Commit Graph

34 Commits

Author SHA1 Message Date
Andrey A. Chernov
70e7b0732a Back out not human readable optimization in prev. commit which shown
to generate 3bytes longer opcode.
2007-11-01 03:18:02 +00:00
Andrey A. Chernov
5dfb7ce5fc Micro-optimization of prev. commit, change
(_c < 0 || _c >= 128) to (_c & ~0x7F)
2007-10-27 22:32:27 +00:00
Andrey A. Chernov
eb75c0a0d0 Small improvement for __isctype(): don't use __mb_sb_limit but
hardcode 128 here instead, since default locale never define anything
above 127 char.
2007-10-27 09:32:32 +00:00
Andrey A. Chernov
367ed4e13d The problem is: currently our single byte ctype(3) functions are broken
for wide characters locales in the argument range >= 0x80 - they may
return false positives.

Example 1: for UTF-8 locale we currently have:
iswspace(0xA0)==1 and isspace(0xA0)==1
(because iswspace() and isspace() are the same code)
but must have
iswspace(0xA0)==1 and isspace(0xA0)==0
(because there is no such character and all others in the range
0x80..0xff for the UTF-8 locale, it keeps ASCII only in the single byte
range because our internal wchar_t representation for UTF-8 is UCS-4).

Example 2: for all wide character locales isalpha(arg) when arg > 0xFF may
return false positives (must be 0).
(because iswalpha() and isalpha() are the same code)

This change address this issue separating single byte and wide ctype
and also fix iswascii() (currently iswascii() is broken for
arguments > 0xFF).
This change is 100% binary compatible with old binaries.

Reviewied by: i18n@
2007-10-13 16:28:22 +00:00
Tim J. Robbins
17211a5f9a Let GCC know that ___runetype(), ___tolower() and ___toupper() are pure
functions, allowing it to generate better code for the <ctype.h> and
<wctype.h> functions. For example, it can now keep _CurrentRuneLocale
in a register across calls to these functions, and can delete calls to
___runetype() if the result is already known or not used.
2004-08-21 07:00:40 +00:00
Tim J. Robbins
de6c9c9d5b Implement wcwidth() as an inline function. 2004-08-12 12:19:11 +00:00
Tim J. Robbins
16133e1530 Move some internal macros and inlines from ctype.h to a new file, _ctype.h,
which has been repo-copied from ctype.h. This will allow us to remove
namespace pollution from <wctype.h> and to make wcwidth() an inline function
without introducing more pollution.
2004-08-12 09:33:47 +00:00
Tim J. Robbins
838faf1c74 Now that <runetype.h> no longer brings in namespace pollution,
bring back the inline functions for the !__BSD_VISIBLE case.
2004-06-23 07:11:39 +00:00
Tim J. Robbins
ddc1eded85 Prefix the names of members of _RuneLocale and its sub-structures
with ``__'' to avoid polluting the namespace. This doesn't change the
documented rune interface at all, but breaks applications that accessed
_RuneLocale directly.
2004-06-23 07:01:44 +00:00
Tim J. Robbins
e267671499 Make isblank() visible in the C99 namespace.
PR:		63371
Submitted by:	Stefan Farfeleder
2004-04-21 13:25:55 +00:00
Mike Barcroft
86eedb4ef4 Solve the <runetype.h> pollution problem by disabling inline
optimizations when a standard has been requested, except when the
inline optimizations are also specifically requested.
2002-09-09 05:38:05 +00:00
Mike Barcroft
423eb945c0 o Fix namespace scope issues in <ctype.h> by using the relatively new
visibility primitives.
o Implement _tolower() and _toupper() POSIX.1-2001 (XSI) macros in
  <ctype.h>.
o Reduce pollution in <runetype.h> by removing typedefs and using
  implementation namespaced types.
o Add a typedef in <rune.h> to compensate for <runetype.h> losing its
  typedefs.

Reviewed by:	bde
2002-09-06 04:22:54 +00:00
Mike Barcroft
abbd890233 o Merge <machine/ansi.h> and <machine/types.h> into a new header
called <machine/_types.h>.
o <machine/ansi.h> will continue to live so it can define MD clock
  macros, which are only MD because of gratuitous differences between
  architectures.
o Change all headers to make use of this.  This mainly involves
  changing:
    #ifdef _BSD_FOO_T_
    typedef	_BSD_FOO_T_	foo_t;
    #undef _BSD_FOO_T_
    #endif
  to:
    #ifndef _FOO_T_DECLARED
    typedef	__foo_t	foo_t;
    #define	_FOO_T_DECLARED
    #endif

Concept by:	bde
Reviewed by:	jake, obrien
2002-08-21 16:20:02 +00:00
Andrey A. Chernov
153bb33daf Move internal defines from here to libc/locale/wcwidth.c 2002-08-19 08:58:51 +00:00
Andrey A. Chernov
dcc3da5862 Properly define SWIDTH1, add autowidth (was SWIDTH1) 2002-08-19 08:50:41 +00:00
Michael C . Wu
c3121a34a0 Add swidth definition for wchar stuff
Submitted by:	clkao@clkao.org
Reviewed by:	keichii
Obtained from:	NetBSD
MFC after:	1 month
2002-08-16 13:42:59 +00:00
Warner Losh
bb28f3c29b Breath deep and take __P out of the system include files.
# This appears to not break X11, but I'm having problems compiling the
# glide part of the server with or without this patch, so I can't tell
# for sure.
2002-03-23 17:24:55 +00:00
Rob Braun
509853bbfd Makes __istype() an inline (and a non inlined) function to avoid C++
compile issues.  std::isspace(' ') was expanding to std::(!!_maskrune...)
which would cause a C++ compile error.  Making __istype() an inline
causes the expansion to be std::__istype() instead, which is valid.

