Re: 3.1.7-pre-2

Re: 3.1.7-pre-2

[Sven Wischnowsky]

Bart Schaefer wrote:

> On Apr 30,  4:45pm, Bart Schaefer wrote:
> } Subject: Re: 3.1.7-pre-1 for the workers
> }
> } I have a couple of issues that came up after the prerelease to resolve.
> } 
> } The "read -q" bug that Sven patched in 10727.
> 
> So ... 10727 makes this work:
> 
> 	zsh -c "read -q '?Can you see this? '" < /dev/null
> 
> But the following still sends the prompt to the redirected stderr, which
> seems wrong to me:
> 
> 	zsh -c "read -q '?Can you see this? '" < /dev/null >& /dev/null
> 
> The goal is to read from the tty only after also prompting there.  Patch
> follows for consideration; I have not committed it, pending commentary.
> The first hunk is the interesting one: It seems to me there's never any
> longer a reason to send the prompt to stderr, but I may be missing some
> case where ZLE is active and yet SHTTY == -1.  (?)

Actually, I was tempted to do the same when I sent the patch, then
tried to act minimally invasive or something like that -- and couldn't 
think of the example you gave.

> The other two hunks just avoid a redundant close().

Oops. If you commit the patch... there is similar code in bin_vared()
in zle_main.c. Implemented by the same idiot. Ahem.


Bye
 Sven


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

Re: 3.1.7-pre-1 for the workers

[Peter Stephenson]

"Bart Schaefer" wrote:
> I just noticed this in the zsh-users announcement:
> 
> Peter Stephenson wrote:
> }
> } This means it's now time to start smoothing the remaining rough corners of
> } for the release, which I hope will be the last before 4.0.1.
> 
> 4.0.1 ?  Are we going to skip 4.0.0 this time?  Why?

Everyone knows versions ending in `0' don't work.  But it's only really
because I'm one of the few people left who believes first things have a `1'
at the end.  I have no strong views.

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>
Work: pws@CambridgeSiliconRadio.com
Web: http://www.pwstephenson.fsnet.co.uk

Re: 3.1.7-pre-1 for the workers

[Oliver Kiddle]

Peter Stephenson wrote:
> 
> "Bart Schaefer" wrote:
> > 4.0.1 ?  Are we going to skip 4.0.0 this time?  Why?
> 
> Everyone knows versions ending in `0' don't work.  But it's only really
> because I'm one of the few people left who believes first things have a `1'
> at the end.  I have no strong views.

I don't really have any strong views either but it seems a bit unusual.
I often think in terms of count from one and index from zero but I'm not
sure which case I'd consider version numbers to be. I think that the
main problem with calling it 4.0.1 is that some people might assume it
is only a minor release. This might affect how it gets reported on sites
like freshmeat and appwatch.com and possibly cause it to be filtered out
by anyone not interested in minor upgrades.

This reminds me: a number of the files in CVS seem to have a version
number which is something like 1.1.1.1. The CVS info page says something
about this having some special significance but I couldn't work out what
exactly. Does anyone know? When 4.0 is released, it might be a good idea
to advance all the files to be version 4.1 anyway, especially if the
1.1.1 versions would cause problems if we created a branch later.

Oliver Kiddle

Re: 3.1.7-pre-1 for the workers

[Bart Schaefer]

On Apr 26,  9:11pm, Oliver Kiddle wrote:
} Subject: Re: 3.1.7-pre-1 for the workers
}
} This reminds me: a number of the files in CVS seem to have a version
} number which is something like 1.1.1.1. The CVS info page says something
} about this having some special significance but I couldn't work out what
} exactly. Does anyone know? When 4.0 is released, it might be a good idea
} to advance all the files to be version 4.1 anyway, especially if the
} 1.1.1 versions would cause problems if we created a branch later.

No, no, no, no, no.

Version numbers in CVS are effectively magical.  They're used by CVS
internally to determine branching properties and miscellaneous other
details related to the functioning of the system.  They no longer even
have any meaningful relationship to RCS revision numbers, from which
they originated.  Forcing CVS to use particular numbers for a revision
in the repository isn't the worst thing you can do to it, but it's not
very far from.

