zsh-3.1.6-dev-20

zsh-3.1.6-dev-20

[Peter Stephenson]

I have uploaded
  ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz
with the changes listed below.  This will need checking quite carefully,
particularly by people who have submitted a lot of patches (guess who that
might mean?)

A few of Sven's most recent patches are missing at his request.

My main remaining priority for 3.1.7 (despite various notes squirreled away
in odd places about things which don't work) is to finish compinstall,
which is now menu driven and sets styles.  I may post an unfinished version
in a desperate attempt to throw the responsibility on other people.  It
won't do anything clever with contexts this time round.


2000-03-23  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Tanaka Akira: 10234, 10236: Completion/X/_xwit: _message passing.

	* Tanaka Akira; 10232: Test/comptest, Test/ztst.zsh, ChangeLog(!):
	minor test fixes and typo.

	* Sven: 10226: Completion/Core/_path_files: quoting problem.

	* Sven: 10225: Completion/Core/_path_files: dot problems with
	files.

	* Tanaka Akira: 10223: Completion/Base/_regex_arguments,
	Completion/X/_x_colormapid, Completion/X/_xwit:
	new X completions.

	* Bart: 10218: Src/prompt.c: possible realloc() in
	putpromptchar().

2000-03-22  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Sven: 10215: Completion/User/_nedit: and this one.

	* Sven: 10214: Completion/Builtins/_cd,
	Completion/Builtins/_zstyle, Completion/User/_pbm,
	Completion/User/_pspdf, Completion/User/_tiff,
	Completion/User/_zdump: some _path_files calls now need loop for
	tags.

	* Sven: 10210: Completion/Commands/_next_tags,
	Completion/Core/_main_complete, Completion/Core/compinit,
	Doc/Zsh/compsys.yo: _next_tags doesn't need to be in completer
	list any more; $compprefuncs array does pre-completion functions.
	
	* Sven: 10209: Completion/Core/_loop, Completion/Core/_try:
	missing functions form 10195.

	* 10204: Completion/Base/_arguments, Completion/Base/_values,
	Completion/Core/_alternative, Completion/Core/_tags,
	Completion/Core/_wanted, Src/Zle/computil.c: additions to 10195.

	* Sven: 10195: Completion/Base/_arguments,
	Completion/Base/_brace_parameter, Completion/Base/_condition,
	Completion/Base/_default, Completion/Base/_describe,
	Completion/Base/_first, Completion/Base/_jobs,
	Completion/Base/_math, Completion/Base/_parameter,
	Completion/Base/_subscript, Completion/Base/_tilde,
	Completion/Base/_values, Completion/Builtins/_arrays,
	Completion/Builtins/_autoload, Completion/Builtins/_bindkey,
	Completion/Builtins/_builtin, Completion/Builtins/_cd,
	Completion/Builtins/_command, Completion/Builtins/_compdef,
	Completion/Builtins/_echotc, Completion/Builtins/_functions,
	Completion/Builtins/_hash, Completion/Builtins/_limits,
	Completion/Builtins/_pids, Completion/Builtins/_popd,
	Completion/Builtins/_sched, Completion/Builtins/_signals,
	Completion/Builtins/_stat, Completion/Builtins/_vars,
	Completion/Builtins/_zftp, Completion/Builtins/_zle,
	Completion/Builtins/_zmodload, Completion/Builtins/_zpty,
	Completion/Builtins/_zstyle, Completion/Commands/_next_tags,
	Completion/Core/_alternative, Completion/Core/_files,
	Completion/Core/_main_complete, Completion/Core/_multi_parts,
	Completion/Core/_options, Completion/Core/_parameters,
	Completion/Core/_path_files, Completion/Core/_requested,
	Completion/Core/_sep_parts, Completion/Core/_set_options,
	Completion/Core/_setup, Completion/Core/_tags,
	Completion/Core/_unset_options, Completion/Core/_wanted,
	Completion/Debian/_apt, Completion/Debian/_deb_packages,
	Completion/Linux/_rpm, Completion/User/_archie,
	Completion/User/_cvs, Completion/User/_dd,
	Completion/User/_domains, Completion/User/_gcc,
	Completion/User/_gdb, Completion/User/_getconf,
	Completion/User/_gprof, Completion/User/_groups,
	Completion/User/_gs, Completion/User/_hosts, Completion/User/_lp,
	Completion/User/_mailboxes, Completion/User/_make,
	Completion/User/_man, Completion/User/_mh, Completion/User/_mount,
	Completion/User/_mutt, Completion/User/_mysql_utils,
	Completion/User/_netscape, Completion/User/_nslookup,
	Completion/User/_pbm, Completion/User/_perl_basepods,
	Completion/User/_perl_builtin_funcs,
	Completion/User/_perl_modules, Completion/User/_ports,
	Completion/User/_rcs, Completion/User/_rlogin,
	Completion/User/_socket, Completion/User/_ssh,
	Completion/User/_stty, Completion/User/_tar,
	Completion/User/_telnet, Completion/User/_tiff,
	Completion/User/_urls, Completion/User/_user_at_host,
	Completion/User/_users, Completion/User/_users_on,
	Completion/User/_whois, Completion/User/_yp,
	Completion/X/_x_color, Completion/X/_x_cursor,
	Completion/X/_x_display, Completion/X/_x_extension,
	Completion/X/_x_font, Completion/X/_x_keysym,
	Completion/X/_x_modifier, Completion/X/_x_window,
	Completion/X/_xmodmap, Completion/X/_xutils, Doc/Zsh/compsys.yo,
	Doc/Zsh/mod_zutil.yo, Etc/completion-style-guide,
	Functions/Zftp/zfcd_match, Functions/Zftp/zfget_match,
	Src/Modules/zutil.c, Src/Zle/computil.c: implement tag aliases,
	which allows neat tricks with styles.

