Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | 39813: fix keybuflen after ungetting extra keys from the end of keybuf in ↵ | Barton E. Schaefer | 2016-11-03 | 1 | -2/+5 |
| | | | | getkeymapcmd(). | ||||
* | 39815: Read input to end on parse error in $(...) inside a string. | Peter Stephenson | 2016-11-03 | 1 | -2/+11 |
| | | | | | This allows ${(z)} to output the whole string, although we can't do word splitting from the error onwards. | ||||
* | 39811: vi-repeat-change must not be the "motion" of vi-change, lest it ↵ | Barton E. Schaefer | 2016-11-02 | 1 | -1/+1 |
| | | | | infinitely repeat itself | ||||
* | 39777: $() is a valid empty command substitution | Peter Stephenson | 2016-10-30 | 1 | -1/+1 |
| | |||||
* | 39758: revise 39704 for array and hash parameters; more POSIXBUITINS tweaks for | Barton E. Schaefer | 2016-10-29 | 2 | -12/+18 |
| | | | | | | | | | | | | | export 39704 was commit 0f5e670, forgot to reference article number in that log. "typeset -p" outputs "typeset" for array and hash parameters, even when exported, because those types can be marked export but are never pushed to the enviroment. For POSIXBUILTINS, "export var" does not implicitly set $var, and its export state is preserved when assigned (but not when explicitly unset). | ||||
* | 39725: more metafication of patterns. | Barton E. Schaefer | 2016-10-25 | 1 | -2/+3 |
| | |||||
* | 39723: metafy string passed to pattern in complist. | Peter Stephenson | 2016-10-25 | 2 | -0/+5 |
| | | | | Also additional safety tokenizing a string with Meta characters. | ||||
* | "typeset -p" uses "export" commands or the "-g" option for parameters that ↵ | Barton E. Schaefer | 2016-10-24 | 1 | -5/+17 |
| | | | | are not local to the current scope | ||||
* | 39715: Handle multibyte characters in complist listing | Peter Stephenson | 2016-10-23 | 1 | -5/+12 |
| | |||||
* | 39688: Skip out of P_EXACTLY pattern loop if P_PURES. | Peter Stephenson | 2016-10-20 | 1 | -2/+2 |
| | | | | | | | If it's really a pure string, there's only one exact match. We could overwrite the next pointer; usually this didn't matter because we'd hit a termination, but if we reallocated we might be looking at garbage. | ||||
* | 39683: Update scan pointers after possible alloc. | Peter Stephenson | 2016-10-20 | 1 | -1/+5 |
| | | | | | Problem could cause next scan index to point into an invalid block when handling meta characters. | ||||
* | 39680: correctly handle case of popping last arena | Barton E. Schaefer | 2016-10-19 | 1 | -1/+2 |
| | | | | | This mattered only when compiling with --enable-stack-allocation because otherwise the heap was never completely empty. | ||||
* | 39678: metafication of printf %q argument and result | Peter Stephenson | 2016-10-19 | 1 | -2/+3 |
| | |||||
* | 39590: zle: Call zle-line-pre-redraw after inserting a completion result | m0viefreak | 2016-10-18 | 1 | -0/+1 |
| | | | | | This is necessary to make sure redrawhook() is called while cycling results in menu completion. | ||||
* | 39500: compdescribe: perform on-screen-width computation using the "nice" ↵ | Daniel Shahaf | 2016-10-16 | 1 | -9/+10 |
| | | | | | | | length, since the matches are rendered by nice*() functions by compadd Remove two "###" question comments relative to the patch posted, since the answer to them is (clearly) negative. | ||||
* | 39625: case needs to reset lastval if no pattern matched | Martijn Dekker | 2016-10-13 | 1 | -2/+6 |
| | |||||
* | 39611: with _arguments sets completion stopped if one of the rest arguments ↵ | Oliver Kiddle | 2016-10-13 | 1 | -1/+2 |
| | | | | starts with a dash | ||||
* | 36108: command -[vV] assumed -p | Peter Stephenson | 2016-10-11 | 1 | -2/+4 |
| | |||||
* | 39599: Don't reset status before "case" execution | Peter Stephenson | 2016-10-10 | 1 | -1/+0 |
| | |||||
* | 39579: Fix string calculation for parameter quoting. | Julien Cretin | 2016-10-06 | 1 | -3/+3 |
| | | | | | Where there was no closing quote the size or position of the null could be wrongly calculated. | ||||
* | 39571: Fix ERR_EXIT bug with && and function. | Peter Stephenson | 2016-10-05 | 1 | -6/+6 |
| | | | | | "foo && bar" inside a function could cause the code outside the function not to perform ERR_EXIT or ERR_RETURN when needed. | ||||
* | 39568: "! <complex-command>" suppresses ERR_EXIT | Peter Stephenson | 2016-10-05 | 1 | -2/+7 |
| | |||||
* | 39566: Improve usefulness of command_not_found_handler. | Peter Stephenson | 2016-10-05 | 1 | -5/+8 |
| | | | | | | | Don't behave as if command not found if return status is non-zero as this may simply be the return status of the replacement command. Let the function report a command not found instead. | ||||
