Next pretest

Next pretest

[Chong Yidong]

I would like to make the next pretest release (23.1.95) this coming
Friday, the 2nd of April.  Let me know if there are no objections, or
any problems that you'd like to see addressed.

RE: Next pretest

[Drew Adams]

> I would like to make the next pretest release (23.1.95) this coming
> Friday, the 2nd of April.  Let me know if there are no objections, or
> any problems that you'd like to see addressed.

How about a Windows binary?

You published 23.1.93 without a binary, giving this as excuse:

> The latest pretest has a couple of serious bugs on Windows, 
> so IMO there is little point making the binary for it.

But you reassured us that you would publish again soon, with those Windows bugs
fixed:

> > Fair enough. Is there a point in publishing a new release 
> > without testing on Windows, in particular, testing the fixes
> > for those serious bugs?
> >
> > Or did you plan to publish another pretest with the bugs fixed?
> > That would be good (normal). That wasn't clear from your mails.
> 
> There should be at least two more pretests before the 23.2 release.

However, you published 23.1.94 without a Windows binary (giving no reason). You
said this on 3/12, but nothing came of it:

> Lennart is having technical problems with the box he compiles 
> on. If he fixes those, I will upload his binaries to alpha.gnu.org.

How many Windows users tested that 94 pretest, which supposedly fixed serious
Windows bugs introduced by pretest 93? Gettin any pretest feedback at all from
Windows users, besides build reports?

There was no 23.1.92 binary either. The last published Windows binary was for
23.1.91, 2010/01/03.

You have lamented the fact that no one has volunteered to build Windows
binaries, but the same day (3/12), Sean Sieger volunteered:

> I've been building and using the previous pretest as well as 
> bzr builds daily for two weeks and everything seems to be
> going well. If it's okay with you, I would like to volunteer
> to compile the Windows binaries and get them (pretests,
> releases and say weekly bzr builds) uploaded to alph.gnu.org
> and ftp.gnu.org.
> 
> In those two weeks I've been careful to read Jason's
> responses to Eli's requests and the _Information for
> Maintainers of GNU Software_ Section 9, `Distributions'.
> 
> There's a couple of questions that I have for Jason before
> I do, but I would like to volunteer to compile the binaries
> for Windows and keep the windows directories current with them.

Whatever became of that? Wasn't it "okay with you" that Sean build and post
binaries?

Re: Next pretest

[Lennart Borgman]

On Sun, Mar 28, 2010 at 9:59 PM, Drew Adams <drew.adams@oracle.com> wrote:
>> I would like to make the next pretest release (23.1.95) this coming
>> Friday, the 2nd of April.  Let me know if there are no objections, or
>> any problems that you'd like to see addressed.
>
> How about a Windows binary?


I will be away. However I could perhaps upload something now to my
site instead. How about that?

BTW I think my checkout is a trunk checkout. How do I get a checkout
from the pretest branch from that? (I am using Launchpad.)

RE: Next pretest

[Drew Adams]

> >> I would like to make the next pretest release (23.1.95) this coming
> >> Friday, the 2nd of April.  Let me know if there are no 
> >> objections, or any problems that you'd like to see addressed.
> >
> > How about a Windows binary?
> 
> I will be away. However I could perhaps upload something now to my
> site instead. How about that?

Every little bit helps, of course. Thanks, Lennart.

However, that's not the real solution. Pretest testers will (and should) look
for Windows binary pretests at the GNU location created for them:

http://alpha.gnu.org/gnu/emacs/pretest/windows/

That's the only purpose of that directory, as stated in the README.

And pretest binaries should correspond to the pretest source code exactly.
Otherwise, we can be comparing apples and oranges (which can mean confusion and
more work for everyone involved).

And `report-emacs-bug' should correctly address the pretest mailing list (that
didn't happen the last time you made a binary available informally for pretest
testing).

So yes, thank you - that will help. But whatever happened to the Windows builds
volunteered by Sean Sieger?