	* Sven: 10197: Src/exec.c, Src/init.c: %_ sometimes got cmdstack
	wrong in source(); $(...) didn't show up at all.

	* Bart: 10193: Src/prompt.c: prompttrunc() didn't track realloc()'s.

	* Bart: 10181: Functions/Prompts/prompt_bart_setup: Bart's Theme
	and Variations.

	* Sven: 10178: Completion/Commands/_next_tags, Doc/Zsh/compsys.yo:
	don't bind _next_tags by default and add some more docs.

	* Sven: 10176: Src/Zle/compresult.c, Src/init.c, Src/parse.c:
	optimize list-packed.

	* Sven: 10174: Doc/Zsh/compsys.yo, Doc/Zsh/compwid.yo: tweaks on
	completer and -V.

	* Bart: 10173: Functions/Prompts/prompt_bart_setup,
	Functions/Prompts/promptinit: Bart's Theme.

	* Sven: 10172: Completion/Core/_main_complete: typo.

	* Bart: 10168: Src/init.c: move hbegin() to get it right in
	precmd.

	* Oliver: 10166: Completion/X/_xt_arguments,
	Completion/User/_nedit: new and improved.

	* Bart: 10165: Doc/Zsh/builtins.yo, Doc/Zsh/compsys.yo: move doc
	for using a completer twice; zcompile doc tweak.

	* Oliver: 10162: Completion/Core/_expand: use nullglob option
	instead of (N) so that user's qualifiers work.

	* Sven: 10159: Src/Zle/compresult.c: wrong calculation of display
	length when not using zsh/complist.

	* Sven: 10156: Src/exec.c: possible invalid field in function
	definition when loading.

	* Zefram: 10152: Src/mem.c: fix ordering ambiguity ANSI wouldn't
	have liked.

	* Bart: 10150: Src/exec.c, Test/51xtrace.ztst: `one last' xtrace
	redirection fix.

	* Sven: 10149: Doc/Zsh/builtins.yo, Src/builtin.c, Src/exec.c,
	Src/parse.c: allows native and ksh-like wordcode function styles.

	* Sven: 10148: Src/Zle/comp.h, Src/Zle/compcore.c, Src/mem.c:
	optimization for added matches where things didn't change.

	* Sven: 10146: Completion/Core/_sort_tags, Completion/Core/_tags,
	Doc/Zsh/compsys.yo: return value of function called by _tags used
	to decide dispositions.

	* Sven: 10145: Doc/Zsh/mod_zle.yo: confess to users that zle input
	is on a stack.

	* Sven: 10144: Doc/Zsh/builtins.yo, Src/Modules/parameter.c,
	Src/builtin.c, Src/exec.c, Src/linklist.c, Src/parse.c: dump
	functions from current shell.

	* Sven: 10143: Src/init.c, Src/math.c: better ignoring of bits
	that shouldn't get evaluated in math expressions.

	* Sven: 10135: Doc/Zsh/compsys.yo, Src/Zle/computil.c: complex
	argument handling expressions.

	* Sven: 10131: Src/Modules/stat.c, Src/Modules/zpty.c,
	Src/Modules/zutil.c, Src/Zle/compctl.c, Src/Zle/complete.c,
	Src/Zle/computil.c, Src/Zle/zle_main.c, Src/Zle/zle_thingy.c,
	Src/parse.c: more consistent zerrnam/zwarnnam usage.

	* Sven: 10129: Doc/Zsh/builtins.yo, Src/builtin.c, Src/parse.c:
	autoload -w can mark all functions in .zwc files for autoloading.

	* Sven: 10128: Doc/Zsh/zle.yo, Src/Zle/zle_main.c, Src/init.c,
	Src/signals.c, Src/zsh.h: zle parameters available readonly in
	traps.

	* Sven: 10127: Doc/Zsh/zle.yo, Src/Zle/compresult.c,
	Src/Zle/iwidgets.list, Src/Zle/zle_main.c, Src/Zle/zle_misc.c,
	Src/Zle/zle_refresh.c, Src/Zle/zle_thingy.c, Src/Zle/zle_tricky.c:
	end-of-list widget to display prompt under command line, plus
	$MARK documentation.

	* Sven: 10126: Src/Zle/compcore.c: quotes in $PREFIX and $SUFFIX.

	* Sven: 10125: Src/math.c, Src/params.c: restore old cruddy
	behaviour removed in 10104, but using recursion counter.

	* Sven: 10120: Src/exec.c, Src/loop.c, Src/parse.c: state not
	tested properly in some constructs.

	* Sven: 10117: Doc/Zsh/mod_zle.yo, Src/Zle/zle_params.c,
	Src/Zle/zle_thingy.c: zle on its own returns zle status;
	MARK parameter; disallow widgets in traps.

	* Oliver: 10116: Completion/Builtins/_source: wasn't looking
	through the path.

	* Sven: 10112: Completion/Base/_condition, Src/Zle/zle_tricky.c:
	fix up altmodische _condition; `[[ ... && ' thought it was in
	command position for completion.

	* Sven: 10108: Completion/Commands/.distfiles,
	Completion/Commands/_next_tags, Completion/Core/_files,
	Completion/Core/_main_complete, Completion/Core/_requested,
	Completion/Core/_wanted, Doc/Zsh/compsys.yo,
	Functions/Misc/.distfiles: _next_tags allows you to cycle through
	sets of completions according to the tags.

	* Sven: 10107: Doc/Zsh/compwid.yo, Src/Zle/complete.c,
	Src/Zle/compmatch.c: ** in match spec allows matching including
	the anchor as well.

