diff options
author | Roland McGrath <roland@gnu.org> | 1996-07-09 23:17:59 +0000 |
---|---|---|
committer | Roland McGrath <roland@gnu.org> | 1996-07-09 23:17:59 +0000 |
commit | 07a4742f9ec2b0587c0d488bb65da2a6faa50fed (patch) | |
tree | 2b10e82cad4053bcbe27447bc3d444f6d2aef5e7 /wcsmbs/wcrtomb.c | |
parent | c150923988933b5db75a974d4cc08cd7f7aaf3dc (diff) | |
download | glibc-07a4742f9ec2b0587c0d488bb65da2a6faa50fed.tar.gz glibc-07a4742f9ec2b0587c0d488bb65da2a6faa50fed.tar.xz glibc-07a4742f9ec2b0587c0d488bb65da2a6faa50fed.zip |
* posix/glob.h (__glob_opendir_hook, __glob_readdir_hook, cvs/libc-960710
__glob_closedir_hook): Remove decls. * sysdeps/generic/machine-gmon.h: Declare mcount_internal. * sysdeps/unix/inet/syscalls.list: Define __ names with weak aliases for send and connect syscalls. * socket/sys/socket.h: New file, taken from non-sysdep parts of linux/sys/socket.h; break sysdeps parts out into socketbits.h. Declare __ names for send and connect. * sysdeps/generic/socketbits.h: New file. * sysdeps/unix/sysv/linux/socketbits.h: New file. * sysdeps/unix/sysv/linux/sys/socket.h: File removed. * sysdeps/generic/sys/socket.h: File removed. * sysdeps/mach/hurd/connect.c: Define __ name and weak alias. * sysdeps/mach/hurd/send.c: Likewise.
Diffstat (limited to 'wcsmbs/wcrtomb.c')
-rw-r--r-- | wcsmbs/wcrtomb.c | 62 |
1 files changed, 42 insertions, 20 deletions
diff --git a/wcsmbs/wcrtomb.c b/wcsmbs/wcrtomb.c index 9069fb105c..eb007a69b9 100644 --- a/wcsmbs/wcrtomb.c +++ b/wcsmbs/wcrtomb.c @@ -1,6 +1,6 @@ /* Copyright (C) 1996 Free Software Foundation, Inc. This file is part of the GNU C Library. -Contributed by Ulrich Drepper, <drepper@gnu.ai.mit.edu> +Contributed by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1996. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as @@ -24,46 +24,68 @@ Boston, MA 02111-1307, USA. */ #define EILSEQ EINVAL #endif +static const wchar_t encoding_mask[] = +{ + ~0x7ff, ~0xffff, ~0x1fffff, ~0x3ffffff +}; + +static const unsigned char encoding_byte[] = +{ + 0xc0, 0xe0, 0xf0, 0xf8, 0xfc +}; +/* The state is for this UTF8 encoding not used. */ static mbstate_t internal; size_t -wcrtomb (s, wc, ps) - char *s; - wchar_t wc; - mbstate_t *ps; +wcrtomb (char *s, wchar_t wc, mbstate_t *ps) { char fake[1]; + size_t written = 0; if (ps == NULL) ps = &internal; - /*************************************************************\ - |* This is no complete implementation. While the multi-byte *| - |* character handling is not finished this will do. *| - \*************************************************************/ - if (s == NULL) { s = fake; wc = L'\0'; } - if (wc == L'\0') + /* Store the UTF8 representation of WC. */ + if (wc < 0 || wc > 0x7fffffff) { - /* FIXME Write any shift sequence to get to *PS == NULL. */ - *ps = 0; - *s = '\0'; + /* This is no correct ISO 10646 character. */ + errno = EILSEQ; + return (size_t) -1; + } + + if (wc < 0x80) + { + /* It's a one byte sequence. */ + if (s != NULL) + *s = (char) wc; return 1; } - /* FIXME For now we don't handle real multi-byte encodings. */ - if ((wc & ~0xff) != 0) + for (written = 2; written < 6; ++written) + if ((wc & encoding_mask[written - 2]) == 0) + break; + + if (s != NULL) { - errno = EILSEQ; - return (size_t) -1; + size_t cnt = written; + s[0] = encoding_byte[cnt - 2]; + + --cnt; + do + { + s[cnt] = 0x80 | (wc & 0x3f); + wc >>= 6; + } + while (--cnt > 0); + s[0] |= wc; } - *s = (char) wc; - return 1; + return written; } |