about summary refs log tree commit diff
path: root/Src
Commit message (Collapse)AuthorAgeFilesLines
* Optimized away a recently-added call to strlen().Wayne Davison2006-02-201-1/+1
|
* Made the realocation of the line buffer in add_match_data() use aWayne Davison2006-02-191-8/+9
| | | | little more optimal size. Also tweaked the code just a little.
* 22284: add --disable-dynamic-nss configure switch to disable use of ↵Clint Adams2006-02-196-13/+38
| | | | getpw*(), getgr*(), initgroups() functions.
* The code in add_match_data() that looks for invalid charactersWayne Davison2006-02-181-25/+26
| | | | now properly parses the string as metafied.
* Made ${(A)=name:=word} (which is an array assignment) not split onWayne Davison2006-02-161-14/+10
| | | | quoted whitespace, just like the new ${1+"$@"} handling.
* Fixed another ${name+word} problem when word consisted of anWayne Davison2006-02-151-29/+19
| | | | array followed by a scalar. Also updated several comments.
* - Fixed: setopt shwordsplit; set 1 '2 3' 4; print -l ${1+"$@"}Wayne Davison2006-02-151-36/+89
| | | | | | - Fixed: setopt shwordsplit; foo='1 2'; print -l ${==foo+$foo} - Fixed: emulate sh; touch '1 2'; print -l ${1:-*\ *} - Improved multsub() and its comments.
* Replaced a flawed look-behind algorithm for backslash detectionWayne Davison2006-02-131-8/+8
| | | | | with one that looks forward (avoiding an accidental quoting of a char after a "\\" sequence).
* Made add_match_data() look for invalid multibyte chars in theWayne Davison2006-02-131-4/+61
| | | | names and substitute $'\123' sequences for them.
* Call makecommaspecial() when we notice we're completing insideWayne Davison2006-02-131-0/+5
| | | | an open brace list.
* Added the makecommaspecial() routine that completion uses whenWayne Davison2006-02-131-0/+14
| | | | it is completing filenames into an open brace list.
* Call ZS_memchr() instead of plain memchr() so that suffix-removalWayne Davison2006-02-091-2/+2
| | | | works again.
* Define ZS_memchr() as either wmemchr() or memchr().Wayne Davison2006-02-091-0/+2
|
* 22200: process text wasn't unmetafied for output.Peter Stephenson2006-02-071-4/+18
|
* 22198: do always set HOME in native emulationPeter Stephenson2006-02-072-5/+25
|
* Added the missing return value to a newly-added "return".Wayne Davison2006-02-071-1/+1
|
* 22195: don't set HOME in the shellPeter Stephenson2006-02-065-16/+20
|
* 22169: hook function zle-keymap-select called when $KEYMAP changesPeter Stephenson2006-01-281-1/+17
|
* Tweaked the code that handles %D so that: (1) we don't free() a NULLWayne Davison2006-01-171-4/+3
| | | | | pointer (we use zsfree() instead), and (2) we avoid a rare bug that could call free() on same memory twice (e.g. "%D{}%D").
* Got rid of unused-variable compiler warning when compiling aWayne Davison2006-01-131-1/+1
| | | | non-multibyte zsh.
* Got rid of some superfluous STOUC() calls (such as the ones I addedWayne Davison2006-01-134-13/+13
| | | | for the arg of calls to nicechar(), which doesn't need this).
* A better version of my last change.Wayne Davison2006-01-131-12/+4
|
* Metafy the output of the ztrftime() string since that's what theWayne Davison2006-01-131-1/+12
| | | | prompt expects and some date names might not be ASCII.
* Values being stored in an array or a hash needed to be metafiedWayne Davison2006-01-131-5/+5
| | | | (using META_DUP) instead of just duplicated via ztrdup().
* If we are saving the ztrftime() value in a variable, we now metafyWayne Davison2006-01-131-1/+1
| | | | the string (using META_DUP) instead of duplicating it via ztrdup().
* The new "eol" var was being set in singlerefresh(), but not accessed.Wayne Davison2006-01-121-1/+1
|
* - When mbrtowc() returns -2 when given all the remaining chars in aWayne Davison2006-01-121-21/+30
| | | | | | | | string, set an end-of-line flag and avoid calling mbrtowc() again for any of the incomplete characters that remain in the string. - Use "mbs" for the multi-byte state variable name (for consistency). - Use the new MB_INVALID and MB_INCOMPLETE defines for the size_t -1 and -2 values (respectively).
* - Use "mbs" for the multi-byte state variable name (for consistency).Wayne Davison2006-01-121-5/+5
| | | | | - Use the new MB_INVALID and MB_INCOMPLETE defines for the size_t -1 and -2 values (respectively).
