cdac7b7d65
clean up our own junk
2007-09-09 18:21:52 +00:00
c1c939e340
• fix memory leaks found by coverity
...
from netbsd via oksh
we had the NULL pointer deref already fixed
• avoid a bogus not-setting the return value of edit.c:x_file_glob()
introduced by the above change in oksh
• escape ? as well (but not ] because that’s wrong)
reminded by cbiere@netbsd via oksh
• Unsetting a non-existent variable is not an error. See
http://www.opengroup.org/onlinepubs/009695399/utilities/unset.html
report from Arkadiusz Miskiewicz; fixed based on
http://cvs.pld-linux.org diff via oksh but modified slightly
• Be more smart waiting for input for non-interactive scripts. Fix
based on a diff from debian: see their bug#296446 (via oksh)
modified slightly
this also fixes cnuke@’s “mksh busy loop” problem, for which I never
received a bug report, but the Debian bug page contains a set of two
scripts to reproduce this before (and no longer after) this commit
• some KNF
• bump version
2007-09-09 18:06:42 +00:00
fe65f616e4
some have u_int but not uint32_t
2007-09-09 12:02:38 +00:00
6f30f395bf
-std=foo can’t be made fatal even with -Werror, live with it
2007-09-09 11:55:44 +00:00
b06d40f8f2
autoscan for uint32_t et al. if <stdint.h> not present
...
this un-special-cases PW32 and should help OSF/1
2007-09-09 11:04:30 +00:00
a1afc4e0a7
gcc 2.8.1 (on Digital Taugtnix a.k.a. alpha-dec-osf4.0d) makes -std=foo
...
failures a warning only
2007-09-09 10:49:20 +00:00
daf019728c
fix typo (blsk -> bksl)
2007-09-07 23:57:14 +00:00
6c47b381c2
testing with the Heirloom sh found this… d’oh
2007-09-07 21:58:40 +00:00
7ddc3d788f
disable annoying privacy invasion, fight the Überwachungsstaat
2007-09-03 09:25:05 +00:00
87eeeb6da0
revert the -std=gnu99–by–default change until programmes are ready, as this
...
is too much annoying (for now) – extra commit so that we have a changeset
2007-08-24 14:25:33 +00:00
c8ac9422f9
• add -std=gnu99 to default CFLAGS
...
• first round of assorted fixes
2007-08-24 14:19:56 +00:00
6b9c277fbc
shut up tcc
2007-08-20 14:12:29 +00:00
b82204aa3d
proto
2007-08-20 14:06:10 +00:00
bc8ba122a7
cast to quieten gcc/obsd/sparc64 __CRAZY=Yes
2007-08-20 13:57:47 +00:00
1e3783f030
employ string pooling techniques to save off a few more bytes
...
(probably more than the new “rename” builtin ever required…)
2007-08-19 23:12:23 +00:00
ca46f3dc98
new builtin “rename” (this name sounds better than “mksh_mv_rescue” ☻)
...
to just call rename(2) directly, e.g. if /bin/mv needs /lib/ld-uClibc.so.0
some assorted code cleanup
2007-08-19 22:06:27 +00:00
49c95a8938
put clear-screen on ESC ^L as suggested by the AT&T ksh people in
...
http://www.opensolaris.org/os/community/on/flag-days/pages/2007081701/
2007-08-18 01:20:28 +00:00
e99f988aae
feature freeze for mksh R31, to be released as part of MirOS #10
2007-08-18 00:22:09 +00:00
f684b0ceb1
Implement the “source” command, as requested by some and agreed bsiegert@
...
In contrast to AT&T ksh93, its semantics are like GNU bash in that it ap-
pends the current working directory to the search path; it is implemented
as a shell alias instead of enhancing funcs.c:shbuiltins[] like in ksh93.
2007-08-13 19:39:21 +00:00
7381c6c274
mention that $RANDOM is _always_ 15-bit unsigned integer, even arc4random’s
2007-08-12 14:01:38 +00:00
f3bb1472df
Debian warning cleanup
2007-08-12 13:52:03 +00:00
499327f7b8
add “set -o arc4random”, RTFM for details
2007-08-12 13:42:23 +00:00
cf2eb697e2
use bounds check for function that calls read
...
prompted by lina output, 10x tarzeau
not a bug
2007-08-06 12:02:39 +00:00
45cadd2abf
make sure we don't show any junk in MKSH_CLS_STRING even if the terminal is
...
not standards compatible
2007-07-31 15:42:50 +00:00
47836e45e7
add a new bindable command “clear-screen”, which can be bound to the ^L key
...
so that archite@midnightbsd won’t have to add evil kludges to oksh again if
they switch their ksh to mksh ☺
both “clear-screen” and “error” aren’t bound; default binding for ^L stays,
as usual, “redraw” (principle of least surprise); however GNU bash converts
also might want to put “bind ^L=clear-screen” into their ~/.mkshrc.