Please believe what I'm telling you.  I recently spent most of a day
fighting with CVS at one of Zanshin's consulting customer sites because
exactly this sort of thing had been done to their repository at one time
in the past.

The Right Way to do versioning in CVS is with symbolic tags.  With the
sole exception of "cvs import -b ..." you should treat the numbering
system as a black box; and the rule for "cvs import -b" is "don't use
this under normal circumstances, but if you must, then always use a
number of the form 1.1.<odd-number-greater-than-2>".

Without going into any great detail ...

Revision 1.1.1 is what's called the "vendor branch."  We have one because
we started from a copy of Tanaka's repository, and he imported zsh as if
it were source from a third-party vendor (which it was at the time he set
his repository up).  File revisions on the vendor branch will always have
four numbers (three dots), beginning with 1.1.1.  So a file with revision
1.1.1.1 is the first real revision on the vendor branch, which means only
that it's a file that has not changed since Tanaka started his repository.
It will not "cause problems" with any CVS operation we might want to do.

Revisions with only two numbers (one dot) are "on the trunk."  Revisions
from the vendor branch (or even other branches, if the file was originally
"cvs add"ed in a sandbox that was checked out on a branch) can be "on the
trunk," too, because CVS doesn't actually assign a number to a revision
until it has a non-empty delta with the revision that preceded it.  The
trunk is where mainline development normally takes place.

When we make a stable release (3.1.7 might count, but maybe we should wait
for 4.0.whatever-PWS-decides) we should probably create a branch that is
explicitly for patches to the 4.0 series, while continuing the next (4.1?)
development version on the trunk.

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

Re: 3.1.7-pre-1 for the workers

[Oliver Kiddle]

Bart Schaefer wrote:

> No, no, no, no, no.
> 
> Version numbers in CVS are effectively magical.  They're used by CVS
> internally to determine branching properties and miscellaneous other

[...much useful info snipped]

Okay, that makes sense, sorry. I'll shut-up about CVS until I know what
I'm talking about now.

On the subject of releases (as opposed to CVS revisions), what is the
situation with 3.0.8? If I haven't missed something, Bart only made a
preliminary release for us to test (some while ago) and there isn't a
final 3.0.8 on the ftp server, yet the latest zsh guide now talks about
the latest stable release being 3.0.8.

Oliver

Re: 3.1.7-pre-1 for the workers

[Bart Schaefer]

On Apr 28, 11:51am, Oliver Kiddle wrote:
} Subject: Re: 3.1.7-pre-1 for the workers
}
} On the subject of releases (as opposed to CVS revisions), what is the
} situation with 3.0.8?

I have a couple of issues that came up after the prerelease to resolve.
I should just get busy and resolve them, sigh.  They are:

The locale globbing thing.  I missed that it had already been changed in
3.1.6+, because I was expecting the strcoll() in glob.c to go away too.

The "read -q" bug that Sven patched in 10727.

The <n> == <n-> globbing thing that Zefram patched in 10444.  I'm tempted
simply to not deal with it, but  the "initial digit gets swallowed" bug
bothers me.

The "no such job" problem as discussed in 10016.

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

3.1.7-pre-2

[Peter Stephenson]

I have uploaded this at
 ftp://ftp.zsh.org/zsh/development/zsh-3.1.7-pre-2.tar.gz

2000-04-30  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* 11033: Completion/Commands/.distfiles,
	Completion/User/.distfiles, Config/version.mk, Etc/NEWS:
	new NEWS and minor fixes for version 3.1.7-pre-2.

2000-04-30  Bart Schaefer  <schaefer@zsh.org>

	* 11032: Src/init.c: Fix signed-ness problem reported by OPK.

2000-04-30  Oliver Kiddle  <opk@zsh.org>

	* 11031: Src/math.c, Src/glob.c, Src/module.c, Src/Zle/compcore.c
	Src/Zle/compresult.c: AIX dependency fixes.

