12608d7933
fix function-local arrays
2011-11-26 00:45:03 +00:00
0d2244da81
use += and do some more mitigation of the alloc problem first
2011-11-25 23:58:04 +00:00
5c1108c5fa
actually… the error paths didn’t se-unset noglob! after looking at the
...
code, until further proof I believe noglob is not even needed ⇒ drop it
2011-11-25 23:54:06 +00:00
5ac24244fd
== inside [[ is not nice, and parsing $(set +o) isn’t either,
...
since we don’t just have [[ -o ]] now, but can use standard $- here
2011-11-25 23:43:59 +00:00
63878e8aef
another couple of verbose checks; the *-global are mksh specific but work;
...
mksh currently fails arrassign-fnc-local (never worked either)
2011-11-25 23:33:42 +00:00
2ff9255f39
add a -f option to test.sh that runs tests of class fastbox
...
plus a test in that category for a bug spotted during development of
https://evolvis.org/plugins/scmgit/cgi-bin/gitweb.cgi?p=shellsnippets/shellsnippets.git;a=blob;f=mksh/sysadmin/getjpics.cgi;h=d74018cbc1ae4e88408267ecc73a6df19aa4c99a;hb=HEAD
2011-11-25 23:29:31 +00:00
dd014625e1
Fix regression introduced by mixing the recursive parser and support
...
for ;| and ;& in TCASE: ;;-less last casepart produced ";\0" in the
SREREAD string which obviously cased reparsing to fail
test from http://www.in-ulm.de/~mascheck/various/cmd-subst/
2011-11-22 18:01:41 +00:00
d7079f3f0e
bump
2011-11-19 21:22:02 +00:00
709524edd1
add explicit casts to uid_t and gid_t of the struct stat slots, for some
...
architectures on Linux have a truly horrible struct stat definition, as
seen in dietlibc’s source code
https://buildd.debian.org/status/fetch.php?pkg=mksh&arch=mipsel&ver=40.2-4&stamp=1321054539
2011-11-19 17:42:24 +00:00
0220aeeb61
disable cd-pe test on Hurd, it also wrecks havoc with check.pl
...
https://buildd.debian.org/status/fetch.php?pkg=mksh&arch=hurd-i386&ver=40.2-4&stamp=1321059202
2011-11-19 17:25:55 +00:00
e20b1295b7
promote x=(a b); x+=(c) to standard mksh functionality at cost of 932 MKSH_SMALL .text bytes on MirBSD/i386
2011-11-11 22:14:19 +00:00
7e719a4cc1
deprecate the UGLY input line beginning with ! hack
2011-11-11 22:14:02 +00:00
cf4e967a5e
fix: "foo" in nroff is quoted, \&"foo" contains literal gänsebeinchen
2011-11-11 22:10:52 +00:00
90afc54ee8
• select(0, …) misbehaves on Interix (by Markus Duft <mduft@gentoo.org>)
...
• select(n, NULL, NULL, NULL, …) behaves the same (by Bruno Haible)
2011-11-09 22:17:26 +00:00
f048ac2f58
disable another symlink-using test on MSYS
2011-11-08 23:10:58 +00:00
c620f865c1
cid 1004E0F5C27271F5B00 broke integer-base handling of built-in parameters
...
mksh -c 'echo a=$RANDOM; RANDOM=0x12; echo b=$RANDOM' # behaviour changed
2011-11-08 22:07:15 +00:00
f92108dccc
make MSYS match Cygwin more closely; ignore its broken <stdint.h>
2011-11-05 23:39:02 +00:00
92b60a9a5e
forgotten in cid 1004EA5BF612F516747 to take '\n' presence as $'…' quoting reason
2011-10-26 20:46:16 +00:00
fed31331f4
rename more identifiers with trailing underscores
2011-10-25 22:36:39 +00:00
cd7dab250b
also Xcheck_grow_ may not end with an underscore
2011-10-25 22:25:07 +00:00
5893a36bad
document “export -” accident found by Snader_LB
2011-10-24 19:41:13 +00:00
3b9ec6c3c4
also emit newline as backslash-n to align with GNU bash (and to keep everything on one line)
2011-10-24 19:40:58 +00:00
5416abb2d0
when printing for-reentry parameter values, use dgk’s $'…' for non-ASCII characters
2011-10-24 19:40:46 +00:00
de75cbdb33
fix spello in comment, by <Snader_LB:#!/bin/mksh>
2011-10-16 00:18:56 +00:00
904dc3dae4
oh lookie, oksh fixed something we did in 2007… ⇒ pointless rcsid-in-comment sync
2011-10-11 19:06:07 +00:00
d6bf9c197b
better handling of eval and CPPFLAGS in build system
2011-10-07 19:51:44 +00:00
9eaf6a8e12
remaining sync with oksh (pointless rcsid, no code diff)
2011-10-07 19:51:29 +00:00
c2f994f6f1
unbreak test-build with printf
2011-10-07 19:51:17 +00:00
23afb6dbb8
merge “upper bound Emacs mode command repeat by input line length” from
...