2007-07-31 15:29:40 +00:00
73fa4622ef
do the rlim_t for tcc cast dance for Debian GNU/kFreeBSD
2007-07-31 13:55:26 +00:00
496b8f4dc3
• new way of checking for mknod & friends, due to tcc vs glibc weirdness
...
• bump vsn for the code restructuring
2007-07-31 11:11:25 +00:00
809c1446b2
some steps towards building with tcc
2007-07-31 10:42:15 +00:00
9884efef00
• implement some (not yet tested) kludges for tcc
...
• define $CPP internally
2007-07-31 10:17:52 +00:00
735ec12dfd
• TenDRA (testsuite pass)
...
• Fabrice Bellard’s tcc (not supported yet, may follow later)
reminded by tarzeau (Gürkan Sengün)
2007-07-31 10:04:46 +00:00
3e5b234965
these are actually cool people, I think
2007-07-26 22:38:31 +00:00
b9da9d4ea3
if the shell doesn't know the cwd, “builtin pwd” won't either
2007-07-26 13:37:21 +00:00
8c958b410d
small textual nits
2007-07-26 13:35:58 +00:00
b8a693ed68
mention ${foo:2:3} limitations
2007-07-26 13:32:16 +00:00
3718a00106
• test on HURD (gnubber)
...
• fix unreachable code (break stmt) cought by suncc on yofuh's E420
• bump vsn to today
2007-07-26 13:23:52 +00:00
7374f12176
some more not-so working platforms
...
can't extend the working ones better since my UWIN decided to break
after MSVC++ Expreß seems to have “upgraded” some libraries… *sigh*
2007-07-24 21:54:46 +00:00
b20e5a90ce
initial support for PW32, not quite working yet (same issues as Minix 3 and Plan 9, I think)
2007-07-24 21:47:14 +00:00
e37577de42
fix another errexit bug: unwind() is __dead, so the pseudo-signal was never
...
delivered to the process
• regression test by Clint Pachl, verified against Debian ksh93 by me
• place to fix it in the code discovered by Otto Moerbeek
2007-07-24 11:22:04 +00:00
f2a52c65fa
… and fix that bug.
...
Ｇｕｖｆ ｐｂｚｚｖｇ ｈａｘａｂｊｖａｔｙｌ ｆｃｂａｆｂｅｒｑ
ｏｌ ＮｈｅＶＦＣ ＴｏＥ， ｆｖａｐｒ ｊｒ ｈｆｒ ｚｘｆｕ，
ｎａｑ ｚｌ ｂｇｕｒｅ ｇｎｆｘｆ ｎｅｒ ｊｎｖｇｖａｔ．．．
2007-07-23 14:28:52 +00:00
7c0209771c
more testcase – currently broken… had thought of this but apparently not done
2007-07-23 14:10:48 +00:00
b09b3621e2
• we had an unused variable leftover
...
• make warning-free for both gcc and xlC
2007-07-22 14:01:50 +00:00
748d09122a
de-experimental AIX
2007-07-22 13:47:10 +00:00
a1433b7122
• Build.sh: always prepend well-known signal names, so that some signals
...
will not come up weirdly (e.g. on AIX: SIGSAK (SIGMAX?), SIGIO (SIGAIO?),
SIGURG (SIGIOINT?)), and add a few more while here
• check.t, sh.h: bump
2007-07-22 13:46:15 +00:00
cf5fc9eac2
some null/NUL
2007-07-22 13:38:26 +00:00
979406bba7
• support IBM xlC on AIX
...
• fix all bugs it could find ☺
2007-07-22 13:34:52 +00:00
211ecedf4e
• move __GNUC__ test quite to the bottom, since I fear other compilers,
...
such as icc, define it too
• provide a Rebuild.sh skipping the mirtoconf checks – devel only!
2007-07-22 13:08:54 +00:00
29d4125e3c
switch $CC default from gcc to cc, as we support most native compilers now,
...
and – with Daniel's words – das ist doch kein Rumgeleenockse hier!
2007-07-17 19:41:26 +00:00
56a8da33cf
• make the "if called as sh, set FPOSIX" not !SMALL-only
...
• bump to R30-gamma, feature freeze
2007-07-17 13:56:51 +00:00
823f5c2a17
hopefully fix large file stuff for real
2007-07-09 11:19:55 +00:00
4422137a88
introduce SADELIM_MAKE matching style – ${foo:S/bar/baz/flags} –
...
to complement SADELIM_BASH, as before – ${foo/bar/baz} –
but ifdef'd out for now, so probably no coding style. also new
is a flags array (not declared yet)
2007-07-07 22:29:36 +00:00
tg