	* Sven: 10106: Functions/Misc/zrecompile, Src/exec.c, Src/parse.c,
	Src/zsh.h: recompilation of wordcode files and related bugs.

	* Sven: 10105: Src/lex.c: single exclamation marks in conditions.

	* Sven: 10104: Src/math.c, Src/params.c: retrieve parameters in
	math evaluations more directly rather than recursively calling
	matheval(), which was kind of stupid, so stupid I could	quite
	possible have put it in myself, but I don't think I did this time.

	* Sven: 10103: Src/glob.c: bad handling of ./ and ../ in
	closures.

	* Bart: 10098: configure.in, Src/cond.c: check for AIX's
	faccessx() and neaten configure library function tests.

	* Oliver: 10088: Completion/User/_groups,
	Completion/Base/_condition: better handling of files after
	conditions, checking of use of NIS.

	* Adam: 10087: Completion/User/_perl_modules,
	Completion/User/_perl: new and enhanced perl completions.

	* Oliver: 10073: Src/cond.c: /dev/fd emulation for access tests.

	* Adam: 10072: Completion/User/_mysql_utils: new completion
	function for something with suspcicious `SQL' bit in its name.

	* Bart: 10071: Src/exec.c: xtrace output redirection problem.

	* Sven: 10059: Src/builtin.c, Src/lex.c, Src/parse.c, Src/text.c,
	Src/zsh.h: bugs found by ksh tests: typeset -Ai didn't create
	assocs (still doesn't create ints), some heredoc difficulties, |&
	difficulties.

	* Alexandre: 10058: Completion/User/_prcs, Completion/User/_diff,
	Completion/User/_diff_options: another solution for the GNU diff
	completion problem.

	* Bart: 10056: Etc/MACHINES: RLIM_INFINITY incompatibility with
	some Linux/GNU libraries.

	* Sven: 10044: Completion/User/_diff, Completion/User/_use_lo:
	gdiff should use diff completion, pretty obviously.

	* Sven: 10041: Doc/Zsh/builtins.yo, Doc/Zsh/files.yo,
	Doc/Zsh/func.yo, Src/init.c, Src/parse.c: enhancements for digest
	files including digests of sourced files.  Or maybe digestion of
	sauce.

	* Sven: 10040: Completion/User/_diff_options: can test for GNU
	diff.

	* Oliver: 10029: Src/utils.c: bad handling of '^^' in keystring
	parsing.

	* Andrej: 10028: Completion/Builtins/_cd: remove local directory
	from cdpath.

	* Sven: 10025: Functions/Misc/nslookup, Src/Modules/zpty.c:
	optimizations for pty reads.

	* Sven: 10017: Doc/Zsh/builtins.yo, Src/lex.c, Src/parse.c:
	fix buffer problem for autoloading and unportable macro.

	* Oliver: 10013: Completion/Builtins/_compdef,
	Completion/Builtins/_zpty, Completion/User/_getconf: two new
	functions and one updated.

	* Andrej: 10011: Completion/User/_sudo: options for recent sudo.

	* Bart: 10009: Completion/Base/_subscript: didn't get arrays the
	right length.

	* Bart: 10006: configure.in, Src/utils.c: use mktemp() if
	available.

	* Bart: 9992: Functions/Zle/predict-on: tweaks.

	* Bart: 9990: Test/07cond.ztst: compatibility fix for chmod g+s.

	* Bart: 9982: Test/53completion.ztst, Test/Makefile.in,
	Test/comptest, Test/ztst.zsh: better handling of uninstalled
	modules.

	* Andrej: 9981: Src/Modules/zpty.c: SYSV-related pty stuff.

	* Bart: 9980: Src/Makefile.in: separate mymods.conf and xmods.conf
	dependencies.

	* Sven: 9966: Doc/Zsh/builtins.yo, Doc/Zsh/func.yo, Src/parse.c:
	require .zwc ending for digest files.

	* Bart: 9963: Test/11glob.ztst: existing globbing tests run under
	new test harness.

	* Bart: 9961: INSTALL, Src/Makefile.in, Test/53completion.ztst,
	Test/comptest: errors trying to load zpty for static shell in
	tests; INSTALL didn't describe mymods.conf properly.

	* Tanaka Akira: 9960: Test/comptest, Test/53completion.ztst: avoid
	zpty timing problem.

	* Sven: 9958: Src/parse.c: sharing strings made test fail with
	new wordcode.

	* Tanaka Akira: 9957: Test/comptest: clear to end.

	* Sven: 9953: Src/Zle/complist.c: was incorrectly omitting some
	display sequences.

	* Tanaka Akira: 9952: Test/comptest: compatibility, quoting and
	debugging fixes.

	* Sven: 9951: Completion/Core/_multi_parts: was producing
	mysterious blank entries.

	* Sven: 9948: Src/exec.c, Src/parse.c: stripkshdef() was broken.

	* Sven: 9947: Completion/Core/compdump, Completion/Core/compinit,
	Doc/Zsh/builtins.yo, Doc/Zsh/func.yo, Src/builtin.c, Src/cond.c,
	Src/exec.c, Src/glob.c, Src/lex.c, Src/loop.c, Src/math.c,
	Src/mem.c, Src/parse.c, Src/text.c, Src/utils.c, Src/zsh.h:
	Wordcode compilation for functions using .zwc or `digest
	functions' in $fpath.

	* Bart: 9946: Src/subst.c: return type in 9937.

	* Bart: 9945: Src/prompt.c: truncation messed up non-literal
	parts of the prompt.

	* Bart: 9942: Src/prompt.c: truncation in true part of ternary
	expression incorrectly showed false part and closing parenthesis.