2000-04-30  Bart Schaefer  <schaefer@zsh.org>

	* 11029: Src/hashtable.c, Src/signals.c: Back out more bits of
	11015 made obsolete by 11026.

2000-04-30  Clint Adams  <schizo@debian.org>

	* 11027: Josip Rodin: Completion/User/_joe:
	completion for joe.

2000-04-30  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* 11026: Src/builtin.c, Src/hashtable.c, Src/signals.c,
	Test/08traps.ztst: fix unfunction problem for localtraps by
	always copying traps to be saved, hence decoupling saved and actual
	traps.

2000-04-30  Bart Schaefer  <schaefer@zsh.org>

	* 11024: Completion/Commands/_complete_debug, Doc/Zsh/compsys.yo:
	Completion widget that enables and captures xtrace output.

2000-04-29  Adam Spiers  <adam@spiers.net>

	* 11019: Doc/Zsh/compsys.yo: fix bug in _description example

	* 11017: Completion/Linux/_rpm: fixes for --rmsource and
	--recompile options, and for breakage caused by 10772

2000-04-28  Bart Schaefer  <schaefer@zsh.org>

	* 11015: Src/builtin.c, Src/hashtable.c, Src/signals.c:
	`unfunction TRAPxxx' now works with localtraps in effect.

2000-04-28  Peter Stephenson  <pws@cambridgesiliconradio.com>

	* pws: 11013: Completion/Core/compinstall: more breaks and
	a completion context.

2000-04-28  Bart Schaefer  <schaefer@zsh.org>

	* 11012: Doc/Zsh/intro.yo, Doc/Zsh/manual.yo: Reorder menus
	to match sections in 11008.

2000-04-28  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 11008: Doc/zsh.yo, Doc/Zsh/compctl.yo, Doc/Zsh/compsys.yo,
 	Doc/Zsh/compwid.yo, Doc/Zsh/modules.yo, Doc/Zsh/seealso.yo,
 	Doc/Zsh/zle.yo: move compctl docs behind compsys, mention
 	`zcompile -z' in compsys
	
	* 11007: Src/Zle/compmatch.c: misplaced cursor with partial-word
 	matching
	
	* Tanaka: 11001: Completion/User/_enscript, Completion/User/_lp,
	Completion/User/_printers: printer-detection moved to new
	_printers; used by _enscript

	* 11003: Src/Zle/computil.c: fix for reporting multiple actions in
 	comparguments, options with non-optional arguments
	
2000-04-28  Peter Stephenson  <pws@cambridgesiliconradio.com>

	* pws: 10999: Completion/Core/compinstall: some idiot didn't
	get searching of alternative files for definitions to work properly.

2000-04-28  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10992: Src/Zle/complist.c: avoid wrapping around both
 	horizontally and vertically at the same time
	
2000-04-27  Bart Schaefer  <schaefer@zsh.org>

	* Tanaka Akira: 10980: Completion for enscript.

2000-04-27  Peter Stephenson  <pws@pwstephenson.fsnet.co.uk>

	* 10985: Completion/Core/compinstall: _oldlist, list-prompt and
	matcher-list enhancements.

2000-04-27  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10979: originally from Felix: 10894: Completion/Core/.distfiles,
	Completion/Core/_history, Doc/Zsh/compsys.yo: new _history completer
	
	* 10978: Src/Zle/complist.c: better wrap around for cursor
	movement widgets; avoid flicker when leaving menu-selection
	
	* 10977: and Andrej: 10976: Completion/User/_make: use $words[1]
 	to call make in _make and throw away stderr and use _call
	
	* 10974: Completion/Base/_arguments: ensure $PWD is prepended to
 	cache-names for long options in _arguments
	
2000-04-27  Peter Stephenson  <pws@cambridgesiliconradio.com>

	* 10971: Src/pattern.c: noglobdots didn't work with the PAT_ANY
	flag, which meant that **/ always globbed dots.