* | 39561: missing unqueue_signals() when ZSH_HEAP_DEBUG | Barton E. Schaefer | 2016-10-04 | 1 | -2/+7 |
| | |||||
* | 39550: reset signal queue in recursiveedit() | Barton E. Schaefer | 2016-10-03 | 1 | -0/+6 |
| | |||||
* | 39548: DEBUG for queueing_enabled | Barton E. Schaefer | 2016-10-03 | 2 | -3/+32 |
| | |||||
* | 39547: handle zero delta in calc_timeout() | Barton E. Schaefer | 2016-10-03 | 1 | -1/+1 |
| | |||||
* | 39545: Add some missing unqueue_signals(). | Peter Stephenson | 2016-10-03 | 10 | -4/+24 |
| | | | | All of these are added simply to fit existing logic in other branches. | ||||
* | 39521: Refactor start of execcmd(). | Peter Stephenson | 2016-10-03 | 2 | -46/+97 |
| | | | | | | By splitting into _analyse and _exec execpline2() has easier access to the state at the start of execution. Use this to ensure we fork if this is a builtin with no arguments. | ||||
* | 39540: "! command" should suppress ERR_EXIT and ERR_RETURN | Peter Stephenson | 2016-10-02 | 1 | -2/+6 |
| | |||||
* | 39509: in the event the current directory has been removed, use chasedots ↵ | Barton E. Schaefer | 2016-09-30 | 1 | -4/+21 |
| | | | | semantics for "cd ..". | ||||
* | 39519: restore missing PM_EXPORT flags. | Peter Stephenson | 2016-09-30 | 1 | -0/+1 |
| | | | | This was missing when exporting using USE_SET_UNSET_ENV coce variant. | ||||
* | 39517: back off 39502 (WC_ASSIGN causes fork in pipe). | Peter Stephenson | 2016-09-30 | 1 | -1/+1 |
| | | | | This isn't a robust fix as WC_ASSIGNs simply precede the main wordcode. | ||||
* | 39498: use PRIVILEGED option to decide on problematic parameter imports | Peter Stephenson | 2016-09-30 | 2 | -4/+4 |
| | |||||
* | Call the pre-redraw hook if there is text in the buffer on init | Mikael Magnusson | 2016-09-30 | 1 | -1/+13 |
| | |||||
* | 39507: TMPSUFFIX for =(...) | Barton E. Schaefer | 2016-09-29 | 1 | -0/+9 |
| | |||||
* | 39470: failure to open a supposedly unique temp file name should result in ↵ | Barton E. Schaefer | 2016-09-29 | 2 | -5/+15 |
| | | | | | | an error Also band-aid for signal-related race conditions in temp file name generation | ||||
* | 39502: Fork for assignment in LHS of pipeline. | Peter Stephenson | 2016-09-29 | 1 | -1/+2 |
| | | | | | | | foo=bar | stuff left the value of foo set to bar as we didn't realise we needed to fork. | ||||
* | unposted: Update findcmd() in Zle. | Peter Stephenson | 2016-09-29 | 2 | -5/+7 |
| | | | | Unfinished business from previous patch. | ||||
* | 34943: Fixes for "command" with multiple options. | Peter Stephenson | 2016-09-29 | 3 | -39/+126 |
| | | | | | | These need to combine properly, and alos "command -p" with either -v or -V needs to search for builtins and then using the default system path. | ||||
* | unposted: change '\0' to NULL to silence spurious compile warning. | Bart Schaefer | 2016-09-28 | 1 | -1/+1 |
| | | | | | | params.c:830:13: warning: expression which evaluates to zero treated as a null pointer constant of type 'char *' [-Wnon-literal-null-conversion] *envp = '\0'; | ||||
* | 39460: Don't import PS4 if running as root. | Peter Stephenson | 2016-09-28 | 2 | -3/+31 |
| | | | | | There was an exploit in bash using SHELLOPTS to turn on xtrace, however this can't happen in zsh, so this is simply a precaution. | ||||
* | unposted: remove extra close-paren from d6a6b4ff | Bart Schaefer | 2016-09-27 | 1 | -1/+1 |
| | |||||
* | 39369 (tweaked cf 39371): when calling an internal widget, set bindk because ↵ | Oliver Kiddle | 2016-09-27 | 1 | -0/+4 |
| | | | | some widgets use it to determine how they should act | ||||
* | 39437: use list_pipe_pid in assignment for clarity | Barton E. Schaefer | 2016-09-27 | 1 | -1/+1 |
| | |||||
* | 39448: reset REPLY in select on user action | Martijn Dekker | 2016-09-26 | 1 | -0/+2 |
| | |||||
* | 39436: Pass on status of SIGINT better. | Peter Stephenson | 2016-09-25 | 2 | -1/+3 |
| | | | | | | Set lastval to 128 + SIGINT on interrupt. Don't execute builtin if already interrupted at that point. | ||||
* | 39435: Further fix for pgrp of funny pipelines. | Peter Stephenson | 2016-09-25 | 1 | -1/+2 |
| | | | | | Don't set gleader of SUBJOB at the point of creation if the SUPERJOB has no processes yet. | ||||
* | 39412: Fix directory completion when $PWD:h contains parentheses, which are ↵ | Daniel Shahaf | 2016-09-23 | 1 | -1/+4 |
| | | | | interpreted as globbing metacharacters. | ||||
* | 39383: compadd: Restrict previous patch to the case where $PREFIX includes ↵ | Daniel Shahaf | 2016-09-20 | 1 | -2/+4 |
| | | | | characters both from the -P prefix and from the (unprefixed) candidate completion word. |