	* Sven: 9937: Src/subst.c: parsing problem with parameter
	substitution which showed up in file completion.

	* Sven: 9938: Src/Modules/parameter.c: problem assigning to
	dirstack.

	* Tanaka Akira: 9936: Src/Zle/complist.c, Test/comptest,
	Test/53completion.ztst: tests for completion using zpty.

	* Tanaka Akira: 9935: Test/52zregexparse.ztst: tests for
	zregexparse.

2000-03-21  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Sven: 9920: Completion/Core/.distfiles, Completion/Core/_call:
	the function needed for 9897.

	* Bart: 9912: Completion/User/_mailboxes: further improvements.

	* Sven: 9904: Src/utils.c: getquery() and rmstarsilent.

	* Sven: 9902: Completion/Core/_path_files: _files -g problem.

	* Sven: 9900, 9903: Src/exec.c: parsing oddity in here documents.

	* 9897: Completion/Base/_arguments, Completion/Builtins/_pids,
	Completion/Linux/_rpm, Completion/User/_a2ps,
	Completion/User/_archie, Completion/User/_diff_options,
	Completion/User/_finger, Completion/User/_groups,
	Completion/User/_gs, Completion/User/_killall,
	Completion/User/_lp, Completion/User/_telnet,
	Completion/User/_users_on, Completion/User/_yp,
	Completion/X/_x_font, Completion/X/_x_window, Doc/Zsh/compsys.yo:
	use _call function and `command' style to override command
	used in various completions.

	* Sven: 9895: Completion/Base/_combination,
	Completion/Builtins/_zstyle, Completion/User/_telnet,
	Doc/Zsh/compsys.yo: usage of combination styles.

	* Sven: 9893: Completion/Core/_multi_parts: return values.

	* Sven: 9892: Completion/Core/_main_complete,
	Completion/Core/_path_files, Completion/Core/compdump,
	Completion/Core/compinit: improved fix for 9886; typo in
	_main_complete; compdump uses (q) flag; remove compstyle; update
	compconf.

	* Bart: 9891: Completion/User/_cvs: kshautoload problem.

	* Bart: 9889: acconfig.h, configure.in, Src/exec.c,
	Src/prototypes.h: from 3.0.8: detect mknod(), minor redirection
	tweaks.

	* Bart: 9886: Completion/Core/_path_files: matcher problem.

	* Tanaka Akira: 9885: Completion/User/_dvi, Completion/User/_lynx,
	Completion/X/_xutils: extra arguments for completion functions.

	* Alexandre: 9882: Completion/User/_prcs: error messages.

	* Sven: 9880: Completion/Core/_path_files: error message with
	~unknown.

	* Sven: 9879: Completion/User/_tilde_files: use of magicequalsubst
	with completion.

	* Sven: 9877: Src/parse.c: wordcode bug in functions with failed
	builtins.

	* Sven: 9876: Completion/Base/_combination,
	Completion/User/_rlogin, Completion/User/_ssh,
	Completion/User/_user_at_host, Doc/Zsh/compsys.yo: more flexible
	handling of separators in combinations.

	* Sven: 9870: Completion/Core/_path_files,
	Completion/User/.distfiles, Completion/User/_dd,
	Completion/User/_tilde_files, Src/Zle/compresult.c: better
	handling of paths beginning ~ in various places.

	* Sven: 9865: Completion/Builtins/_zstyle,
	Completion/Core/_main_complete, Completion/Core/_setup,
	Doc/Zsh/compsys.yo, Src/Zle/compcore.c, Src/Zle/compresult.c:
	implement prefer-ignored style to prefer alternative matches
	to going on to another completer.

	* Sven: 9863: Src/parse.c, Src/text.c: wordcode problems, using
	old code.

	* Sven: 9861: Completion/Core/_main_complete: interaction of
	_main_complete and _oldlist.

	* Sven: 9860: Src/Zle/compctl.c: storing of match and display list.

	* Sven: 9855: Completion/Core/_multi_parts, Doc/Zsh/compsys.yo:
	_multi_parts accepts -f.

2000-02-23  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Sven: 9845: Functions/Zle/incremental-complete-word: another
	context change.

	* Sven: 9844: Functions/Zle/predict-on: rationalize 9837.

	* Sven: 9840: config.sub: test for alphaev6 systems.

	* Sven: 9839: Src/Builtins/sched.c, Src/Modules/example.c,
	Src/Modules/mapfile.c, Src/Modules/parameter.c,
	Src/Modules/zftp.c, Src/Modules/zprof.c, Src/Modules/zpty.c,
	Src/Modules/zutil.c, Src/Zle/compcore.c, Src/Zle/compctl.c,
	Src/Zle/complete.c, Src/Zle/complist.c, Src/Zle/compresult.c,
	Src/Zle/computil.c, Src/Zle/zle_hist.c, Src/Zle/zle_main.c,
	Src/Zle/zle_misc.c, Src/Zle/zle_thingy.c, Src/Zle/zle_tricky.c,
	Src/Zle/zleparameter.c, Src/builtin.c, Src/cond.c, Src/exec.c,
	Src/glob.c, Src/hist.c, Src/init.c, Src/jobs.c, Src/lex.c,
	Src/linklist.c, Src/loop.c, Src/main.c, Src/math.c, Src/mem.c,
	Src/module.c, Src/params.c, Src/parse.c, Src/pattern.c,
	Src/prompt.c, Src/signals.c, Src/subst.c, Src/utils.c, Src/zsh.h:
	Eliminate `useheap' variable and consequent HEAPALLOC/PERMALLOC
	usage throughout the shell (hooray!)  All memory allocation is
	either explicitly on the heap or permanent.