2000-04-27  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10967: Completion/Base/_arguments: fix for automatic
 	option-argument completion with partially typed option in _arguments
	
	* 10965: and Felix: 10952: Completion/User/_urls,
 	Doc/Zsh/compsys.yo: document _wanted -C; fix for call to _wanted
 	in _urls
	
	* 10963: Src/Zle/complete.c: fix for parse error with
 	left-anchored patterns
	
	* 10961: Src/Zle/complist.c: make menu-selection handle
 	undefined-key
	
	* 10959: Completion/Core/_main_complete, Doc/Zsh/compsys.yo,
 	Doc/Zsh/mod_complist.yo, Src/params.c, Src/Zle/compcore.c,
 	Src/Zle/complete.c, Src/Zle/complist.c, Src/Zle/compresult.c:
 	renamed parameters for menu selection to MENUSELECT, MENUPROMPT
 	and MENUSCROLL; no special value for LISTMAX to turn on list
 	scrolling, this is done by setting LISTPROMPT now; default values
 	for LISTPROMPT and MENUPROMPT when set but empty
	
2000-04-27  Adam Spiers <adam@spiers.net>

	* 10949: Completion/User/_perl_modules: print warning when
	perl not found on PATH forces guessing of @INC

2000-04-26  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10933: Completion/Commands/_bash_completions,
 	Completion/Commands/_history_complete_word: ensure options are set
 	as expected in _history_complete_word and _bash_completions
	
	* Adam Spiers: 10892: Completion/User/_perl_modules: ensure perl
	is there and better pattern for searching modules

	* Tanaka Akira: 10890: Completion/User/_gunzip,
	Completion/User/_gzip, Completion/User/_zcat: new completion for
	GNU zip commands

	* Tanaka Akira: 10885: Completion/X/_xset: new completion function
	for xset

	* 10926: Completion/X/_x_color, Doc/Zsh/compwid.yo, Src/Zle/complete.c,
 	Src/Zle/compmatch.c: allow to anchor on gaps between strings in
 	matching control specs
	
	* 10925: Src/Zle/compmatch.c: faster test for `*'-patterns in
 	matching control; prefer direct character matches over match specs
 	in recursive invocations of match_str()
	
	* 10924: Src/Zle/compmatch.c: fix for matching control; improve
 	merging cline lists
	
	* 10923: Completion/Core/_main_complete: better default value for
 	list-prompt style
	
	* 10922: Completion/Core/_path_files: file-sort made $fignore fail
	
2000-04-25  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10917: Src/Zle/complist.c, Src/Zle/compresult.c: don't show
 	empty hidden matches even in menu-selection
	
	* 10916: Completion/Core/_main_complete, Src/Zle/complist.c:
 	re-calculate number of line with mark when needed; make default
 	prompts more consistent
	
	* 10915: Src/Zle/computil.c: forgotten fix for _arguments using
 	multiple actions
	
	* 10912: Src/Zle/compresult.c: don't always ignore backslashes in
 	paths when testing file type
	
	* 10910: Completion/Base/_tilde, Completion/Core/_all_labels,
 	Completion/Core/_requested: small fix for _requested to be able to
 	use _all_labels from within it
	
	* 10909: Completion/Core/_expand, Src/Zle/compcore.c,
 	Src/Zle/compresult.c: report prefix/suffix of parameter expansion
 	in IPREFIX/ISUFFIX; make _expand use them to be able to expand $foo
	
	* 10908: Completion/Base/.distfiles, Completion/Base/_argument_sets,
 	Completion/Base/_arguments, Completion/Base/_describe,
 	Completion/Builtins/_bindkey, Completion/Builtins/_compdef,
 	Completion/Builtins/_emulate, Completion/Builtins/_zpty,
 	Completion/Core/_tags, Doc/Zsh/compsys.yo,
 	Etc/completion-style-guide, Src/Zle/computil.c: make _arguments
 	use more than one action when appropriate; add _argument_sets to
 	complete different sets of arguments and options for the same command
	
2000-04-24  Bart Schaefer  <schaefer@zsh.org>

	* 10900: Src/Makefile.in: Replace a dependency on Makefile with a
	more accurate dependency on config.status.

