widthtable); widthtable.ti_shift += 4; /* correct for 4 entries per */ - +#endif #if 0 printf("optimizing comb3 table..\n"); @@ -553,7 +590,6 @@ int main(int argc, char **argv) dump_table_data(&cttable); dump_table_data(&ultable); dump_table_data(&combtable); - dump_table_data(&widthtable); } printf("verifying for %s...\n", *argv); @@ -715,8 +751,8 @@ int main(int argc, char **argv) fprintf(fp, "#endif /* WANT_WCuplow_diff_data */\n\n"); - output_table(fp, "comb", &combtable); - output_table(fp, "width", &widthtable); +/* output_table(fp, "comb", &combtable); */ +/* output_table(fp, "width", &widthtable); */ fclose(fp); } diff --git a/extra/locale/locale_mmap.h b/extra/locale/locale_mmap.h index 1ed7a2094..3db490407 100644 --- a/extra/locale/locale_mmap.h +++ b/extra/locale/locale_mmap.h @@ -8,7 +8,7 @@ #define WCctype_TBL_LEN (WCctype_II_LEN + WCctype_TI_LEN + WCctype_UT_LEN) #define WCuplow_TBL_LEN (WCuplow_II_LEN + WCuplow_TI_LEN + WCuplow_UT_LEN) #define WCuplow_diff_TBL_LEN (2 * WCuplow_diffs) -#define WCcomb_TBL_LEN (WCcomb_II_LEN + WCcomb_TI_LEN + WCcomb_UT_LEN) +/* #define WCcomb_TBL_LEN (WCcomb_II_LEN + WCcomb_TI_LEN + WCcomb_UT_LEN) */ #endif #undef __PASTE2 @@ -41,7 +41,7 @@ typedef struct { const unsigned char tblwctype[WCctype_TBL_LEN]; const unsigned char tblwuplow[WCuplow_TBL_LEN]; const int16_t tblwuplow_diff[WCuplow_diff_TBL_LEN]; - const unsigned char tblwcomb[WCcomb_TBL_LEN]; +/* const unsigned char tblwcomb[WCcomb_TBL_LEN]; */ /* width?? */ #endif /* __WCHAR_ENABLED */ diff --git a/include/iconv.h b/include/iconv.h new file mode 100644 index 000000000..e17360335 --- /dev/null +++ b/include/iconv.h @@ -0,0 +1,54 @@ +/* Copyright (C) 1997, 1998, 1999, 2000 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#ifndef _ICONV_H +#define _ICONV_H 1 + +#include +#define __need_size_t +#include + +#ifndef __UCLIBC_HAS_LOCALE__ +#error Attempted to include iconv.h when uClibc built without locale support. +#endif + +__BEGIN_DECLS + +/* Identifier for conversion method from one codeset to another. */ +typedef void *iconv_t; + + +/* Allocate descriptor for code conversion from codeset FROMCODE to + codeset TOCODE. */ +extern iconv_t iconv_open (__const char *__tocode, __const char *__fromcode) + __THROW; + +/* Convert at most *INBYTESLEFT bytes from *INBUF according to the + code conversion algorithm specified by CD and place up to + *OUTBYTESLEFT bytes in buffer at *OUTBUF. */ +extern size_t iconv (iconv_t __cd, char **__restrict __inbuf, + size_t *__restrict __inbytesleft, + char **__restrict __outbuf, + size_t *__restrict __outbytesleft); + +/* Free resources allocated for descriptor CD for code conversion. */ +extern int iconv_close (iconv_t __cd) __THROW; + +__END_DECLS + +#endif /* iconv.h */ diff --git a/ldso/util/Makefile b/ldso/util/Makefile index dcd91dbaf..35c2dceab 100644 --- a/ldso/util/Makefile +++ b/ldso/util/Makefile @@ -74,5 +74,5 @@ ldd.target: ldd.c $(STRIPTOOL) -x -R .note -R .comment $@ clean: - rm -f $(TARGETS) *.o *~ core ./elf.h + rm -f $(TARGETS) *.o *~ core ./elf.h *.target diff --git a/libc/misc/locale/locale.c b/libc/misc/locale/locale.c index 0ff359d3d..0875a4e5b 100644 --- a/libc/misc/locale/locale.c +++ b/libc/misc/locale/locale.c @@ -454,8 +454,8 @@ void _locale_init(void) = (const unsigned char *) &__locale_mmap->tblwuplow; __global_locale.tblwuplow_diff = (const uint16_t *) &__locale_mmap->tblwuplow_diff; - __global_locale.tblwcomb - = (const unsigned char *) &__locale_mmap->tblwcomb; +/* __global_locale.tblwcomb */ +/* = (const unsigned char *) &__locale_mmap->tblwcomb; */ /* width?? */ #endif /* __WCHAR_ENABLED */ diff --git a/libc/misc/time/time.c b/libc/misc/time/time.c index 378854adc..838a3a025 100644 --- a/libc/misc/time/time.c +++ b/libc/misc/time/time.c @@ -90,6 +90,9 @@ * Defining __TIME_TZ_OPT_SPEED will cause a tzset() to keep a copy of the * last TZ setting string and do a "fast out" if the current string is the * same. + * + * Nov 21, 2002 Fix an error return case in _time_mktime. + * */ @@ -2022,6 +2025,10 @@ time_t _time_mktime(struct tm *timeptr, int store_on_success) localtime_r(&t, (struct tm *)p); + if (t < 0) { + return -1; + } + if (store_on_success) { memcpy(timeptr, p, sizeof(struct tm)); } diff --git a/libc/misc/wchar/Makefile b/libc/misc/wchar/Makefile index 7ac27b6ea..acc852195 100644 --- a/libc/misc/wchar/Makefile +++ b/libc/misc/wchar/Makefile @@ -23,12 +23,17 @@ TOPDIR=../../../ include $(TOPDIR)Rules.mak +TARGET_CC = $(TOPDIR)extra/gcc-uClibc/$(TARGET_ARCH)-uclibc-gcc MSRC1= wchar.c MOBJ1= btowc.o wctob.o mbsinit.o mbrlen.o mbrtowc.o wcrtomb.o mbsrtowcs.o \ wcsrtombs.o _wchar_utf8sntowcs.o _wchar_wcsntoutf8s.o \ __mbsnrtowcs.o __wcsnrtombs.o wcwidth.o wcswidth.o +ifeq ($(UCLIBC_HAS_LOCALE),y) + MOBJ1 += iconv.o +endif + MSRC2= wstdio.c MOBJ2= fwide.o \ fgetwc.o getwchar.o fgetws.o \ @@ -37,10 +42,6 @@ MOBJ2= fwide.o \ # getwc (fgetwc alias) getwc_unlocked (fgetwc_unlocked alias) # putwc (fputwc alias) putwc_unlocked (fputwc_unlocked alias) -# wcwidth wcswidth -# wcstod wcstof wcstold -# wcstol wcstoul wcstoq wcstouq wcstoll wcstoull -# fwprintf wprintf swprintf vfwprintf vwprintf vswprintf # fwscanf wscanf swscanf vfwscanf vwscanf vswscanf # wcsftime @@ -61,6 +62,10 @@ $(MOBJ2): $(MSRC2) $(CC) $(CFLAGS) -DL_$* $< -c -o $*.o $(STRIPTOOL) -x -R .note -R .comment $*.o +iconv.target: wchar.c + $(TARGET_CC) $(CFLAGS) -DL_iconv_main wchar.c -o $@ + $(STRIPTOOL) -x -R .note -R .comment $@ + clean: - rm -f *.[oa] *~ core + rm -f *.[oa] *~ core iconv.target diff --git a/libc/misc/wchar/wchar.c b/libc/misc/wchar/wchar.c index 7da3b1a50..a6e5fc616 100644 --- a/libc/misc/wchar/wchar.c +++ b/libc/misc/wchar/wchar.c @@ -1,3 +1,4 @@ + /* Copyright (C) 2002 Manuel Novoa III * * This library is free software; you can redistribute it and/or @@ -79,6 +80,12 @@ * framework. Minimally tested at the moment, but the stub C-locale * version (which most people would probably be using) should be fine. * + * Nov 21, 2002 + * + * Revert the wc<->mb changes from earlier this month involving the C-locale. + * Add a couple of ugly hacks to support *wprintf. + * Add a mini iconv() and iconv implementation (requires locale support). + * * Manuel */ @@ -97,13 +104,17 @@ #ifdef __UCLIBC_HAS_LOCALE__ #define ENCODING (__global_locale.encoding) -#ifdef __UCLIBC_MJN3_ONLY__ -#warning implement __CTYPE_HAS_UTF_8_LOCALES! +#ifndef __CTYPE_HAS_UTF_8_LOCALES +#warning __CTYPE_HAS_UTF_8_LOCALES not set! #endif -#define __CTYPE_HAS_UTF_8_LOCALES #else -#undef __CTYPE_HAS_8_BIT_LOCALES -#undef __CTYPE_HAS_UTF_8_LOCALES +#define ENCODING (__ctype_encoding_7_bit) +#ifdef __CTYPE_HAS_8_BIT_LOCALES +#error __CTYPE_HAS_8_BIT_LOCALES is defined! +#endif +#ifdef __CTYPE_HAS_UTF_8_LOCALES +#error __CTYPE_HAS_UTF_8_LOCALES is defined! +#endif #undef L__wchar_utf8sntowcs #undef L__wchar_wcsntoutf8s #endif @@ -114,7 +125,7 @@ #define UTF_8_MAX_LEN 3 #endif -/* #define KUHN */ +#define KUHN 1 /* Implementation-specific work functions. */ @@ -157,17 +168,14 @@ wint_t btowc(int c) #else /* __CTYPE_HAS_8_BIT_LOCALES */ -#ifdef __CTYPE_HAS_UTF_8_LOCALES - if (ENCODING == __ctype_encoding_utf8) { - return (((unsigned int)c) < 0x80) ? c : WEOF; - } -#endif /* __CTYPE_HAS_UTF_8_LOCALES */ - #ifdef __UCLIBC_HAS_LOCALE__ - assert(ENCODING == __ctype_encoding_7_bit); -#endif + assert((ENCODING == __ctype_encoding_7_bit) + || (ENCODING == __ctype_encoding_utf8)); +#endif /* __UCLIBC_HAS_LOCALE__ */ - return (((unsigned int)c) <= UCHAR_MAX) ? c : WEOF; + /* If we don't have 8-bit locale support, then this is trivial since + * anything outside of 0-0x7f is illegal in C/POSIX and UTF-8 locales. */ + return (((unsigned int)c) < 0x80) ? c : WEOF; #endif /* __CTYPE_HAS_8_BIT_LOCALES */ } @@ -188,17 +196,18 @@ int wctob(wint_t c) #else /* __CTYPE_HAS_8_BIT_LOCALES */ -#ifdef __CTYPE_HAS_UTF_8_LOCALES - if (ENCODING == __ctype_encoding_utf8) { - return ((c >= 0) && (c < 0x80)) ? c : EOF; - } -#endif /* __CTYPE_HAS_UTF_8_LOCALES */ - #ifdef __UCLIBC_HAS_LOCALE__ - assert(ENCODING == __ctype_encoding_7_bit); -#endif + assert((ENCODING == __ctype_encoding_7_bit) + || (ENCODING == __ctype_encoding_utf8)); +#endif /* __UCLIBC_HAS_LOCALE__ */ + + /* If we don't have 8-bit locale support, then this is trivial since + * anything outside of 0-0x7f is illegal in C/POSIX and UTF-8 locales. */ + + /* TODO: need unsigned version of wint_t... */ +/* return (((unsigned int)c) < 0x80) ? c : WEOF; */ + return ((c >= 0) && (c < 0x80)) ? c : EOF; - return ((c >= 0) && (c <= UCHAR_MAX)) ? c : EOF; #endif /* __CTYPE_HAS_8_BIT_LOCALES */ } @@ -359,9 +368,19 @@ size_t _wchar_utf8sntowcs(wchar_t *__restrict pwc, size_t wn, assert(ps != NULL); incr = 1; - if (!pwc) { + /* NOTE: The following is an AWFUL HACK! In order to support %s in + * wprintf, we need to be able to compute the number of wchars needed + * for the mbs conversion, not to exceed the precision specified. + * But if dst is NULL, the return value is the length assuming a + * sufficiently sized buffer. So, we allow passing of (wchar_t *) ps + * as pwc in order to flag that we really want the length, subject + * to the restricted buffer size and no partial conversions. + * See mbsnrtowcs() as well. */ + if (!pwc || (pwc == ((wchar_t *)ps))) { + if (!pwc) { + wn = SIZE_MAX; + } pwc = wcbuf; - wn = SIZE_MAX; incr = 0; } @@ -406,7 +425,7 @@ size_t _wchar_utf8sntowcs(wchar_t *__restrict pwc, size_t wn, } BAD: #ifdef DECODER - wc = 0xfffd; + wc = 0xfffdU; goto COMPLETE; #else ps->mask = mask; @@ -635,9 +654,19 @@ size_t __mbsnrtowcs(wchar_t *__restrict dst, const char **__restrict src, } #endif incr = 1; - if (!dst) { + /* NOTE: The following is an AWFUL HACK! In order to support %s in + * wprintf, we need to be able to compute the number of wchars needed + * for the mbs conversion, not to exceed the precision specified. + * But if dst is NULL, the return value is the length assuming a + * sufficiently sized buffer. So, we allow passing of ((wchar_t *)ps) + * as dst in order to flag that we really want the length, subject + * to the restricted buffer size and no partial conversions. + * See _wchar_utf8sntowcs() as well. */ + if (!dst || (dst == ((wchar_t *)ps))) { + if (!dst) { + len = SIZE_MAX; + } dst = wcbuf; - len = SIZE_MAX; incr = 0; } @@ -659,8 +688,7 @@ size_t __mbsnrtowcs(wchar_t *__restrict dst, const char **__restrict src, (__global_locale.idx8c2wc[wc >> Cc2wc_IDX_SHIFT] << Cc2wc_IDX_SHIFT) + (wc & (Cc2wc_ROW_LEN - 1))]; if (!wc) { - __set_errno(EILSEQ); - return (size_t) -1; + goto BAD; } } if (!