Re: Next pretest

[Christoph]

On 3/28/2010 2:18 PM, Lennart Borgman wrote:
> On Sun, Mar 28, 2010 at 9:59 PM, Drew Adams<drew.adams@oracle.com>  wrote:
>    
>>> I would like to make the next pretest release (23.1.95) this coming
>>> Friday, the 2nd of April.  Let me know if there are no objections, or
>>> any problems that you'd like to see addressed.
>>>        
>> How about a Windows binary?
>>      
> I will be away. However I could perhaps upload something now to my
> site instead. How about that?
>    

I can build the Windows binaries for the pretest and upload them to the 
site.

I will have a look at the documentation tonight and see if I need some 
help for this the first time.

Just let me know if you want me to do this.

Christoph

Re: Next pretest

[Lennart Borgman]

On Sun, Mar 28, 2010 at 10:36 PM, Drew Adams <drew.adams@oracle.com> wrote:
>> >> I would like to make the next pretest release (23.1.95) this coming
>> >> Friday, the 2nd of April.  Let me know if there are no
>> >> objections, or any problems that you'd like to see addressed.
>> >
>> > How about a Windows binary?
>>
>> I will be away. However I could perhaps upload something now to my
>> site instead. How about that?
>
> Every little bit helps, of course. Thanks, Lennart.
>
...
> And pretest binaries should correspond to the pretest source code exactly.
> Otherwise, we can be comparing apples and oranges (which can mean confusion and
> more work for everyone involved).


Thanks. I am including a file with something like this now:

;; bzr checkout info for Emacs unpatched version
(defconst emacs-trunk-bzr-revision-id
"schwab@linux-m68k.org-20100321115749-axlxbdz2vt3sdbf6")
(defconst emacs-trunk-bzr-revno 99712)
(defconst emacs-trunk-bzr-date "2010-03-21 12:57:49 +0100")
(defconst emacs-trunk-bzr-branch-nick "trunk")

I suggested that should be recommended in the build process and
included in bug reports, but I got no responce.


> And `report-emacs-bug' should correctly address the pretest mailing list (that
> didn't happen the last time you made a binary available informally for pretest
> testing).


I have forgotten. Was I using the wrong branch, or?

Re: Next pretest

[Chong Yidong]

"Drew Adams" <drew.adams@oracle.com> writes:

> You have lamented the fact that no one has volunteered to build Windows
> binaries, but the same day (3/12), Sean Sieger volunteered:
>
>> I've been building and using the previous pretest as well as
>> bzr builds daily for two weeks and everything seems to be
>> going well. If it's okay with you, I would like to volunteer
>> to compile the Windows binaries and get them (pretests,
>> releases and say weekly bzr builds) uploaded to alph.gnu.org
>> and ftp.gnu.org.

> Whatever became of that? Wasn't it "okay with you" that Sean build and
> post binaries?

The problem of Windows binaries availability is on my radar, and we have
been working on it.  Sean has been having extensive problems uploading
his binaries, which we are currently trying to resolve with the FSF
sysadmins.

RE: Next pretest

[Drew Adams]

> The problem of Windows binaries availability is on my radar, 
> and we have been working on it.  Sean has been having
> extensive problems uploading his binaries, which we are
> currently trying to resolve with the FSF sysadmins.

That's good to hear. Thanks for the progress report and thanks for working on
this.

Re: Next pretest

[Eli Zaretskii]

> From: Lennart Borgman <lennart.borgman@gmail.com>
> Date: Sun, 28 Mar 2010 22:48:22 +0200
> Cc: Chong Yidong <cyd@stupidchicken.com>, Sean Sieger <sean.sieger@gmail.com>,
> 	emacs-devel@gnu.org
> 
> > And pretest binaries should correspond to the pretest source code exactly.
> > Otherwise, we can be comparing apples and oranges (which can mean confusion and
> > more work for everyone involved).
> 
> 
> Thanks. I am including a file with something like this now:
> 
> ;; bzr checkout info for Emacs unpatched version
> (defconst emacs-trunk-bzr-revision-id
> "schwab@linux-m68k.org-20100321115749-axlxbdz2vt3sdbf6")
> (defconst emacs-trunk-bzr-revno 99712)
> (defconst emacs-trunk-bzr-date "2010-03-21 12:57:49 +0100")
> (defconst emacs-trunk-bzr-branch-nick "trunk")

This is a trunk checkout, whereas Drew (see above) wants a pretest
binary, which should be built from the emacs-23 branch.