	* Sven: 9838: Src/Modules/parameter.c, Src/Zle/computil.c,
	Src/Zle/zle_main.c, Src/builtin.c, Src/cond.c, Src/exec.c,
	Src/glob.c, Src/hashtable.c, Src/init.c, Src/jobs.c, Src/lex.c,
	Src/linklist.c, Src/loop.c, Src/params.c, Src/parse.c,
	Src/signals.c, Src/signals.h, Src/subst.c, Src/text.c,
	Src/version.h, Src/zsh.h: parser now emits wordcode directly; a
	whole heap of optimisations of the execution tree.

	* Sven: 9837: Completion/User/_gunzip, Functions/Zle/predict-on:
	no. of colons in predict-on, typo in _gunzip.

	* Sven: 9835: Completion/Builtins/_zstyle,
	Completion/Core/_main_complete, Doc/Zsh/compsys.yo: document
	single-ignored style, better handling of descriptions with
	warnings about no matches.

	* Bart: 2930: Completion/User/_ssh: basic completion for remote
	files.

2000-02-22  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Sven: 9831: Completion/Core/_multi_parts: better separator
	removal.

	* Sven: 9830: Completion/X/_x_arguments: reverse 9824, but use
	_default instead of _files if no special arguments.

	* Bart: 9828: Completion/User/_mailboxes: $i -> "$i" missed out.

2000-02-21  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* Clint: 9824: Completion/X/_x_arguments: `*/X11/*' was too broad
	a definition for X commands taking standard arguments to be
	useful.

	* Clint: 9823: configure.in: open fifo read/write in test to avoid
	possible (unconfirmed) hang.

	* Bart: 9822: Completion/User/_mailboxes: add completions for mush
	and tkrat; allow pine to complete mh folders.

	* Sven: 9816: Completion/Core/_multi_parts,
	Completion/User/_mailboxes: improved version of 9812/9815.

	* Bart: 9812, 9815: Completion/User/_mailboxes,
	Completion/User/_mutt: improved mailbox completion for all sorts
	of types, currently only used for mutt.

	* Sven: 9810: Completion/Core/_multi_parts: difficulties with
	separator as initial character.

	* Sven: 9809: Completion/Builtins/_popd,
	Completion/Core/_approximate, Doc/Zsh/compsys.yo: documentation
	improvements and typo fixes.

	* Tanaka Akira/Sven: 9807: Completion/Core/_main_complete:
	single-ignored style, as yet undocumented, for handling of
	alternate match set.

	* Sven: 9806: Completion/Core/_multi_parts: problems with literal
	(parenthesised) arrays.

	* Sven: 9805: Completion/Core/_path_files: difficulties with
	suffixes when expanding paths.

	* Sven: 9802: Src/utils.c: no need to quote !'s in ${(qq)...}
	because they're not active in single quotes.

	* Bart: 9799: Src/utils.c: better interpretation of typeahead when
	querying user e.g. for spell checking.

	* Adam: 9798: Doc/Makefile.in: use `texi' and `html' for targets
	to avoid duplicated name complaint by gmake.

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>

Re: zsh-3.1.6-dev-20

[Tanaka Akira]

In article <E12Yc4C-0002qM-00.2000-03-24-21-56-29@cmailg3.svr.pol.co.uk>,
  Peter Stephenson <pws@pwstephenson.fsnet.co.uk> writes:

> I have uploaded
>   ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz
> with the changes listed below.  This will need checking quite carefully,
> particularly by people who have submitted a lot of patches (guess who that
> might mean?)

* 10233 is not applied.

* The description for users-hosts-ports is doubled in
  Doc/Zsh/compsys.yo.  (9913 pointed out this.)

* Src/text.c is different to my version.  Where is the change come from?

| Index: Src/text.c
| ===================================================================
| RCS file: /projects/zsh/zsh/Src/text.c,v
| retrieving revision 1.1.1.19
| retrieving revision 1.1.1.18
| diff -u -r1.1.1.19 -r1.1.1.18
| --- Src/text.c	2000/03/25 00:22:02	1.1.1.19
| +++ Src/text.c	2000/03/23 04:11:14	1.1.1.18
| @@ -122,8 +122,7 @@
|      tlim = tbuf + tsiz;
|      tindent = 1;
|      tjob = 0;
| -    if (prog->len)
| -	gettext2(&s);
| +    gettext2(&s);
|      *tptr = '\0';
|      untokenize(tbuf);
|      return tbuf;
| @@ -195,6 +194,9 @@
|  	struct {
|  	    int par;
|  	} _cond;
| +	struct {
| +	    Wordcode end;
| +	} _subsh;
|      } u;
|  };
|  
| @@ -330,8 +332,10 @@
|  	    if (!s) {
|  		taddstr("( ");
|  		tindent++;
| -		tpush(code, 1);
| +		n = tpush(code, 1);
| +		n->u._subsh.end = state->pc + WC_SUBSH_SKIP(code);
|  	    } else {
| +		state->pc = s->u._subsh.end;
|  		tindent--;
|  		taddstr(" )");
|  		stack = 1;
| @@ -341,8 +345,10 @@
|  	    if (!s) {
|  		taddstr("{ ");
|  		tindent++;
| -		tpush(code, 1);
| +		n = tpush(code, 1);
| +		n->u._subsh.end = state->pc + WC_CURSH_SKIP(code);
|  	    } else {
| +		state->pc = s->u._subsh.end;
|  		tindent--;
|  		taddstr(" }");
|  		stack = 1;

* 9988 is applied (by hand because it's encoded by quoted-printable?)
  but it has a problem.