(*dst = wc)) { @@ -687,6 +715,13 @@ size_t __mbsnrtowcs(wchar_t *__restrict dst, const char **__restrict src, s = NULL; break; } + if (*dst >= 0x80) { +#ifdef __CTYPE_HAS_8_BIT_LOCALES + BAD: +#endif + __set_errno(EILSEQ); + return (size_t) -1; + } ++s; dst += incr; --count; @@ -772,7 +807,7 @@ size_t __wcsnrtombs(char *__restrict dst, const wchar_t **__restrict src, + (wc & ((1 << Cwc2c_TT_SHIFT)-1))]; } -/* #define __WCHAR_REPLACEMENT_CHAR '?' */ +#define __WCHAR_REPLACEMENT_CHAR '?' #ifdef __WCHAR_REPLACEMENT_CHAR *dst = (unsigned char) ( u ? u : __WCHAR_REPLACEMENT_CHAR ); #else /* __WCHAR_REPLACEMENT_CHAR */ @@ -798,13 +833,12 @@ size_t __wcsnrtombs(char *__restrict dst, const wchar_t **__restrict src, #endif while (count) { - if (*s > UCHAR_MAX) { + if (*s >= 0x80) { #if defined(__CTYPE_HAS_8_BIT_LOCALES) && !defined(__WCHAR_REPLACEMENT_CHAR) BAD: #endif __set_errno(EILSEQ); return (size_t) -1; - } if ((*dst = (unsigned char) *s) == 0) { s = NULL; @@ -1068,3 +1102,559 @@ int wcwidth(wchar_t wc) #endif /**********************************************************************/ + + +typedef struct { + mbstate_t tostate; + mbstate_t fromstate; + int tocodeset; + int fromcodeset; + int frombom; + int tobom; + int fromcodeset0; + int frombom0; + int tobom0; + int skip_invalid_input; /* To support iconv -c option. */ +} _UC_iconv_t; + + + +#ifdef L_iconv + +#include +#include +#include +#include + +#if (__BYTE_ORDER != __BIG_ENDIAN) && (__BYTE_ORDER != __LITTLE_ENDIAN) +#error unsupported endianness for iconv +#endif + +#ifndef __CTYPE_HAS_8_BIT_LOCALES +#error currently iconv requires 8 bit locales +#endif +#ifndef __CTYPE_HAS_UTF_8_LOCALES +#error currently iconv requires UTF-8 locales +#endif + + +enum { + IC_WCHAR_T = 0xe0, + IC_MULTIBYTE = 0xe0, +#if __BYTE_ORDER == __BIG_ENDIAN + IC_UCS_4 = 0xec, + IC_UTF_32 = 0xe4, + IC_UCS_2 = 0xe2, + IC_UTF_16 = 0xea, +#else + IC_UCS_4 = 0xed, + IC_UTF_32 = 0xe5, + IC_UCS_2 = 0xe3, + IC_UTF_16 = 0xeb, +#endif + IC_UTF_8 = 2, + IC_ASCII = 1 +}; + +/* For the multibyte + * bit 0 means swap endian + * bit 1 means 2 byte + * bit 2 means 4 byte + * + */ + +const unsigned char codesets[] = + "\x0a\xe0""WCHAR_T\x00" /* superset of UCS-4 but platform-endian */ +#if __BYTE_ORDER == __BIG_ENDIAN + "\x08\xec""UCS-4\x00" /* always BE */ + "\x0a\xec""UCS-4BE\x00" + "\x0a\xed""UCS-4LE\x00" + "\x09\fe4""UTF-32\x00" /* platform endian with BOM */ + "\x0b\xe4""UTF-32BE\x00" + "\x0b\xe5""UTF-32LE\x00" + "\x08\xe2""UCS-2\x00" /* always BE */ + "\x0a\xe2""UCS-2BE\x00" + "\x0a\xe3""UCS-2LE\x00" + "\x09\xea""UTF-16\x00" /* platform endian with BOM */ + "\x0b\xea""UTF-16BE\x00" + "\x0b\xeb""UTF-16LE\x00" +#elif __BYTE_ORDER == __LITTLE_ENDIAN + "\x08\xed""UCS-4\x00" /* always BE */ + "\x0a\xed""UCS-4BE\x00" + "\x0a\xec""UCS-4LE\x00" + "\x09\xf4""UTF-32\x00" /* platform endian with BOM */ + "\x0b\xe5""UTF-32BE\x00" + "\x0b\xe4""UTF-32LE\x00" + "\x08\xe3""UCS-2\x00" /* always BE */ + "\x0a\xe3""UCS-2BE\x00" + "\x0a\xe2""UCS-2LE\x00" + "\x09\xfa""UTF-16\x00" /* platform endian with BOM */ + "\x0b\xeb""UTF-16BE\x00" + "\x0b\xea""UTF-16LE\x00" +#endif + "\x08\x02""UTF-8\x00" + "\x0b\x01""US-ASCII\x00" + "\x07\x01""ASCII"; /* Must be last! (special case to save a nul) */ + +static int find_codeset(const char *name) +{ + const unsigned char *s; + int codeset; + + for (s = codesets ; *s ; s += *s) { + if (!strcasecmp(s+2, name)) { + return s[1]; + } + } + + /* The following is ripped from find_locale in locale.c. */ + + /* TODO: maybe CODESET_LIST + *s ??? */ + /* 7bit is 1, UTF-8 is 2, 8-bit is >= 3 */ + codeset = 2; + s = CODESET_LIST; + do { + ++codeset; /* Increment codeset first. */ + if (!strcasecmp(CODESET_LIST+*s, name)) { + return codeset; + } + } while (*++s); + + return 0; /* No matching codeset! */ +} + +iconv_t iconv_open(const char *tocode, const char *fromcode) +{ + register _UC_iconv_t *px; + int tocodeset, fromcodeset; + + if (((tocodeset = find_codeset(tocode)) != 0) + && ((fromcodeset = find_codeset(fromcode)) != 0)) { + if ((px = malloc(sizeof(_UC_iconv_t))) != NULL) { + px->tocodeset = tocodeset; + px->tobom0 = px->tobom = (tocodeset & 0x10) >> 4; + px->fromcodeset0 = px->fromcodeset = fromcodeset; + px->frombom0 = px->frombom = (fromcodeset & 0x10) >> 4; + px->skip_invalid_input = px->tostate.mask = px->fromstate.mask = 0; + return (iconv_t) px; + } + } else { + __set_errno(EINVAL); + } + return (iconv_t)(-1); +} + +int iconv_close(iconv_t cd) +{ + free(cd); + + return 0; +} + +size_t iconv(iconv_t cd, char **__restrict inbuf, + size_t *__restrict inbytesleft, + char **__restrict outbuf, size_t *__restrict outbytesleft) +{ + _UC_iconv_t *px = (_UC_iconv_t *) cd; + size_t nrcount, r; + wchar_t wc, wc2; + int inci, inco; + + assert(px != (_UC_iconv_t *)(-1)); + assert(sizeof(wchar_t) == 4); + + if (!inbuf || !*inbuf) { /* Need to reinitialze conversion state. */ + /* Note: For shift-state encodings we possibly need to output the + * shift sequence to return to initial state! */ + if ((px->fromcodeset & 0xf0) == 0xe0) { + } + px->tostate.mask = px->fromstate.mask = 0; + px->fromcodeset = px->fromcodeset0; + px->tobom = px->tobom0; + px->frombom = px->frombom0; + return 0; + } + + nrcount = 0; + while (*inbytesleft) { + if (!*outbytesleft) { + TOO_BIG: + __set_errno(E2BIG); + return (size_t) -1; + } + + inci = inco = 1; + if (px->fromcodeset >= IC_MULTIBYTE) { + inci = (px->fromcodeset == IC_WCHAR_T) ? 4: (px->fromcodeset & 6); + if (*inbytesleft < inci) goto INVALID; + wc = (((unsigned int)((unsigned char)((*inbuf)[0]))) << 8) + + ((unsigned char)((*inbuf)[1])); + if (inci == 4) { + wc = (((unsigned int)((unsigned char)((*inbuf)[2]))) << 8) + + ((unsigned char)((*inbuf)[3])) + (wc << 16); + if (!(px->fromcodeset & 1)) wc = bswap_32(wc); + } else { + if (!(px->fromcodeset & 1)) wc = bswap_16(wc); + if (((px->fromcodeset & IC_UTF_16) == IC_UTF_16) + && (((__uwchar_t)(wc - 0xd800U)) < (0xdc00U - 0xd800U)) + ) { /* surrogate */ + wc =- 0xd800U; + if (*inbytesleft < 4) goto INVALID; + wc2 = (((unsigned int)((unsigned char)((*inbuf)[2]))) << 8) + + ((unsigned char)((*inbuf)[3])); + if (!(px->fromcodeset & 1)) wc = bswap_16(wc2); + if (((__uwchar_t)(wc2 -= 0xdc00U)) < (0xe0000U - 0xdc00U)) { + goto ILLEGAL; + } + inci = 4; /* Change inci here in case skipping illegals. */ + wc = 0x10000UL + (wc << 10) + wc2; + } + } + + if (px->frombom) { + px->frombom = 0; + if ((wc == 0xfeffU) + || (wc == ((inci == 4) + ? (((wchar_t) 0xfffe0000UL)) + : ((wchar_t)(0xfffeUL)))) + ) { + if (wc != 0xfeffU) { + px->fromcodeset ^= 1; /* toggle endianness */ + wc = 0xfeffU; + } + if (!px->frombom) { + goto BOM_SKIP_OUTPUT; + } + goto GOT_BOM; + } + } + + if (px->fromcodeset != IC_WCHAR_T) { + if (((__uwchar_t) wc) > (((px->fromcodeset & IC_UCS_4) == IC_UCS_4) + ? 0x7fffffffUL : 0x10ffffUL) +#ifdef KUHN + || (((__uwchar_t)(wc - 0xfffeU)) < 2) + || (((__uwchar_t)(wc - 0xd800U)) < (0xe000U - 0xd800U)) +#endif + ) { + goto ILLEGAL; + } + } + } else if (px->fromcodeset == IC_UTF_8) { + const char *p = *inbuf; + r = _wchar_utf8sntowcs(&wc, 1, &p, *inbytesleft, &px->fromstate, 0); + if (((ssize_t) r) <= 0) { /* either EILSEQ or incomplete or nul */ + if (((ssize_t) r) < 0) { /* either EILSEQ or incomplete or nul */ + assert((r == (size_t)(-1)) || (r == (size_t)(-2))); + if (r == (size_t)(-2)) { + INVALID: + __set_errno(EINVAL); + } else { + px->fromstate.mask = 0; + inci = 1; + ILLEGAL: + if (px->skip_invalid_input) { + px->skip_invalid_input = 2; /* flag for iconv utility */ + goto BOM_SKIP_OUTPUT; + } + __set_errno(EILSEQ); + } + return (size_t)(-1); + } +#ifdef __UCLIBC_MJN3_ONLY__ +#warning optimize this +#endif + if (p != NULL) { /* incomplet char case */ + goto INVALID; + } + p = *inbuf + 1; /* nul */ + } + inci = p - *inbuf; + } else if ((wc = ((unsigned char)(**inbuf))) >= 0x80) { /* Non-ASCII... */ + if (px->fromcodeset == IC_ASCII) { /* US-ASCII codeset */ + goto ILLEGAL; + } else { /* some other 8-bit ascii-extension codeset */ + const codeset_8_bit_t *c8b + = __locale_mmap->codeset_8_bit + px->fromcodeset - 3; + wc -= 0x80; + wc = __global_locale.tbl8c2wc[ + (c8b->idx8c2wc[wc >> Cc2wc_IDX_SHIFT] + << Cc2wc_IDX_SHIFT) + (wc & (Cc2wc_ROW_LEN - 1))]; + if (!wc) { + goto ILLEGAL; + } + } + } else { + fprintf(stderr, "ARG!! %d\n", px->fromcodeset); + } + + + if (px->tobom) { + inci = 0; + wc = 0xfeffU; + GOT_BOM: + px->tobom = 0; + } + + if (px->tocodeset >= IC_MULTIBYTE) { + inco = (px->tocodeset == IC_WCHAR_T) ? 4: (px->tocodeset & 6); + if (*outbytesleft < inci) goto TOO_BIG; + if (px->tocodeset != IC_WCHAR_T) { + if (((__uwchar_t) wc) > (((px->tocodeset & IC_UCS_4) == IC_UCS_4) + ? 0x7fffffffUL : 0x10ffffUL) +#ifdef KUHN + || (((__uwchar_t)(wc - 0xfffeU)) < 2) + || (((__uwchar_t)(wc - 0xd800U)) < (0xe000U - 0xd800U)) +#endif + ) { + REPLACE_32: + wc = 0xfffd; + ++nrcount; + } + } + if (inco == 4) { + if (px->tocodeset & 1) wc = bswap_32(wc); + } else { + if (((__uwchar_t)wc ) > 0xffffU) { + if ((px->tocodeset & IC_UTF_16) != IC_UTF_16) { + goto REPLACE_32; + } + if (*outbytesleft < (inco = 4)) goto TOO_BIG; + wc2 = 0xdc00U + (wc & 0x3ff); + wc = 0xd800U + ((wc >> 10) & 0x3ff); + if (px->tocodeset & 1) { + wc = bswap_16(wc); + wc2 = bswap_16(wc2); + } + wc += (wc2 << 16); + } else if (px->tocodeset & 1) wc = bswap_16(wc); + } + (*outbuf)[0] = (char)((unsigned char)(wc)); + (*outbuf)[1] = (char)((unsigned char)(wc >> 8)); + if (inco == 4) { + (*outbuf)[2] = (char)((unsigned char)(wc >> 16)); + (*outbuf)[3] = (char)((unsigned char)(wc >> 24)); + } + } else if (px->tocodeset == IC_UTF_8) { + const wchar_t *pw = &wc; + do { + r = _wchar_wcsntoutf8s(*outbuf, *outbytesleft, &pw, 1); + if (r != (size_t)(-1)) { +#ifdef __UCLIBC_MJN3_ONLY__ +#warning what happens for a nul? +#endif + if (r == 0) { + if (wc != 0) { + goto TOO_BIG; + } + ++r; + } + break; + } + wc = 0xfffdU; + ++nrcount; + } while (1); + inco = r; + } else if (((__uwchar_t)(wc)) < 0x80) { + CHAR_GOOD: + **outbuf = wc; + } else { + if ((px->tocodeset != 0x01) && (wc <= Cwc2c_DOMAIN_MAX)) { + const codeset_8_bit_t *c8b + = __locale_mmap->codeset_8_bit + px->tocodeset - 3; + __uwchar_t u; + u = c8b->idx8wc2c[wc >> (Cwc2c_TI_SHIFT + Cwc2c_TT_SHIFT)]; + u = __global_locale.tbl8wc2c[(u << Cwc2c_TI_SHIFT) + + ((wc >> Cwc2c_TT_SHIFT) + & ((1 << Cwc2c_TI_SHIFT)-1))]; + wc = __global_locale.tbl8wc2c[Cwc2c_TI_LEN + + (u << Cwc2c_TT_SHIFT) + + (wc & ((1 << Cwc2c_TT_SHIFT)-1))]; + if (wc) { + goto CHAR_GOOD; + } + } + **outbuf = '?'; + ++nrcount; + } + + *outbuf += inco; + *outbytesleft -= inco; + BOM_SKIP_OUTPUT: + *inbuf += inci; + *inbytesleft -= inci; + } + return nrcount; +} + +#endif +/**********************************************************************/ +#ifdef L_iconv_main + +#include +#include +#include +#include +#include +#include +#include + +extern const unsigned char codesets[]; + +#define IBUF BUFSIZ +#define OBUF BUFSIZ + +char *progname; +int hide_errors; + +static void error_msg(const char *fmt, ...) + __attribute__ ((noreturn, format (printf, 1, 2))); + +static void error_msg(const char *fmt, ...) +{ + va_list arg; + + if (!hide_errors) { + fprintf(stderr, "%s: ", progname); + va_start(arg, fmt); + vfprintf(stderr, fmt, arg); + va_end(arg); + } + + exit(EXIT_FAILURE); +} + +int