Emacs is coming back - please test

Emacs is coming back - please test

[Steffen Sledz]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Here are the results of my first try to build official emacs packages of
version 22.1. Because this is my first work as a maintainer it is marked
as experimental. It would be nice if someone could check them and give
me some feedback.

The base package including everything to run without X11 and the sources:

<http://www.zone42.org/cygwin/emacs/emacs-22.1-1.tar.bz2>
<http://www.zone42.org/cygwin/emacs/emacs-22.1-1-src.tar.bz2>
<http://www.zone42.org/cygwin/emacs/setup.hint>

The X11 binary:

<http://www.zone42.org/cygwin/emacs/emacs-X11/emacs-X11-22.1-1.tar.bz2>
<http://www.zone42.org/cygwin/emacs/emacs-X11/setup.hint>

The lisp sources (including LEIM sources):

<http://www.zone42.org/cygwin/emacs/emacs-el/emacs-el-22.1-1.tar.bz2>
<http://www.zone42.org/cygwin/emacs/emacs-el/setup.hint>

LEIM is obsolete as an own package. This is an empty package moving
emacs-leim to _obsolete:

<http://www.zone42.org/cygwin/emacs/emacs-leim/emacs-leim-22.1-1.tar.bz2>
<http://www.zone42.org/cygwin/emacs/emacs-leim/setup.hint>

You can also install directly from:

<http://www.zone42.org/cygwin/setup.ini>

Steffen Sledz
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGkWP8lSfCk8EgCzgRAj85AKCsx90gMPFPvfzygZbih7gAw63E9QCgr9z2
UsroyoIBYJL4zGLl4xTYPdw=
=/nvy
-----END PGP SIGNATURE-----

Re: Emacs is coming back - please test

[Eric Blake]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

According to Steffen Sledz on 7/8/2007 4:23 PM:
> Here are the results of my first try to build official emacs packages of
> version 22.1. Because this is my first work as a maintainer it is marked
> as experimental. It would be nice if someone could check them and give
> me some feedback.

I will be trying them out this week.  But first, a suggestion.

Currently, emacs and emacs-X11 both provide /usr/bin/emacs.exe.  Which
means whichever package is installed last wins.  I would much rather that
you used the 'alternatives' package, and instead of directly providing
/usr/bin/emacs.exe, instead provide two differently-named binaries, and
use the postinstall script to install symlinked alternatives where the X11
version has a higher default priority, but where the user can use the
alternatives package to choose which of the two they prefer.  Then,
upgrading just the emacs package won't wipe out the emacs-X11 package
(which has happened to me in the past).  See autoconf for an example of
how the alternatives package can let you install symlinks to alternative
versions, so that competing packages don't install to the same filename.

- --
Don't work too hard, make some time for fun as well!

Eric Blake             ebb9@byu.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGkidD84KuGfSFAYARAiMoAJ9Je0aVo2AGq8F/+slGpMEF52hlHwCgsvsY
o40xJg2roXzueJOuvX46Op8=
=LOnX
-----END PGP SIGNATURE-----

Re: Emacs is coming back - please test

[Steffen Sledz]

On Mon, 9 Jul 2007, Eric Blake wrote:
> I will be trying them out this week.

Thx

> But first, a suggestion.
> Currently, emacs and emacs-X11 both provide /usr/bin/emacs.exe.  Which 
> means whichever package is installed last wins....

If I made it right they don't. Package emacs provides 
/usr/bin/emacs-nox.exe instead. But I'll have a look at that symlink 
alternatives stuff.

Steffen

Re: Emacs is coming back - please test

[Steffen Sledz]

Eric Blake schrieb:
> Currently, emacs and emacs-X11 both provide /usr/bin/emacs.exe.  Which
> means whichever package is installed last wins.  I would much rather that
> you used the 'alternatives' package, and instead of directly providing
> /usr/bin/emacs.exe, instead provide two differently-named binaries, and
> use the postinstall script to install symlinked alternatives where the X11
> version has a higher default priority, but where the user can use the
> alternatives package to choose which of the two they prefer.  Then,
> upgrading just the emacs package won't wipe out the emacs-X11 package
> (which has happened to me in the past).  See autoconf for an example of
> how the alternatives package can let you install symlinks to alternative
> versions, so that competing packages don't install to the same filename.