| % cvs diff -w -u -r1.1.1.6 -r1.1.1.5 Src/Modules/zpty.c
| Index: Src/Modules/zpty.c
| ===================================================================
| RCS file: /projects/zsh/zsh/Src/Modules/zpty.c,v
| retrieving revision 1.1.1.6
| retrieving revision 1.1.1.5
| diff -u -w -u -r1.1.1.6 -r1.1.1.5
| --- Src/Modules/zpty.c	2000/03/25 00:22:04	1.1.1.6
| +++ Src/Modules/zpty.c	2000/03/14 11:18:00	1.1.1.5
| @@ -165,7 +165,7 @@
|  
|  #else /* ! __osf__ */
|  
| -#if __SVR4
| +#if defined(__SVR4) || defined(sinix)
|  
|  #include <sys/stropts.h>
|  
| @@ -174,11 +174,12 @@
|  {
|      int mfd, sfd;
|      char *name;
| +    int ret;
|  
|      if ((mfd = open("/dev/ptmx", O_RDWR)) < 0)
|  	return 1;
|  
| -    if (!(name = ptsname(mfd)) || grantpt(mfd) || unlockpt(mfd)) {
| +    if (grantpt(mfd) || unlockpt(mfd) || !(name = ptsname(mfd))) {
|  	close(mfd);
|  	return 1;
|      }

Since ret is undeclared, it cannot be compiled.

% make
...
gcc -c -I. -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -DHAVE_CONFIG_H -DMODULE -Wall -Wno-implicit -Wmissing-prototypes -ggdb -fpic -o zpty..o zpty.c
zpty.c: In function `get_pty':
zpty.c:189: `ret' undeclared (first use in this function)
zpty.c:189: (Each undeclared identifier is reported only once
zpty.c:189: for each function it appears in.)

* zsh-users_2973 is not applied. (Because Geoff is not commented?)
-- 
Tanaka Akira

Re: zsh-3.1.6-dev-20

[Peter Stephenson]

Tanaka Akira wrote:
> * 10233 is not applied.

That's because Sven said `if the tag alias changes make it', and I assume
he was talking about the latest set which he didn't want in yet.  If those
appear next time (and I expect they will), this will be added too.

> * The description for users-hosts-ports is doubled in
>   Doc/Zsh/compsys.yo.  (9913 pointed out this.)

O.K., as Bart told me that explains the drift I was getting in line numbers
in Sven's patches for the file.

> * Src/text.c is different to my version.  Where is the change come from?

10179, which was an attempt to fix something which apparently didn't have
any effect.  I will take advice on whether to apply it or not.

> * 9988 is applied (by hand because it's encoded by quoted-printable?)

Sort of; I applied the first version by hand because of the problem that
made the quoted-printable version necessary.  (My father is a retired
lawyer.)

>   but it has a problem.
>
> Since ret is undeclared, it cannot be compiled.

Whoops, that's because I'm not compiling in the SVR4 environment.  Sorry,
I've applied Bart's update.

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>

Bart's cleareod patch users/2973 (Was: zsh-3.1.6-dev-20)

[Geoff Wing]

Tanaka Akira <akr@jaist.ac.jp> typed:
:* zsh-users_2973 is not applied. (Because Geoff is not commented?)

Sorry, the message made it onto the disk and into the mailing list archive
but for some reason didn't get posted into my news system so I missed it.

Excerpts from Bart's message:
:I believe this has to do with the always_last_prompt option -- the code
:that prints the prompt wants to be sure that there's a clean slate below
:the prompt where it can display completion listings.  However, it's a
:bit too aggressive; it performs the clear-eod whenever it's "safe" to do
:so (that is, whenever there's no completion listing there already that
:needs to continue to be seen) and not just whenever it's necessary to do
:so (which it -almost- never is when always_last_prompt is not set).

Usually it's unnecessary.  However, most of the commands people run are
external to the shell and their output cannot be guaranteed to leave the
cursor in an appropriate position (i.e. below any output the external
command has generated).  Thus we clear a ``workspace'' and now we can
utilise an area we know is blank.  CLEAREOD is the quickest way.  Bart's
change makes it affect only the area we currently use.  The problem with
this (though maybe other people won't consider it a problem) is that it
does not leave any buffer space (e.g. an extra blank line) between zsh
output and any garbage that may be lying around on the screen.  Always
putting an extra blank line in will cause some extra/unnecessary scrolling
if we're at the bottom of the screen (and usually we can't reliably
determine where we are on the screen).  So the options are:
1) CLEAREOD                                (commonly: ``^[[J'')
2) CLEAREOL in our area                    (commonly: ``^[[K'' on each line) 
3) CLEAREOL in our area plus an extra line

Now there are parts of the screen refresh which are a bit aggressive, viz.
if we've done a CLEAREOD we still might do CLEAREOL further on - however
I haven't considered all the display permutations to determine whether it
is worth removing them yet.

Regards,
-- 
Geoff Wing : <gcw@pobox.com>     Work URL: http://www.primenet.com.au/
Rxvt Stuff : <gcw@rxvt.org>      Ego URL : http://pobox.com/~gcw/
Zsh Stuff  : <gcw@zsh.org>       Phone   : (Australia) 0413 431 874

Re: zsh-3.1.6-dev-20

[Bart Schaefer]

On Mar 25, 10:35am, Tanaka Akira wrote:
} Subject: Re: zsh-3.1.6-dev-20
}
} In article <E12Yc4C-0002qM-00.2000-03-24-21-56-29@cmailg3.svr.pol.co.uk>,
}   Peter Stephenson <pws@pwstephenson.fsnet.co.uk> writes:
} 
} > I have uploaded
} >   ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz
} 
} * 10233 is not applied.

That's because 10230 and 10231 are not applied either.  Will 10233 work
without those others?

} * Src/text.c is different to my version.  Where is the change come from?

