Commit Graph

495 Commits

Author SHA1 Message Date
tg 93ab969fdc compat.h does have the status of a system header though... 2006-08-22 22:22:07 +00:00
tg 5b2be8517f save more whitespace and make it compile again (oops) 2006-08-22 22:21:20 +00:00
tg a87ff74cbd * version.h doesn't need an RCS ID
* compat.h does, add it to misc.c, always include it, move if NEED_COMPAT
* major whitespace cleanup in sh.h
2006-08-22 22:16:04 +00:00
tg e20d82c4d1 * rename the (generated) Test.sh to test.sh
* point to it, TFM and the sample more
2006-08-22 22:01:59 +00:00
tg 7161eeb35e Plan 9 stuff belongs into its branch 2006-08-18 18:59:49 +00:00
tg 5ee451f98c restrict RANDOM to a positive 31-bit integer number, for use with modulo op 2006-08-18 18:48:26 +00:00
tg 79ba3570f0 add a note about why this check can fail 2006-08-18 13:41:09 +00:00
tg b131ca909d as long as we're using arc4random(3) to generate $RANDOM,
let it be an uint32_t, like nbsh(1)
2006-08-18 13:40:16 +00:00
tg 11282be59d compat.h has a section 'extra headers', move inttypes.h there 2006-08-16 00:00:24 +00:00
tg 78cb0de69c shuffle code around for better structure 2006-08-15 23:56:32 +00:00
tg 98c4dc04a1 interix doesn't have <inttypes.h>, openbsd no <stdint.h>, the old chaos.
so add <inttypes.h> only for the one platform which needs it today, APE.
2006-08-15 23:54:09 +00:00
tg f34f24fefc be more semantical 2006-08-15 23:45:53 +00:00
tg 1e313c282c minor cleanups from the porting work 2006-08-15 23:43:30 +00:00
tg 083f19dc53 * document the feature-not-bug status of pipeline-subshell-1
* fix spelling in regression test
* expand test-stlt-1 by reverse logic, just to make sure
2006-08-14 20:41:21 +00:00
tg 2936b81597 check the last command of a pipeline is REALLY executed in a subshell 2006-08-14 20:36:39 +00:00
tg 1c9a925139 rewrite stripcom/Lstripcom function, making use of the 'last command
of a pipeline is executed in a subshell' trick, don't use co-process
(because that's verbose twice in an interactive shell profile), move
the noglob block into the while and rid the redundant +o and local.
2006-08-14 20:32:44 +00:00
tg 7c083ff950 improve English 2006-08-12 21:49:53 +00:00
tg 3ddbd5e7fa optimise with a goto 2006-08-12 20:32:33 +00:00
tg b208f76525 fix quoting levels of backthingies 2006-08-12 20:30:22 +00:00
tg 36b7c841d3 some help from the autoconf portable shell docs 2006-08-12 20:19:36 +00:00
tg 70babe6b0c fix and simplify further 2006-08-12 20:01:35 +00:00
tg d1dfe9657f simplify further 2006-08-12 19:53:39 +00:00
tg cca3f93e59 a probably-Solaris-/bin/sh-workable improved algorithm 2006-08-12 19:51:09 +00:00
tg 0eb31f6a82 optimise signames.inc output (shrinks binary size) 2006-08-12 19:38:44 +00:00
tg e7fab65d5b -D_ANSI_SOURCE no longer needed, verified on all platforms which
actually need signame.inc
2006-08-12 19:26:20 +00:00
tg 2950eea78d mirbsdksh -> mksh 2006-08-12 18:49:21 +00:00
tg 9bda688f59 sanitise variable handling and nroff vs gnroff detection 2006-08-12 18:48:39 +00:00
tg 24000a4fb2 gensigs.sh was called with $SHELL (or /bin/sh if $SHELL was a csh),
which failed for Han Boetes using zsh as user shell (to test). Now,
the code is integrated into Build.sh as-is, and Test.sh uses $SHELL
no longer either but the mksh just built, in the shebang line.
2006-08-12 18:43:55 +00:00
tg 3b194eebb7 point people to the IRC channel (#mksh/OPN) 2006-08-11 13:46:32 +00:00
tg 1100be9300 re-implement "set -o posix" which doesn't do much ;)
but turns off 'braceexpand' when turned on as side effect,
just like oksh/pdksh.
document "set -o sh" too.
2006-08-09 20:44:16 +00:00
tg e62ca90db2 wording improvements 2006-08-09 20:34:42 +00:00
tg 500c89635f fix an omission to delete, and remove confusing ancient comment 2006-08-09 20:27:18 +00:00
tg c874727a5c commit a test for posix mode (turning off braceexpand when turned on)
which mksh doesn't currently pass (oksh passes this test)
2006-08-09 20:21:08 +00:00
tg 49d627c9f7 correct the description of "abort" 2006-08-08 20:19:14 +00:00
tg 4661992d5b remove the 'stuff' functionality altogether,
ksh93 doesn't have it, and I don't know, nor
do I care about what "the BRL ^T mini-systat
feature" is.
2006-08-08 20:17:22 +00:00
tg 5176249248 * bind ^T to transpose-chars as documented in the manual page,
not to stuff, even if TIOCSTI is available
* it's transpose-chars not transpose

