Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | 37092: make nested ${(P)name} properly refer to parameter on return | Peter Stephenson | 2015-11-11 | 9 | -69/+138 |
| | |||||
* | missed log entry | Peter Stephenson | 2015-11-11 | 1 | -0/+5 |
| | |||||
* | 37091: clear lastline and lastlinesz when freeing | Peter Stephenson | 2015-11-11 | 1 | -0/+2 |
| | |||||
* | 37084: add implementation check to zsh/parm/private test | Peter Stephenson | 2015-11-10 | 2 | -0/+8 |
| | |||||
* | Merge branch 'master' of git://git.code.sf.net/p/zsh/code | Barton E. Schaefer | 2015-11-09 | 2 | -5/+10 |
|\ | | | | | | | | | Conflicts: ChangeLog | ||||
| * | 37085: fix doc: DEBUG_BEFORE_CMD is set by default | Jun-ichi Takimoto | 2015-11-09 | 2 | -5/+10 |
| | | |||||
* | | Jun T. proofreading | Barton E. Schaefer | 2015-11-09 | 2 | -2/+6 |
|/ | |||||
* | 37081: new module zsh/param/private for private-scoped parameters in functions | Barton E. Schaefer | 2015-11-08 | 6 | -2/+955 |
| | |||||
* | 37080: use paramtab abstraction more consistently, add explanatory comments | Barton E. Schaefer | 2015-11-08 | 3 | -6/+14 |
| | |||||
* | 37079: better description of "optstr" field in struct builtin as used in ↵ | Barton E. Schaefer | 2015-11-08 | 2 | -1/+8 |
| | | | | BUILTIN() macro | ||||
* | 37074: extend previous fix to over whitespace at end | Peter Stephenson | 2015-11-07 | 3 | -46/+73 |
| | |||||
* | 37073: another SH_WORD_SPLIT problem. | Peter Stephenson | 2015-11-06 | 3 | -6/+94 |
| | | | | | In cases like x${:- y} the space was simply removed instead of being used for splitting. | ||||
* | 37062: tentative mechanism for git commit descriptions. | Peter Stephenson | 2015-11-04 | 2 | -9/+39 |
| | | | | | Mechanism for forcing completion system into verbose mode is subject to change. | ||||
* | 35303: remove Completion/Unix/Command/_systemd | Daniel Hahler | 2015-11-02 | 2 | -470/+5 |
| | | | | | systemd's upstream version seems to be maintained well, and e.g. on Arch Linux the version from Zsh gets removed in favour of it. | ||||
* | 37038: add -l option to emulate to list options in emulations | Peter Stephenson | 2015-11-01 | 4 | -9/+69 |
| | |||||
* | 37029: update options and new tr completion | Oliver Kiddle | 2015-10-31 | 5 | -14/+82 |
| | |||||
* | 37028: new scl completion | Oliver Kiddle | 2015-10-31 | 2 | -0/+40 |
| | |||||
* | 36993: zed: fix argument parsing | Eric Cook | 2015-10-31 | 2 | -21/+14 |
| | |||||
* | unposted (cf. 36998,36999): undo 36956 / restore 34451 with expanded comment ↵ | Barton E. Schaefer | 2015-10-31 | 3 | -6/+10 |
| | | | | about the flip-flopping | ||||
* | 37032: Temporarily revert 36959. | Daniel Shahaf | 2015-10-31 | 2 | -1/+8 |
| | |||||
* | 37027: allow subword context to discriminate between words | Peter Stephenson | 2015-10-30 | 3 | -5/+14 |
| | |||||
* | unposted: Typo in ChangeLog | Daniel Shahaf | 2015-10-30 | 1 | -1/+1 |
| | |||||
* | unposted (after 37018): Fix typo in error message. | Daniel Shahaf | 2015-10-30 | 3 | -2/+7 |
| | |||||
* | 37022: add GLOB_STAR_SHORT option to abbreviate ** and *** | Peter Stephenson | 2015-10-30 | 6 | -18/+60 |
| | |||||
* | 37018: Make WARNCREATEGLOBAL more consistent. | Peter Stephenson | 2015-10-29 | 4 | -42/+45 |
| | | | | | Wd don't need separate math handling any more, and can make it output the function name in all cases. | ||||
* | missed ChangeLog entry | Peter Stephenson | 2015-10-29 | 1 | -0/+2 |
| | |||||
* | unposted: ignore .orig files | Peter Stephenson | 2015-10-29 | 1 | -0/+1 |
| | |||||
* | 37014: Improved internal parameter setting. | Peter Stephenson | 2015-10-29 | 9 | -29/+94 |
| | | | | | | | | Enhance WARNCREATEGLOBAL to work in many more cases. Don't create REPLY as an integer if it didn't previously exist as one, even if the value to be set is integral, as this is likely to mess up later uses of REPLY. | ||||
* | 36983 + 36990: new completions for head and tail | Jun-ichi Takimoto | 2015-10-29 | 3 | -0/+115 |
| | |||||
* | 36994: declare local REPLY for use with "zle .read-command"; use ↵ | Barton E. Schaefer | 2015-10-27 | 2 | -3/+9 |