They came (or rather, did NOT come) from 10179, which was not applied.

} * 9988 is applied (by hand because it's encoded by quoted-printable?)
}   but it has a problem.

Below is what I have (diffed against -dev-20).  There's an extra hunk in
there that converts leading spaces to tabs in some of the indentation;
I don't know if that's a mailing artifact or an emacs-editing artifact
on my part.

It also appears that 9897 and 10089 were not properly merged into _groups.

Index: Completion/User/_groups
===================================================================
@@ -6,7 +6,8 @@
 
 if ! zstyle -a ":completion:${curcontext}:" groups groups; then
   (( $+_cache_groups )) ||
-      if (( ${+commands[ypcat]} )) && tmp=$(ypcat group.byname); then
+      if (( ${+commands[ypcat]} )) &&
+	  tmp=$(_call groups ypcat group.byname); then
         : ${(A)_cache_groups:=${${(f)tmp}%%:*}} # If you use YP
       else
         : ${(A)_cache_groups:=${${(s: :)$(</etc/group)}%%:*}}

Index: Src/Modules/zpty.c
===================================================================
@@ -165,7 +165,7 @@
 
 #else /* ! __osf__ */
 
-#if __SVR4
+#if defined(__SVR4) || defined(sinix)
 
 #include <sys/stropts.h>
 
@@ -174,11 +174,12 @@
 {
     int mfd, sfd;
     char *name;
+    int ret;
 
     if ((mfd = open("/dev/ptmx", O_RDWR)) < 0)
 	return 1;
 
-    if (!(name = ptsname(mfd)) || grantpt(mfd) || unlockpt(mfd)) {
+    if (grantpt(mfd) || unlockpt(mfd) || !(name = ptsname(mfd))) {
 	close(mfd);
 	return 1;
     }
@@ -186,31 +187,31 @@
 	close(mfd);
 	return 1;
     }
-   if ((ret = ioctl(sfd, I_FIND, "ptem")) != 1)
-      if (ret == -1 || ioctl(sfd, I_PUSH, "ptem") == -1) {
-          close(mfd);
-          close(sfd);
-          return 1;
-      }
-   if ((ret = ioctl(sfd, I_FIND, "ldterm")) != 1)
-      if (ret == -1 || ioctl(sfd, I_PUSH, "ldterm") == -1) {
-          close(mfd);
-          close(sfd);
-          return 1;
-      }
-   if ((ret = ioctl(sfd, I_FIND, "ttcompat")) != 1)
-      if (ret == -1 || ioctl(sfd, I_PUSH, "ttcompat") == -1) {
-          close(mfd);
-          close(sfd);
-          return 1;
-      }
+    if ((ret = ioctl(sfd, I_FIND, "ptem")) != 1)
+       if (ret == -1 || ioctl(sfd, I_PUSH, "ptem") == -1) {
+	   close(mfd);
+	   close(sfd);
+	   return 1;
+       }
+    if ((ret = ioctl(sfd, I_FIND, "ldterm")) != 1)
+       if (ret == -1 || ioctl(sfd, I_PUSH, "ldterm") == -1) {
+	   close(mfd);
+	   close(sfd);
+	   return 1;
+       }
+    if ((ret = ioctl(sfd, I_FIND, "ttcompat")) != 1)
+       if (ret == -1 || ioctl(sfd, I_PUSH, "ttcompat") == -1) {
+	   close(mfd);
+	   close(sfd);
+	   return 1;
+       }
     *master = mfd;
     *slave = sfd;
 
     return 0;
 }
 
-#else /* ! (defined(__SVR4) || defined(sinix)) */
+#else /* ! (defined(__SVR4) || defind(sinix)) */
 
 static int
 get_pty(int *master, int *slave)



-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: zsh-3.1.6-dev-20

[Bart Schaefer]

On Mar 25,  4:46pm, Bart Schaefer wrote:
} Subject: Re: zsh-3.1.6-dev-20
}
} Index: Src/Modules/zpty.c
} ===================================================================
}  
} -#else /* ! (defined(__SVR4) || defined(sinix)) */
} +#else /* ! (defined(__SVR4) || defind(sinix)) */

Of course there's no reason to introduce my silly spelling error into
the baseline.  I should have spotted that.

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: zsh-3.1.6-dev-20

[Chmouel Boudjnah]

Peter Stephenson <pws@pwstephenson.fsnet.co.uk> writes:

> I have uploaded
>   ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz
> with the changes listed below.  This will need checking quite carefully,
> particularly by people who have submitted a lot of patches (guess who that
> might mean?)
> 

i got this

yodl -o zsh.texi -I. -w ztexi.yo version.yo zsh.yo; \
test -f zsh.texi

yodl fatal error: file version.yo unreadable (No such file or directory)
make: *** [zsh.texi] Error 1

-- 
MandrakeSoft Inc                http://www.mandrakesoft.com
Pasadena, CA USA                                  --Chmouel

Re: zsh-3.1.6-dev-20

[Bart Schaefer]

On Mar 25,  8:53am, Chmouel Boudjnah wrote:
} Subject: Re: zsh-3.1.6-dev-20
}
} i got this
} 
} yodl fatal error: file version.yo unreadable (No such file or directory)
} make: *** [zsh.texi] Error 1

version.yo is built from Config/version.mk by this code in Doc/Makefile.in:

version.yo: $(sdir_top)/Config/version.mk
	( \
	  echo 'STARTDEF()'; \
	  echo 'def(version)(0)('$(VERSION)')'; \
	  echo 'def(date)(0)('$(VERSION_DATE)')'; \
	  echo 'ENDDEF()#' | tr '#' '\\'; \
	) > $(sdir)/version.yo


You must be having some kind of configure-time problem ?