I'm sorry, but I can't find any binary symlinking stuff in autoconf2.5
postinstall, just handling of man and info pages. Did you mean another
package?

Steffen

PS: What are the arguments against the method I chose
(/usr/bin/emacs-nox.exe in emacs and /usr/bin/emacs.exe in emacs-X11)?
That's also the way the linux distributions I now handle it.

Re: Emacs is coming back - please test

[Eric Blake]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

According to Steffen Sledz on 7/8/2007 4:23 PM:
> Here are the results of my first try to build official emacs packages of
> version 22.1. Because this is my first work as a maintainer it is marked
> as experimental. It would be nice if someone could check them and give
> me some feedback.
> 
> The base package including everything to run without X11 and the sources:
> 
> <http://www.zone42.org/cygwin/emacs/emacs-22.1-1.tar.bz2>
> <http://www.zone42.org/cygwin/emacs/emacs-22.1-1-src.tar.bz2>
> <http://www.zone42.org/cygwin/emacs/setup.hint>

Why does this depend on libintl1?  Did you mean libintl8?

Why did the package grow from 8meg to 17meg?  Not necessarily a problem,
if there is a good reason for it, but it seems a bit odd.

> 
> The X11 binary:
> 
> <http://www.zone42.org/cygwin/emacs/emacs-X11/emacs-X11-22.1-1.tar.bz2>
> <http://www.zone42.org/cygwin/emacs/emacs-X11/setup.hint>

The binary worked for me on XP, with cygwin 1.5.24-2.  If you can adjust
your packaging to use alternatives (see my earlier mail), then I think you
are ready to have your package uploaded to cygwin.com, and that you
deserve a gold star for taking on this giant project.

Not so obvious is that since emacs has been changed so extensively between
21.x and 22.1, that all my .emacs customizations needed tweaking to align
to the new way of doing things (for example,
read-file-name-completion-ignore-case is new, whereas I used to use
completion-ignore-case).  But that is not a packaging bug, just upstream
progress.

> 
> The lisp sources (including LEIM sources):
> 
> <http://www.zone42.org/cygwin/emacs/emacs-el/emacs-el-22.1-1.tar.bz2>
> <http://www.zone42.org/cygwin/emacs/emacs-el/setup.hint>

Why did the package shrink from 4.5meg down to .6meg?  Could this account
for some of the increase in the size of the base emacs package?

> 
> LEIM is obsolete as an own package. This is an empty package moving
> emacs-leim to _obsolete:
> 
> <http://www.zone42.org/cygwin/emacs/emacs-leim/emacs-leim-22.1-1.tar.bz2>
> <http://www.zone42.org/cygwin/emacs/emacs-leim/setup.hint>

Looks correct for doing the obsoletion.

> 
> You can also install directly from:
> 
> <http://www.zone42.org/cygwin/setup.ini>
> 
> Steffen Sledz

- --
Don't work too hard, make some time for fun as well!

Eric Blake             ebb9@byu.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGki0Q84KuGfSFAYARAhEkAJ4wrMiUeWITiLS5jwv9YB7I7j/XwgCg18JO
r3emUpyNpx3a0V+Ue3Do9Xs=
=vYz5
-----END PGP SIGNATURE-----

Re: Emacs is coming back - please test

[Steffen Sledz]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Eric Blake schrieb:
> Why did the package grow from 8meg to 17meg?  Not necessarily a problem,
> if there is a good reason for it, but it seems a bit odd.

The emacs.exe binary itself is about 15MB (stripped). ;-)

> Why did the package shrink from 4.5meg down to .6meg?  Could this account
> for some of the increase in the size of the base emacs package?

It was a packaging error. A lot of files were missing. Thanx for the
hint. I'll fix this in the next release.

Steffen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGmme6lSfCk8EgCzgRAjnLAKC4p9/NVItzdHzCotfxlIqODroR8gCgkyg3
qq0ba+Hg4UhetRjDgdqwifs=
=9w/F
-----END PGP SIGNATURE-----

Re: Emacs is coming back - please test

[Dr. Volker Zell]

>>>>> Steffen Sledz writes:

    > Here are the results of my first try to build official emacs packages of
    > version 22.1. Because this is my first work as a maintainer it is marked
    > as experimental. It would be nice if someone could check them and give
    > me some feedback.