Re: Next pretest

[Sean Sieger]

Eli Zaretskii <eliz@gnu.org> writes:

    This is a trunk checkout, whereas Drew (see above) wants a pretest
    binary, which should be built from the emacs-23 branch.

I'll keep trying to get the 23.1.94 binary uploaded and hope that by
Friday I'll be able to upload the 23.1.95 pretest.

Re: Next pretest

[Eli Zaretskii]

> From: Sean Sieger <sean.sieger@gmail.com>
> Date: Mon, 29 Mar 2010 11:41:40 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
>     This is a trunk checkout, whereas Drew (see above) wants a pretest
>     binary, which should be built from the emacs-23 branch.
> 
> I'll keep trying to get the 23.1.94 binary uploaded and hope that by
> Friday I'll be able to upload the 23.1.95 pretest.

Thank you for your efforts.

Re: Next pretest

[Glenn Morris]

Chong Yidong wrote:

> any problems that you'd like to see addressed.

It would be good to deal with the six remaining "serious" bugs:

http://debbugs.gnu.org/severity:serious

2056 is NS-only, but seems to bother a lot of people

2503 is the Windows version of the same, maybe it is not as widespread
     or really "serious"

3347 I don't know

4431 message versus mail mode is hopefully better now, and perhaps is
     already as good is it is going to get

4656 Personally I think we should perhaps just remove many of these
     files, especially test.py

5162 probably was fixed by Andreas's change (note does not apply to
     Windows), or at least is no longer "serious"

Re: Next pretest

[Leo]

On 2010-03-29 19:33 +0100, Glenn Morris wrote:
> 2056 is NS-only, but seems to bother a lot of people
>
> 2503 is the Windows version of the same, maybe it is not as widespread
>      or really "serious"

The Emacs mac-port has no such bug maybe something can be taken from
there.

Leo

bug#4656:

[Chong Yidong]

Glenn Morris <rgm@gnu.org> writes:

> The only remaining issue is the test/cedet/tests/ directory.
>
> EML says:
>
>     Many of the test files are just scratch files where I put random
>     code snippets for manual testing.
>
>     All the test files in semantic/tests that are part of the
>     automated suite are referred to in semantic/semantic-ia-utest.el,
>     semantic-lex-spp.el, and semantic-utest-c.el.

I have removed some of the files for which provenance is uncertain.  For
those clearly written by Eric (after checking the CEDET CVS logs), I
have added a copyright header.

There are now only three files that do not have a copyright header:

  testfriends.cpp
  testnsp.cpp
  testsppcomplete.c

These are all very short snippets, and I do not think we need to worry
about these.  Adding a copyright header would double the length of each
of these files.

In any case, the test/ directory is not distributed with the tarball, so
I would not consider this bug a blocker for the 23.2 release.

Re: Next pretest

[Chong Yidong]

Glenn Morris <rgm@gnu.org> writes:

> 4431 message versus mail mode is hopefully better now, and perhaps is
>      already as good is it is going to get

This is pretty much done.  I've finished the necessary documentation
changes over the weekend, and any remaining bugs can be dealt with on a
case by case basis.  I'll close this bug.

> 4656 Personally I think we should perhaps just remove many of these
>      files, especially test.py

I've just fixed this, mostly.  Discussion about what to do with the
remaining ~25 line snippet unit test files, copyright-wise, can go into
the thread for that bug.  Since the tests are not distributed with the
tarball anyway, I'm downgrading the severity to "normal".

> 5162 probably was fixed by Andreas's change (note does not apply to
>      Windows), or at least is no longer "serious"

I'll check, and comment on the bug.

Re: Next pretest

[Glenn Morris]

Chong Yidong wrote:

> I've just fixed this, mostly.  Discussion about what to do with the
> remaining ~25 line snippet unit test files, copyright-wise, can go into
> the thread for that bug.  Since the tests are not distributed with the
> tarball anyway, I'm downgrading the severity to "normal".

Oh, sorry; for some reason I thought "tests/" was in the tarfile.
Thanks.