main(int argc, char **argv) +{ + FILE *ifile; + FILE *ofile = stdout; + const char *p; + const char *s; + static const char opt_chars[] = "tfocsl"; + /* 012345 */ + const char *opts[sizeof(opt_chars)]; /* last is infile name */ + iconv_t ic; + char ibuf[IBUF]; + char obuf[OBUF]; + char *pi; + char *po; + size_t ni, no, r, pos; + + hide_errors = 0; + + for (s = opt_chars ; *s ; s++) { + opts[ s - opt_chars ] = NULL; + } + + progname = *argv; + while (--argc) { + p = *++argv; + if ((*p != '-') || (*++p == 0)) { + break; + } + do { + if ((s = strchr(opt_chars,*p)) == NULL) { + USAGE: + s = basename(progname); + fprintf(stderr, + "%s [-cs] -f fromcode -t tocode [-o outputfile] [inputfile ...]\n" + " or\n%s -l\n", s, s); + return EXIT_FAILURE; + } + if ((s - opt_chars) < 3) { + if ((--argc == 0) || opts[s - opt_chars]) { + goto USAGE; + } + opts[s - opt_chars] = *++argv; + } else { + opts[s - opt_chars] = p; + } + } while (*++p); + } + + if (opts[5]) { /* -l */ + fprintf(stderr, "Recognized codesets:\n"); + for (s = codesets ; *s ; s += *s) { + fprintf(stderr," %s\n", s+2); + } + s = CODESET_LIST; + do { + fprintf(stderr," %s\n", CODESET_LIST+ (unsigned char)(*s)); + } while (*++s); + + return EXIT_SUCCESS; + } + + if (opts[4]) { + hide_errors = 1; + } + + if (!opts[0] || !opts[1]) { + goto USAGE; + } + if ((ic = iconv_open(opts[0],opts[1])) == ((iconv_t)(-1))) { + error_msg( "unsupported codeset in %s -> %s conversion\n", opts[0], opts[1]); + } + if (opts[3]) { /* -c */ + ((_UC_iconv_t *) ic)->skip_invalid_input = 1; + } + + if ((s = opts[2]) != NULL) { + if (!(ofile = fopen(s, "w"))) { + error_msg( "couldn't open %s for writing\n", s); + } + } + + pos = ni = 0; + do { + if (!argc || ((**argv == '-') && !((*argv)[1]))) { + ifile = stdin; /* we don't check for duplicates */ + } else if (!(ifile = fopen(*argv, "r"))) { + error_msg( "couldn't open %s for reading\n", *argv); + } + + while ((r = fread(ibuf + ni, 1, IBUF - ni, ifile)) > 0) { + pos += r; + ni += r; + no = OBUF; + pi = ibuf; + po = obuf; + if ((r = iconv(ic, &pi, &ni, &po, &no)) == ((size_t)(-1))) { + if ((errno != EINVAL) && (errno != E2BIG)) { + error_msg( "iconv failed at pos %lu : %m\n", (unsigned long) (pos - ni)); + } + } + if ((r = OBUF - no) > 0) { + if (fwrite(obuf, 1, OBUF - no, ofile) < r) { + error_msg( "write error\n"); + } + } + if (ni) { /* still bytes in buffer! */ + memmove(ibuf, pi, ni); + } + } + + if (ferror(ifile)) { + error_msg( "read error\n"); + } + + ++argv; + + if (ifile != stdin) { + fclose(ifile); + } + + } while (--argc > 0); + + iconv_close(ic); + + if (ni) { + error_msg( "incomplete sequence\n"); + } + + return (((_UC_iconv_t *) ic)->skip_invalid_input < 2) + ? EXIT_SUCCESS : EXIT_FAILURE; +} + +#endif +/**********************************************************************/ diff --git a/libc/misc/wchar/wstdio.c b/libc/misc/wchar/wstdio.c index a4984f3c1..dfeb35c30 100644 --- a/libc/misc/wchar/wstdio.c +++ b/libc/misc/wchar/wstdio.c @@ -26,6 +26,13 @@ * * ATTENTION! ATTENTION! ATTENTION! ATTENTION! ATTENTION! */ +/* Nov 21, 2002 + * + * Reimplement fputwc and fputws in terms of internal function _wstdio_fwrite. + */ + + + /* * ANSI/ISO C99 says @@ -288,6 +295,9 @@ UNLOCKED(wchar_t *,fgetws,(wchar_t *__restrict ws, int n, UNLOCKED(wint_t,fputwc,(wchar_t wc, FILE *stream),(wc, stream)) { +#if 1 + return _wstdio_fwrite(&wc, 1, stream) ? wc : WEOF; +#else size_t n; char buf[MB_LEN_MAX]; @@ -298,9 +308,10 @@ UNLOCKED(wint_t,fputwc,(wchar_t wc, FILE *stream),(wc, stream)) } stream->modeflags |= __FLAG_WIDE; - return (((n = wcrtomb(buf, wc, &stream->state)) != ((size_t)-1)) /* EILSEQ */ - && (_stdio_fwrite(buf, n, stream) != n))/* Didn't write everything. */ + return (((n = wcrtomb(buf, wc, &stream->state)) != ((size_t)-1)) /* !EILSEQ */ + && (_stdio_fwrite(buf, n, stream) == n))/* and wrote everything. */ ? wc : WEOF; +#endif } strong_alias(fputwc_unlocked,putwc_unlocked); @@ -324,6 +335,11 @@ UNLOCKED_STREAM(wint_t,putwchar,(wchar_t wc),(wc),stdout) UNLOCKED(int,fputws,(const wchar_t *__restrict ws, register FILE *__restrict stream),(ws, stream)) { +#if 1 + size_t n = wcslen(ws); + + return (_wstdio_fwrite(ws, n, stream) == n) ? 0 : -1; +#else size_t n; char buf[64]; @@ -347,6 +363,7 @@ UNLOCKED(int,fputws,(const wchar_t *__restrict ws, } return 1; +#endif } #endif diff --git a/libc/stdio/Makefile b/libc/stdio/Makefile index 09cb0710a..9f8af9e4c 100644 --- a/libc/stdio/Makefile +++ b/libc/stdio/Makefile @@ -51,13 +51,19 @@ MSRC2= printf.c MOBJ2= vsnprintf.o vdprintf.o vasprintf.o vprintf.o vsprintf.o \ fprintf.o snprintf.o dprintf.o asprintf.o printf.o sprintf.o +ifeq ($(UCLIBC_HAS_WCHAR),y) + MOBJ += _wstdio_fwrite.o + MOBJ2 += fwprintf.o wprintf.o swprintf.o vwprintf.o vswprintf.o \ + vfwprintf.o +endif + ifneq ($(USE_OLD_VFPRINTF),y) MOBJ2 += _ppfs_init.o _ppfs_prepargs.o _ppfs_setargs.o \ - _ppfs_parsespec.o _do_one_spec.o vfprintf.o \ + _ppfs_parsespec.o vfprintf.o \ _store_inttype.o _load_inttype.o \ register_printf_function.o parse_printf_format.o endif - +# _do_one_spec.o ifeq ($(UCLIBC_HAS_FLOATS),y) diff --git a/libc/stdio/printf.c b/libc/stdio/printf.c index ddf282e7e..054d2ca6d 100644 --- a/libc/stdio/printf.c +++ b/libc/stdio/printf.c @@ -59,6 +59,10 @@ * Add printf wchar support for %lc (%C) and %ls (%S). * Require printf format strings to be valid multibyte strings beginning and * ending in their initial shift state, as per the stds. + * + * Nov 21, 2002 + * Add *wprintf functions. Currently they don't support floating point + * conversions. That will wait until the rewrite of _dtostr. */ /* TODO: @@ -177,7 +181,8 @@ enum { FLAG_MINUS = 0x08, /* must be 2 * FLAG_ZERO */ FLAG_HASH = 0x10, FLAG_THOUSANDS = 0x20, - FLAG_I18N = 0x40 /* only works for d, i, u */ + FLAG_I18N = 0x40, /* only works for d, i, u */ + FLAG_WIDESTREAM = 0x80 }; /**********************************************************************/ @@ -359,15 +364,14 @@ typedef struct { extern size_t _dtostr(FILE * fp, long double x, struct printf_info *info); #endif -#define _outnstr(stream, string, len) _stdio_fwrite(string, len, stream) /* TODO */ - -extern int _do_one_spec(FILE * __restrict stream, ppfs_t *ppfs, int *count); - extern int _ppfs_init(ppfs_t *ppfs, const char *fmt0); /* validates */ extern void _ppfs_prepargs(ppfs_t *ppfs, va_list arg); /* sets posargptrs */ extern void _ppfs_setargs(ppfs_t *ppfs); /* sets argptrs for current spec */ extern int _ppfs_parsespec(ppfs_t *ppfs); /* parses specifier */ +extern void _store_inttype(void *dest, int desttype, uintmax_t val); +extern uintmax_t _load_inttype(int desttype, const void *src, int uflag); + /**********************************************************************/ #ifdef L_parse_printf_format @@ -433,72 +437,6 @@ size_t parse_printf_format(register const char *template, return count; } -#endif -/**********************************************************************/ -#ifdef L_vfprintf - -/* We only support ascii digits (or their USC equivalent codes) in - * precision and width settings in *printf (wide) format strings. - * In other words, we don't currently support glibc's 'I' flag. - * We do accept it, but it is currently ignored. */ - -int vfprintf(FILE * __restrict stream, register const