    > The base package including everything to run without X11 and the sources:

    > <http://www.zone42.org/cygwin/emacs/emacs-22.1-1.tar.bz2>
    > <http://www.zone42.org/cygwin/emacs/emacs-22.1-1-src.tar.bz2>
    > <http://www.zone42.org/cygwin/emacs/setup.hint>

I splitted out the following conflicting files from my xemacs package:

xemacs-tags:            (conflicts with ctags-5.6 and now with upcoming emacs-22.1)
  usr/bin/ctags.exe
  usr/bin/etags.exe
  usr/share/man/man1/ctags.1.gz
  usr/share/man/man1/etags.1.gz

xemacs-emacs-common:    (used to conflict with emacs-21.2 now with upcoming emacs-22.1)
  /usr/bin/b2m.exe
  /usr/bin/rcs-checkin

    > Steffen Sledz

Ciao
  Volker

Re: Emacs is coming back - please test

[Steffen Sledz]

On Tue, 10 Jul 2007, Dr. Volker Zell wrote:
> I splitted out the following conflicting files from my xemacs package:

What's the right way to handle such conflicts?

> 
> xemacs-tags:            (conflicts with ctags-5.6 and now with upcoming emacs-22.1)
>   usr/bin/ctags.exe
>   usr/bin/etags.exe
>   usr/share/man/man1/ctags.1.gz
>   usr/share/man/man1/etags.1.gz

Exclude from emacs and make ctags a requirement?

> xemacs-emacs-common:    (used to conflict with emacs-21.2 now with upcoming emacs-22.1)
>   /usr/bin/b2m.exe
>   /usr/bin/rcs-checkin

No idea.

Regards,
Steffen

Re: Emacs is coming back - please test

[Corinna Vinschen]

On Jul 10 11:24, Steffen Sledz wrote:
> On Tue, 10 Jul 2007, Dr. Volker Zell wrote:
> > I splitted out the following conflicting files from my xemacs package:
> 
> What's the right way to handle such conflicts?
> 
> > 
> > xemacs-tags:            (conflicts with ctags-5.6 and now with upcoming emacs-22.1)
> >   usr/bin/ctags.exe
> >   usr/bin/etags.exe
> >   usr/share/man/man1/ctags.1.gz
> >   usr/share/man/man1/etags.1.gz
> 
> Exclude from emacs and make ctags a requirement?

Exclude, yes.  A requirement, no.  ctags/etags are no requirement
to edit files.

> 
> > xemacs-emacs-common:    (used to conflict with emacs-21.2 now with upcoming emacs-22.1)
> >   /usr/bin/b2m.exe
> >   /usr/bin/rcs-checkin
> 
> No idea.

That's one between Volker and you.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Re: Emacs is coming back - please test

[Dr. Volker Zell]

>>>>> Corinna Vinschen writes:

    > On Jul 10 11:24, Steffen Sledz wrote:
    >> On Tue, 10 Jul 2007, Dr. Volker Zell wrote:
    >> > I splitted out the following conflicting files from my xemacs package:
    >> 
    >> What's the right way to handle such conflicts?
    >> 
    >> > 
    >> > xemacs-tags:            (conflicts with ctags-5.6 and now with upcoming emacs-22.1)
    >> >   usr/bin/ctags.exe
    >> >   usr/bin/etags.exe
    >> >   usr/share/man/man1/ctags.1.gz
    >> >   usr/share/man/man1/etags.1.gz
    >> 
    >> Exclude from emacs and make ctags a requirement?

    > Exclude, yes.  A requirement, no.  ctags/etags are no requirement
    > to edit files.

Maybe you can bundle them in an emacs-tags package.
    
    >> 
    >> > xemacs-emacs-common:    (used to conflict with emacs-21.2 now with upcoming emacs-22.1)
    >> >   /usr/bin/b2m.exe
    >> >   /usr/bin/rcs-checkin
    >> 
    >> No idea.

    > That's one between Volker and you.

An emacs-xemacs-common package ?
    
    > Corinna

Ciao
  Volker
  

Re: Emacs is coming back - please test

[Corinna Vinschen]

On Jul 10 18:21, Dr. Volker Zell wrote:
> >>>>> Corinna Vinschen writes:
>     > On Jul 10 11:24, Steffen Sledz wrote:
>     >> Exclude from emacs and make ctags a requirement?
> 
>     > Exclude, yes.  A requirement, no.  ctags/etags are no requirement
>     > to edit files.
> 
> Maybe you can bundle them in an emacs-tags package.

We already have a ctags package.  Having another package providing
ctags/etags is wrong and only helps confusing users and setup.  Since
they are provided by its own ctags package, they should just be dropped
in other packages.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Re: Emacs is coming back - please test

[Steffen Sledz]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 10 Jul 2007, Corinna Vinschen wrote:
> > > xemacs-emacs-common:    (used to conflict with emacs-21.2 now with upcoming emacs-22.1)
> > >   /usr/bin/b2m.exe
> > >   /usr/bin/rcs-checkin
> > 
> > No idea.
> 
> That's one between Volker and you.

It looks like Volker had made this xemacs-emacs-common package just for 
this purpose, because it only contains these two files. So I can remove 
them from emacs.  :)

Steffen

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: For info see http://quantumlab.net/pine_privacy_guard/

iD8DBQFGk8hLlSfCk8EgCzgRAtu8AJ9xgHVYSML3qiZ8TpVPJ/KDywvL3gCfaT4L
1G+pTj+ytAgc4Mt5IeGiq74=
=4JCt
-----END PGP SIGNATURE-----

Re: Emacs is coming back - please test

[Reini Urban]

Steffen Sledz schrieb:
> It looks like Volker had made this xemacs-emacs-common package just for 
> this purpose, because it only contains these two files. So I can remove 
> them from emacs.  :)

And put it into requires: of your setup.hint.

Re: Emacs is coming back - please test

[idirectscm]

On Jul 10 18:44, Corinna Vinschen wrote:

>> On Jul 10 18:21, Dr. Volker Zell wrote:
>> >>>>> Corinna Vinschen writes:
>>     > On Jul 10 11:24, Steffen Sledz wrote:
>>     >> Exclude from emacs and make ctags a requirement?
>>
>>     > Exclude, yes.  A requirement, no.  ctags/etags are no requirement
>>     > to edit files.
>>
>> Maybe you can bundle them in an emacs-tags package.
>
> We already have a ctags package.  Having another package providing
> ctags/etags is wrong and only helps confusing users and setup.  Since
> they are provided by its own ctags package, they should just be dropped
> in other packages.

Fedora 7 handles this conflict as follows:

   - Exhuberant Ctags (aka, "ctags") provides the 'ctags' binary, which
     has an command-line switch, '-e', for generating Emacs Lisp tag files.

   - Emacs 22 provides the 'etags' binary, along with 'gctags' (GNU ctags)

Commentary:
My understanding is that etags "belongs" to emacs, i.e., it predates
exuberant ctags (after all, "etags" generates tags for the Emacs Lisp
programming language).  So, emacs's provision of etags supercedes
any other package's providing of a program with the same name.  If there
is a conflict between the tags generated by Emacs 'etags' and the tags
generated by the Exuberant Ctags 'ctags -e' or 'etags', then I would
expect that the Emacs etags-generated file would take precedence, i.e.,
it will likely be more in sync with whatever evolution there might be
in the Emacs Lisp programming language.

 From an end-user perspective, there are the following considerations:

    - Someone who installs emacs, but has not installed exuberant
      ctags should be able to expect that etags will be available.

    - Someone who wants to use 'ctags' should not be required to
      install emacs.  It should be possible to get this utility by
      installing only the exuberant ctags package.

Because of these two reasons, it seems to me that the Fedora 7 solution
is a reasonable one, and would be better than preventing emacs from
providing the etags utility program.

---
Mark Harig
i d i r e c t s c m AT a i m DAUGHT c o m (minus the spaces, and with 
punctuation for the "words")