bug spotted by hondza who wondered that on ^T the line repeats itself forever
2006-08-08 20:07:53 +00:00
tg e2e1a01610 use $SHELL for Test.sh too 2006-08-02 14:17:13 +00:00
tg 174b4c5a1d the 'delete' key does *not* act as Ctrl-D (EOT, cf. ascii(7)) 2006-08-02 14:02:55 +00:00
tg 58e5606710 bind xterm-xfree86 and wscons (NetBSD®) wsvt25/vt220 too 2006-08-02 14:00:19 +00:00
tg 52357a30be this is mksh R28 (beta)
I think of implementing multiline editing (bash-style) for R28 final,
but other than that, the new features already suffice for a new major
2006-08-02 13:34:03 +00:00
tg c96e31a707 implement a hack whereas you can now bind functions (and macros)
to keys with a trailing tilde (after prefix + control char); the
function is then invoked after prefices, control char and one more
character X are entered; if X is not a tilde, it is ungetc(3)d
2006-08-02 13:32:18 +00:00
tg b99aa06d9b save whitespace 2006-08-02 12:54:48 +00:00
tg 7771520c05 rid junk 2006-08-02 12:50:20 +00:00
tg 99007d97e7 cry if the 'bind' editing command gets too long arguments
minor KNF while here
2006-08-02 12:49:04 +00:00
tg e869099ba3 some thoughts on the two expected-fail tests
change one from bug to feature
2006-08-02 12:38:45 +00:00
tg bef8651936 * Build.sh: fix manpage generation defaults under Cygwin
* mksh.1: rework prompt ($PS1) section, simplify example,
  point to packaged dot.mkshrc example
* mksh.1: fix description of 'redraw' regarding redrawal
  of prompts longer than one screen line
* all: bump version/date
2006-08-02 11:50:30 +00:00
tg 790c312e9f fix the prompt redrawal routines for multi-line prompts, prompts
with embedded newline (\n), and multi-line prompts with embedded
newlines. single-line prompts should be not affected; prompts with
embedded carriage return (\r) should be fixed by this as well.
also fix prompt_redraw comment (can't remove the variable though,
since it's required if the shell inserts a newline after the prompt
herself).
tested on cygwin and interix
2006-08-02 11:33:37 +00:00
tg ad43cfa6cd rename 'count' into the more appropriate 'columns' 2006-08-02 10:42:30 +00:00
tg df54522223 * Build.sh: fix accidental CR-LF catastrophe
(yeah, that's what you get from developing with a
  16-bit MS-DOS(R) executable of your favourite text editor)
* lex.c, mksh.1: do not print the delimiting character for
  not-to-be-counted character sequences (i.e. ANSI escapes)
  in prompts any more, mostly because ASCII 01h is printable
  on both Interix and Cygwin and I'm lazy
  (this also fixes prompt width counting if a printable character
  such as 'x' is used as delimiting character)
* lex.c: through printing the prompt character by character,
  also print the delimited sequences if skipping, fixes some
  prompt redrawal not honouring colours stuff while not totally
  redrawing the entire prompt
2006-08-02 10:41:03 +00:00
tg ce8fd25903 fix build on Cygwin, has been broken for quite some time apparently 2006-08-02 10:02:21 +00:00