2000-04-20  Sven Wischnowsky  <wischnow@informatik.hu-berlin.de>

	* 10881: Functions/Misc/zrecompile: don't remove *all* .zwc files,
 	keep the first one
	
	* 10868: Src/Zle/complist.c: fix off-by-one error in scrolled
 	menu-selections
	
	* 10861: Completion/Bsd/.distfiles, Completion/Bsd/_bsd_pkg,
 	Completion/Bsd/_cvsup, Completion/Bsd/_kld: new completions for
 	FreeBSD commands
	
	* 10854: Completion/Commands/_expand_word, Completion/Core/_expand,
 	Doc/Zsh/compsys.yo: default values for styles in _expand_word,
 	listing expansions with ^Xd, _expand uses add-space style
	
2000-04-20  Oliver Kiddle  <opk@zsh.org>

	* 10857: Completion/AIX/.distfiles Completion/AIX/_floppy
	Completion/AIX/_lsdev Completion/AIX/_lslv Completion/AIX/_lspv
	Completion/AIX/_lsvg Completion/AIX/_physical_volumes
	Completion/AIX/_smit Completion/AIX/_volume_groups: new completions
	for AIX commands

	* 10848: Completion/Builtins/_print, Completion/Core/.distfiles,
	Completion/Core/_file_descriptors: Add completion for file
	descriptors

-- 
Peter Stephenson <pws@pwstephenson.fsnet.co.uk>
Work: pws@CambridgeSiliconRadio.com
Web: http://www.pwstephenson.fsnet.co.uk

Re: 3.1.7-pre-2

[Bart Schaefer]

On Apr 30,  4:45pm, Bart Schaefer wrote:
} Subject: Re: 3.1.7-pre-1 for the workers
}
} I have a couple of issues that came up after the prerelease to resolve.
} 
} The "read -q" bug that Sven patched in 10727.

So ... 10727 makes this work:

	zsh -c "read -q '?Can you see this? '" < /dev/null

But the following still sends the prompt to the redirected stderr, which
seems wrong to me:

	zsh -c "read -q '?Can you see this? '" < /dev/null >& /dev/null

The goal is to read from the tty only after also prompting there.  Patch
follows for consideration; I have not committed it, pending commentary.
The first hunk is the interesting one: It seems to me there's never any
longer a reason to send the prompt to stderr, but I may be missing some
case where ZLE is active and yet SHTTY == -1.  (?)

The other two hunks just avoid a redundant close().

Index: Src/builtin.c
===================================================================
RCS file: /extra/cvsroot/zsh/zsh-3.1/Src/builtin.c,v
retrieving revision 1.86
diff -u -r1.86 builtin.c
--- builtin.c	2000/04/30 18:54:07	1.86
+++ builtin.c	2000/05/01 02:32:28
@@ -3420,8 +3420,8 @@
 	     *readpmpt && *readpmpt != '?'; readpmpt++);
 	if (*readpmpt++) {
 	    if (keys || isatty(0)) {
-		zputs(readpmpt, (haso ? shout : stderr));
-		fflush(haso ? shout : stderr);
+		zputs(readpmpt, shout);
+		fflush(shout);
 	    }
 	    readpmpt[-1] = '\0';
 	}
@@ -3461,8 +3461,7 @@
 	    else
 		settyinfo(&shttyinfo);
 	    if (haso) {
-		close(SHTTY);
-		fclose(shout);
+		fclose(shout);	/* close(SHTTY) */
 		shout = oshout;
 		SHTTY = -1;
 	    }
@@ -3494,8 +3493,7 @@
 
 	    /* dispose of result appropriately, etc. */
 	    if (haso) {
-		close(SHTTY);
-		fclose(shout);
+		fclose(shout);	/* close(SHTTY) */
 		shout = oshout;
 		SHTTY = -1;
 	    }

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