oksh (slightly different implementation) and remark better plans (which
doesn’t mean bounding the repeat counter is bad, though…)
2011-10-07 19:45:08 +00:00
84096a0a86
Message-ID: <CAOw7k5gBmt+WqzQxgK8xrjMvCh3iBHcQ7cU78cy1WbzRf8d_Hg@mail.gmail.com>
...
just to make sure, got a copyright licence for this
2011-09-13 18:46:46 +00:00
9782f6b4d1
• access(2) is broken in at least kFreeBSD 9.0 as “modern” OS, so bring
...
back the wrapper code as well as refactor most other code calling it
• apparently, names can’t end in ‘_’ or contain ‘__’ anywhere…
2011-09-07 15:24:22 +00:00
577c918beb
patch most of Jerker Bäck’s concerns out, unless not applicable
2011-08-27 18:06:52 +00:00
5d9c88ceeb
jg71 reported -DMKSH_ASSUME_UTF8=* breaks defining stristr
2011-08-27 17:30:07 +00:00
da80c56acb
12:23⎜<cnuke:#bosng> angefangen damit, dass es kein uname unter OPENSTEP gibt :)
2011-08-26 10:24:51 +00:00
6f76d7f2fe
fix… pasto, I think, from some time ago, spotted by cnuke@
...
12:17⎜<cnuke:#bosng> OPENSTEP hat kein dirname und $srcdir/sh.h wird dann /sh.h
2011-08-26 10:20:28 +00:00
308290e2b6
dict.leo.org says this is correct
2011-08-13 22:19:41 +00:00
80223417d7
followup fix for 1004D90467358D6B13C regarding escaped substitutions
...
found in, out of all things, Debian/m68k, by Wouter Verhelst (Yoe)
fix hacked at DebConf 11, Бања Лука, Република Српска, Босна и Херцеговина
2011-07-26 16:57:28 +00:00
a7566387cf
optimise an if away, and possibly even the function bodies…
2011-07-20 23:47:29 +00:00
6ce68e906f
now that I got Hash.cs compiled and run, I’m amazed; make NZAAT (and NZAT!) all green with a small tweak; bit distribution and χ2 look OK as well
2011-07-18 00:35:46 +00:00
f51883e350
eek, only be a ksh_nice if we HAVE_NICE
2011-07-16 23:37:58 +00:00
ac814611f1
somebody (who is probably inlining) complains about memset with the
...
third argument being 0; maybe (I think can only be this) fix that
2011-07-16 18:15:45 +00:00
1b9bee46fc
check later whether the final link may succeed, and include "sh.h" → catch inlining bugs in system headers fighting with -c combine
2011-07-16 18:03:04 +00:00
0168f6dc48
some distributions declare nice(3) with warn-unused-result; annoy mksh users running into this so they complain to their distros
2011-07-16 17:41:50 +00:00
09db3d67a2
finish merging from R40-stable; set version number to yesterday though
2011-07-16 17:08:52 +00:00
620faa3dc9
‣ merge from mksh R40-stable
...
drop heredoc-weird-3 while here, it’s not deterministic by design(?)
2011-07-16 17:08:35 +00:00
e03b485d75
‣ merge from mksh R40-stable
...
move /etc/{,suid_}profile to /system/etc/ for Android (but do not make
the location of /etc configurable); rewrite manpage section about
/etc/{,suid_}profile, .profile, .mkshrc
2011-07-16 17:08:21 +00:00
b0ee967d00
‣ merge from mksh R40-stable
...
first part of long-overdue manpage improvements
2011-07-16 17:08:04 +00:00
d6445aa007
‣ merge from mksh R40-stable
...
document the unportability of the echo utility;
arekm's non-bug report inspired this
2011-07-16 17:07:48 +00:00
8d1edbd95e
‣ merge from mksh R40-stable
...
use common macro bodies shf_{get,put}c_ for definition
of shf_{put,get}c as functions (MKSH_SMALL) or macros, respectively
2011-07-16 17:07:35 +00:00
851f8fb7d8
introduce MKSH_NO_EXTERNAL_CAT skipping the “call /bin/cat if we have an
...
option” nonsense, intended for Android, Plan 9, and other systems which
honour that cat -v is deemed harmful
2011-07-07 21:24:53 +00:00
tg