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: zsh-3.1.6-dev-20

[Bart Schaefer]

On Mar 23,  9:56pm, Peter Stephenson wrote:
} Subject: zsh-3.1.6-dev-20
}
} I have uploaded
}   ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz

Hmm; version.mk still has

VERSION_DATE='February 23, 2000'

There's also one change in Src/text.c that I can't explain ... I missed
it before because of the 10179 diffs ... it looks sensible, but it's not
from any article mentioned in ChangeLog ... the `+' lines are -dev-20:

@@ -122,7 +122,8 @@
     tlim = tbuf + tsiz;
     tindent = 1;
     tjob = 0;
-    gettext2(&s);
+    if (prog->len)
+	gettext2(&s);
     *tptr = '\0';
     untokenize(tbuf);
     return tbuf;

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: zsh-3.1.6-dev-20

[Peter Stephenson]

"Bart Schaefer" wrote:
> Hmm; version.mk still has
> 
> VERSION_DATE='February 23, 2000'

Sorry, I was originally going to send a version just before I left, but
didn't when I discovered the wordcode bugs I mentioned around then.

> There's also one change in Src/text.c that I can't explain ... I missed
> it before because of the 10179 diffs ... it looks sensible, but it's not
> from any article mentioned in ChangeLog ... the `+' lines are -dev-20:

> -    gettext2(&s);
> +    if (prog->len)
> +	gettext2(&s);

I think this was when Sven and I were looking at the same termination
problem in the wordcode.  I included his fix, but didn't back off mine
because it looked safer (= more paranoid).  Or something.

P.S.: thanks for all the comments on zshguide, which has now updated to
include the corrections people have sent, but still needs work due to
changes in the shell, and the more detailed explanation of file-patterns.
I think I'll move it to the zsh web site about the time 3.1.7 appears;
what's there seems to be OK for public consumption.

http://www.pwstephenson.fsnet.co.uk/computing/

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>

Re: zsh-3.1.6-dev-20

[Sven Wischnowsky]

Peter Stephenson wrote:

> I have uploaded
>   ftp://ftp.zsh.org/zsh/development/zsh-3.1.6-dev-20.tar.gz
> with the changes listed below.  This will need checking quite carefully,
> particularly by people who have submitted a lot of patches (guess who that
> might mean?)

Whew!

> A few of Sven's most recent patches are missing at his request.

But I hope everyone has applied them now, because I hope to find the
time to send some patches relative to them today... apart from those,
the only missing patches, I can see are Oliver's 10029 (for your
convenience included below) and my 10233 (as mentioned by Tanaka; it
would have worked without 10230 and 10231).


In another message:

> Tanaka Akira wrote:
>
> ...
> 
> > * Src/text.c is different to my version.  Where is the change come from?
> 
> 10179, which was an attempt to fix something which apparently didn't have
> any effect.  I will take advice on whether to apply it or not.

It adds a bit to security, so I would be in favour of applying it.


Bye
 Sven

diff -u -r ../zsh-3.1.6-dev-20/Src/utils.c ./Src/utils.c
--- ../zsh-3.1.6-dev-20/Src/utils.c	Wed Mar 22 21:59:14 2000
+++ ./Src/utils.c	Thu Mar  9 10:57:21 2000
@@ -3198,7 +3198,7 @@
 	} else if (fromwhere == 4 && *s == Snull) {
 	    for (u = t; (*u++ = *s++););
 	    return t + 1;
-	} else if (*s == '^' && !control &&
+	} else if (*s == '^' &&
 		   (fromwhere == 2 || fromwhere == 5 || fromwhere == 6)) {
 	    control = 1;
 	    continue;

--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: zsh-3.1.6-dev-20

[Sven Wischnowsky]

Peter Stephenson wrote:

> ...
> 
> 	* Sven: 10125: Src/math.c, Src/params.c: restore old cruddy
> 	behaviour removed in 10104, but using recursion counter.

Before I forget to ask: of course there are other things that may
recurse endlessly. Functions, source, ${(e)...} come to mind...

Hm. bash doesn't limit function recursion, but ksh does (to 128 nested 
functions).

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: zsh-3.1.6-dev-20

[Sven Wischnowsky]

I forgot...

Peter Stephenson wrote:

> ...
> 
> My main remaining priority for 3.1.7 (despite various notes squirreled away
> in odd places about things which don't work) is to finish compinstall,
> which is now menu driven and sets styles.  I may post an unfinished version
> in a desperate attempt to throw the responsibility on other people.  It
> won't do anything clever with contexts this time round.

I'll have a look (once you've shown it to us ;-) and I'll also have a
look at the Guide (hoping to be able to remember everything that has
changed).

As I said, I think the tag stuff is quite stable now (modulo the
label/alias naming thing). That basically leaves:

- the completer thing (giving completer-aliases/labels(?) in the
  completer style); this would allow to make the contexts for _prefix
  and _ignored somewhat easier to understand, I think
- renaming _try and _loop

for the completion system.


Other things:

- the wordcode stuff; -r/-m versus -M and writing ksh-loaded functions ,
  as discussed (10276)
- maybe the autoload +X[zk] thing (10275)
- the STAT_NOPRINT thing; i.e. should we just test it? (10005 and 10016)


Oh, and 10103 was ok?

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: zsh-3.1.6-dev-20

[Sven Wischnowsky]

I wrote:

> ...
> 
> the only missing patches, I can see are Oliver's 10029 (for your
> convenience included below) and my 10233 (as mentioned by Tanaka; it
> would have worked without 10230 and 10231).

I guess everyone has realised that already... again it was me who had
missed this patch, not Peter.


I'll better creep into some hole in the ground now...

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de