Updates to zsh-4_2-patches branch

Updates to zsh-4_2-patches branch

[Paul Ackersviller]

I promised to do this months ago, but didn't get far over the course
of the winter...  Excuses aside, I've been hard at it for the past
week or two, and have gotten up to about the middle of last year in
merging trunk changes onto the branch.  I started out a little too
agressively in trying to include everything possible, but quickly
dug myself into a hole.  Now that I have the basic strategy down,
I'd like to do some testing for a few days to ensure it all still
works before continuing onto the rest.

How comprehensive is the test suite, and are there any major gotchas?

One area I could use help or at least guidance on is files in Src/Zle/.
I've merged in only a few changes there, when it appeared to be
non-multibyte-related, but is it better to try leaving that completely
alone?  I fear a few multibyte changes that don't belong mat have crept
in, but I'll see how testing goes.

In terms of documenting what I've done, is there any value in sending
all the change numbers to this list?  There will be very many, so I'm
thinking those numbers would better go into the change comments for the
individual cvs revisions.

Re: Updates to zsh-4_2-patches branch

[Peter Stephenson]

Paul Ackersviller wrote:
> I promised to do this months ago, but didn't get far over the course
> of the winter...  Excuses aside, I've been hard at it for the past
> week or two, and have gotten up to about the middle of last year in
> merging trunk changes onto the branch.

Thanks for the work, it's very useful.

> How comprehensive is the test suite, and are there any major gotchas?

The test suite is good at stuff related to scripting and syntax,
moderately good but rather incomplete at the low levels of completion
and line editing, and poor at things which are only testable
interactively like job control.

> One area I could use help or at least guidance on is files in Src/Zle/.
> I've merged in only a few changes there, when it appeared to be
> non-multibyte-related, but is it better to try leaving that completely
> alone?  I fear a few multibyte changes that don't belong mat have crept
> in, but I'll see how testing goes.

Zle has changed so much that it might well be better to leave anything
alone that isn't obviously a standalone fix for a specific problem
that applies cleanly.  We've made various tweaks to simplify interfaces
both internally and to related parts of the main shell such as quoting
to make adding multibyte support easier; these aren't generally
bug fixes and don't need applying.

> In terms of documenting what I've done, is there any value in sending
> all the change numbers to this list?  There will be very many, so I'm
> thinking those numbers would better go into the change comments for the
> individual cvs revisions.

It would be fine if they just appeared in the ChangeLog on the branch.

-- 
Peter Stephenson <p.w.stephenson@ntlworld.com>
Web page now at http://homepage.ntlworld.com/p.w.stephenson/

Re: Updates to zsh-4_2-patches branch

[Geoff Wing]

On Wednesday 2007-03-28 10:21 +1000, Paul Ackersviller output:
:How comprehensive is the test suite, and are there any major gotchas?

------------------------------------------------------------
2007-05-06 01:47  packersv

        * Test/B03print.ztst (1.8.2.1): Merge through trunk revision 1.19
          for 21352, 22758, 23086, 23118, 23119 and a few unposted.
------------------------------------------------------------

23118/23119 weren't brought in.  Are you planning on this still?

./C03traps.ztst fails for me on "Nested `trap ... INT', not triggered".

I'm still looking through the recent stuff (last six weeks or so) so I
may have more questions.

How big is the list of what remains to be done?

Regards,
Geoff

Re: Updates to zsh-4_2-patches branch

[Paul Ackersviller]

On Tue, Oct 30, 2007 at 01:01:41PM +1100, Geoff Wing wrote:
> 2007-05-06 01:47  packersv
> 
>         * Test/B03print.ztst (1.8.2.1): Merge through trunk revision 1.19
>           for 21352, 22758, 23086, 23118, 23119 and a few unposted.
> ------------------------------------------------------------
> 
> 23118/23119 weren't brought in.  Are you planning on this still?

No, thanks for catching that.  I'm syncing with the HEAD mostly just for
completions, I don't know why I did here.  I've reversed out 23118/23119
and had a look for any similar mistakes, but didn't see any in Test/.

> ./C03traps.ztst fails for me on "Nested `trap ... INT', not triggered".

I'll look into that soon.

> I'm still looking through the recent stuff (last six weeks or so) so I
> may have more questions.

Sure, fire away.

> How big is the list of what remains to be done?

There shouldn't be much left now, but I can't give you an accurate answer
until I strip down the branched ChangeLog (probably worth doing anyway).

Re: C03 test failure on zsh-4_2-patches branch

[Paul Ackersviller]

On Fri, Nov 02, 2007 at 04:03:38AM +0000, Paul Ackersviller wrote:
> On Tue, Oct 30, 2007 at 01:01:41PM +1100, Geoff Wing wrote:
> 
> > ./C03traps.ztst fails for me on "Nested `trap ... INT', not triggered".
> 
> I'll look into that soon.

That's gone now, at the cost of backing out 21296 from the branch.

It seemed straightforward enough to put in at the time, however the
builtin.c source file had two (small) merge conflicts that I'd thought
it appeared safe enough to resolve manually.  Apparently that was not
the case, and I'll not try to outsmart myself one more time.  Therefore
I've reversed out 21296 from the three source files and two document ones
on the 4.2 branch.

Just for the record, that is be the only change that had source code
which didn't merge onto the branch completely cleanly.  There's one
document, Doc/Zsh/roadmap.yo, but I'm leaving other changes out.