| | | | | .self-insert instead of .self-insert-unmeta | ||||
* | unposted: fix typo quite -> queue | Barton E. Schaefer | 2015-10-27 | 1 | -1/+1 |
| | |||||
* | 36982: Fix bug with (#cN) patterns and remove redundant description. | Peter Stephenson | 2015-10-27 | 5 | -10/+22 |
| | | | | | We need to restore the current count of matches when returning to match at the point where we previously matched. | ||||
* | 36977: note about difficulties of pattern repetitions | Peter Stephenson | 2015-10-27 | 2 | -0/+13 |
| | |||||
* | 36974: fix some functions with empty argument lists | Peter Stephenson | 2015-10-27 | 6 | -11/+17 |
| | |||||
* | 36906: quite_signals() in ZSH_MEM realloc() | Kamil Dudka | 2015-10-26 | 2 | -2/+10 |
| | |||||
* | 36968: use addmodulefd() to tell the shell about the descriptor of the dbm file | Barton E. Schaefer | 2015-10-26 | 2 | -2/+10 |
| | |||||
* | 36956: revert 34451, mmap() is too slow on MacOS | Barton E. Schaefer | 2015-10-26 | 2 | -0/+11 |
| | |||||
* | users/20825: fix crash when complex completion hooks. | Peter Stephenson | 2015-10-26 | 2 | -0/+23 |
| | | | | | get_undo_current_change() needs protecting against execution in completion environment. | ||||
* | 36962: _git: Fix completion of RHS of refspecs. | Daniel Shahaf | 2015-10-25 | 2 | -1/+5 |
| | | | | | | Before this patch, 'git push $remote :<TAB>' completed branch names (as though by `git branch -a`), instead of completing remote branch names sans the ${remote}/ prefix. | ||||
* | 36964: _git: Complete remotes branch names with slashes correctly. | Daniel Shahaf | 2015-10-25 | 2 | -1/+4 |
| | | | | For example, 'git push remote HEAD:foo/bar' creates such branches. | ||||
* | 36960: _git: Enable slash matcher for more branch and tag completions, ↵ | Daniel Shahaf | 2015-10-25 | 2 | -3/+7 |
| | | | | include 'git branch <TAB>' and 'git tag <TAB>'. | ||||
* | 36959: _git: Offer @~$n as completion of recent commits. | Daniel Shahaf | 2015-10-25 | 2 | -1/+5 |
| | | | | Suggested-by: Oliver Kiddle (users/20705) | ||||
* | 36958: _git: Fix recent commit completion descriptions. | Daniel Shahaf | 2015-10-25 | 2 | -12/+32 |
| | | | | | | The uniquifiers 'HEAD~$n' were incorrect when a recent commit was the second parent of a merge commit. Detect that case and print something correct instead. | ||||
* | unposted: _git: Fix 'commit object name' completion messages. | Daniel Shahaf | 2015-10-25 | 2 | -2/+7 |
| | | | | | | | | Offer the "this alternative is still valid, but I don't know to offer completions for it" for hash completion but not for recent object completion; the former is protected by a _guard and the latter incorrectly appeared also in cases such as 'git log a/b<TAB>' which were not the start of a gitrevisions(7) expression. | ||||
* | 36911: '-optarg' should not match optspec '-opt=' | Jun-ichi Takimoto | 2015-10-25 | 2 | -2/+10 |
| | | | | | Remove the requirement that -xy= should come before -x= in the list of optspecs passed to _arguments. | ||||
* | Merge branch 'master' of git://git.code.sf.net/p/zsh/code | Barton E. Schaefer | 2015-10-24 | 3 | -1/+43 |
|\ | | | | | | | | | | | | | | | Conflicts: ChangeLog Conflicts: ChangeLog | ||||
| * | unposted: _beep completion: Actually hook it for the 'beep' command. | Daniel Shahaf | 2015-10-24 | 2 | -1/+4 |
| | | |||||
| * | 36913 + 36945: vcs_info quilt: Pass patch subject lines to gen-applied-string | Daniel Shahaf | 2015-10-24 | 2 | -0/+33 |
| | | | | | | | | The format used is '${patchname} ${subject}', which is analogous to the git backend. | ||||
| * | 36912: vcs_info quilt: Tolerate being in child of .pc's parent | Daniel Shahaf | 2015-10-24 | 2 | -0/+6 |
| | | | | | | | | | | Without this, gen-applied-string would only be set when .pc is a child of cwd, not when it is a child of an ancestor of cwd. | ||||
* | | 36943: restore scan for reclaimable blocks in freeheap() | Barton E. Schaefer | 2015-10-24 | 2 | -2/+13 |
|/ | | | | | | That scan had been removed by 36834, but testing showed memory usage climbing too high in cases where a new arena was always added at the end of the heap list. |