* - When mbrtowc() returns -2 when given all the remaining chars in aWayne Davison2006-01-124-31/+42
| | | | | | | | string, set an end-of-line flag and avoid calling mbrtowc() again for any of the incomplete characters that remain in the string. - Use "mbs" for the multi-byte state variable name (for consistency). - Use the new MB_INVALID and MB_INCOMPLETE defines for the size_t -1 and -2 values (respectively).
* - The return value of mbsrtowcs() is also a size_t (like mbrtowc()).Wayne Davison2006-01-121-21/+33
| | | | | | | | | - When mbrtowc() returns -2 when given all the remaining chars in a string, set an end-of-line flag and avoid calling mbrtowc() again for any of the incomplete characters that remain in the string. - Use "mbs" for the multi-byte state variable name (for consistency). - Use the new MB_INVALID and MB_INCOMPLETE defines for the size_t -1 and -2 values (respectively).
* - The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-121-61/+71
| | | | | | | | | | | | | assign it to an int and then check for negativity, as that won't work on a system where an int is larger than a size_t. - When mbrtowc() returns -2 when given all the remaining chars in a string, set an end-of-line flag and avoid calling mbrtowc() again for any of the incomplete characters that remain in the string. - Use STOUC() when passing a char value to nicechar(). - Use "mbs" for the multi-byte state variable name (for consistency). - Be sure to reset the mbs state if mbrtowc() returns -1. - Use the new MB_INVALID and MB_INCOMPLETE defines for the size_t -1 and -2 values (respectively).
* Define MB_INCOMPLETE and MB_INVALID for MULTIBYTE_SUPPORT.Wayne Davison2006-01-121-0/+3
|
* Tweaked the code to handle mbrtowc() converting '\0' the same way asWayne Davison2006-01-111-11/+20
| | | | the other callers do. Also, changed the variable name to 'cnt'.
* Decided to use a switch() in mb_niceformat().Wayne Davison2006-01-111-7/+12
|
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-111-6/+6
| | | | | assign it to an int and then check if it's <= 0, as that won't work on a system where an int is larger than a size_t.
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-111-5/+7
| | | | | assign it to an int and then check if it's >= 0, as that won't work on a system where an int is larger than a size_t.
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-111-3/+6
| | | | | assign it to an int and then check if it's < 0, as that won't work on a system where an int is larger than a size_t.
* 22151: bug outputting here-strings from which etc. when already quotedPeter Stephenson2006-01-101-10/+11
|
* Changed the name of the "ret" variable in mb_niceformat() to "cnt"Wayne Davison2006-01-091-8/+7
| | | | | | | because "ret" is usually used for a variable name to hold the return value of the function. Also, changed the test when checking for a \0 to only check if "cnt" is 0, since we must always change a value of 0 to 1.
* Changed the name of the "ret" variable in mb_niceformat() to "cnt"Wayne Davison2006-01-091-8/+8
| | | | | | | because "ret" is usually used for a variable name to hold the return value of the function. Also, changed the test when checking for a \0 to one that checks if "cnt" is 0, since we must always change a value of 0 to 1.
* A test for (size_t)-1 needed to check for (size_t)-2 too.Wayne Davison2006-01-091-1/+1
|
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-091-10/+9
| | | | | | | | | | assign it to an int and then check if it's < 0, as that won't work on a system where an int is larger than a size_t. Also, the code that handled partial multibyte characters (that were assembled from multiple bytes of a metafied string) was not advancing past all the assembled bytes, nor was it handling the decoding of a '\0' char (it looks like it could have infinite looped in that case).
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-091-15/+10
| | | | | | | | assign it to an int and then check if it's > 0, as that won't work on a system where an int is larger than a size_t. Also, we needed to use STOUC() on a char value passed to nicechar(), and we need to clear the mbstate_t object if mbrtowc() returns an error.
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-091-4/+4
| | | | | | assign it to an int and then check if it's > 0, as that won't work on a system where an int is larger than a size_t. Also, we needed to use STOUC() on a char value passed to nicechar().
* The return value of mbrtowc() is a size_t (unsigned), so don'tWayne Davison2006-01-091-3/+3
| | | | | assign it to an int and then check if it's > 0, as that won't work on a system where an int is larger than a size_t.
* Changed mb_niceformat() so that it does not truncate a name thatWayne Davison2006-01-081-15/+14
| | | | | | has an invalid character sequence in the current character set, displaying them as \M-... chars. (Improved version of the patch from workers/22140.)
* Got rid of wcswidth() define for OpenBSD now that wcswidth() isn'tWayne Davison2006-01-081-4/+0
| | | | used.
* Changed wcswidth(&c, 1) to wcwidth(c).Wayne Davison2006-01-081-1/+1
|
* 22124: handle bad multibyte input betterPeter Stephenson2006-01-062-3/+13
|
* 22123: probe for wcswidth() and assume 1 if not foundPeter Stephenson2006-01-061-0/+4
|
generated by cgit-pink 1.4.1 (git 2.36.1) at 2024-05-19 06:25:00 +0000