Reviewed by: jkh
2002-02-05 06:21:34 +00:00
David E. O'Brien
f506ed7467 There is a problem in that one cannot use ctype.h at the same time as parts
of the C++ stdlib.  Our ctype.h uses symbols of the form _<X> to denote the
various character classes.  Our ctype.h also extends the usual ctype.h
offering by adding the "_T" (special) class.  Problem is parts of the STL
also use the symbol "_T" as its parameterized type.  These two uses are
incompatible.

Thus change the form of the symbols used in ctype to something that fixes
the current problem and is less likely to cause conflicts in the future.

Requested by:	Tomoaki NISHIYAMA <tomoaki@biol.s.u-tokyo.ac.jp>
Ok'ed by:	JKH
2000-02-08 07:43:26 +00:00
Alexey Zelkin
94da96bf18 Back up following macros by functions: ishexnumber, isideogram, isnumber,
isphonogram, isrune, isspecial. Fix ordering.

Reviewed by: bde
1999-12-17 15:12:21 +00:00
Andrey A. Chernov
659ad3a4a1 Oops, fix typo in istype definition 1997-09-27 04:30:50 +00:00
Andrey A. Chernov
917065108a 1) Bring '!!' back since we wan't overflow even a char
2) Revive __maskrune, just return mask now
3) Express __istype via !!__maskrune
4) Use __maskrune in digittoint
1997-09-27 04:22:47 +00:00
Andrey A. Chernov
80fd925122 Oops, wrong version of digittoint fix was commited, step back to v1.8
instead just remove '!!' to make digittoint() work and re-arrange
things a little
1997-09-25 22:44:21 +00:00
Andrey A. Chernov
9f05e73cd2 Fix digittoint broken long ago
Redesign to allow digittoint work for runes too
__maskrune removed, __istype become a macro, __runeflags added
1997-09-25 22:27:45 +00:00
Peter Wemm
985ae608d7 Bandaid for the build-breaking reference to _BSD_RUNE_T_. I'm not sure
that this is right, but the old reference defaintely was not.
1997-09-25 03:06:19 +00:00
Julian Elischer
16f76e6f06 Submitted by: Sin'ichiro MIYATANI / Phase One, Inc <siu@phaseone.co.jp>
Basic support for the Shift JIS encoding of japanese.
(and one tiny typo fixed in a comment)
1997-09-24 20:38:12 +00:00
Bruce Evans
cdd84b0211 Fixed longstanding namespace convolution involving rune_t vs wchar_t.
If _ANSI_SOURCE or _POSIX_SOURCE is defined, then <ctype.h> had to
be included before <stddef.h> or <stdlib.h> to get rune_t declared.
Now rune_t is declared perfectly bogusly in all cases when <ctype.h>
is included.

This change breaks similar (but more convoluted) convolutions in the
stddef.h in gcc distributions.  Ports of gcc should avoid using the
gcc headers.
1996-05-01 00:40:10 +00:00
Andrey A. Chernov
f905bd5001 Slightly simplify inlined functions 1996-03-25 13:46:21 +00:00
Andrey A. Chernov
15b31aa05a Fix isspecial/isphonogram, they was swapped
Remove EOF hack, now it is recognized per ANSI/POSIX
Add upper bounds check
Handle all negative chars inside locale functions
1995-11-03 12:25:14 +00:00
Bruce Evans
18c34920f8 Reviewed by: ache and wollman (long ago)
Fix numerous ANSI conformance bugs and other nits.

ctype.h:
o There were no prototypes behind the macros (conformance bug).
o isascii() didn't have enough parentheses (plain bug).
o tolower() and toupper were always static inline (conformance
  bug?  You could undef them and take their address, but this
  gave different addresses in different modules.  You couldn't
  undef them and declare them (correctly) again).  <stdio.h>'s
  treatment of putc() shows one way to handle this problem,
  but it only works because the putc() macro is allowed to
  reevaluate its args.  I used a hack controlled by
  _EXTERNALIZE_CTYPE_INLINES_ to get <ctype.h> to generate the
  code (the previous hack involving _ANSI_LIBRARY_ goes away).
  This has the advantage that the core of the functions is only
  written down once and the disadvantage that another layer of
  functions is required.  The extra layer goes away if inline
  functions are used, leaving only the problem of understanding
  why there are functions named toupper(), __toupper and
  ___toupper() as well as a macro named toupper.
o Nothing seems to define _USE_CTYPE_LIBRARY_.  Eliminate it
o Let the user set _USE_CTYPE_INLINE_ and _DONT_USE_CTYPE_INLINE_
  for full control over inlining.
o The args for the inline functions didn't have enough
  underscores (conformance bug).
o The formatting and ordering was inconsistent (style bug).
o TODO: fix conformance bugs brought by including <runetype.h>.
1995-04-07 11:43:40 +00:00
Andrey A. Chernov
ead39f6e45 Move inline args out of user namespace.
Obtained from: 1.x
1995-04-07 09:56:10 +00:00
Andrey A. Chernov
09b133a008 Remove EOF handling after Bruce explanation. This step returns
to 4.4 way to not allow EOF in ctype and now all signed chars
(including '\377' which becomes EOF) converted to (unsigned char) properly.
1994-10-09 11:18:44 +00:00
Andrey A. Chernov
262fb20771 Handle EOF case in all macros by ANSI standard.
Cast all ints < 0 to (unsigned char) to fix common problem
with sign extention on signed char.
1994-10-08 17:36:44 +00:00
Rodney W. Grimes
59deaec541 BSD 4.4 Lite Include Sources 1994-05-24 09:57:34 +00:00