Re: Changes committed gnus/contrib (ChangeLog compface.el)

Re: Changes committed gnus/contrib (ChangeLog compface.el)

[Romain Francoise]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> Move the version of ELisp-based uncompface program from the lisp directory
> because of the copyright problem.

This is problematic since (unless I missed something) files in contrib/
are not automatically byte-compiled.  I upgraded my No Gnus copy and was
surprised to see that articles with X-Face headers were rendered _very_
slowly[1] on my not so slow system...  Byte-compiling compface.el fixed
the problem, rendering is instantaneous again.

Would it be a good idea to byte-compile files in contrib as well?

-- 
Romain Francoise <romain@orebokech.com> | You know that old saying,
it's a miracle -- http://orebokech.com/ | that you always hurt the ones
                                        | you love? Well it works both
                                        | ways.

Footnotes:
[1]  ELP results:

    Function Name           Call Count  Elapsed Time  Average Time
    ======================  ==========  ============  ============
    uncompface              1           1.495175      1.495175
    uncompface-internal     1           1.495155      1.495155
    uncompface-uncompress   241         0.1217170000  0.0005050497
    uncompface-big-mul-add  408         0.031415      7.699...e-05
    uncompface-big-pop      319         0.0303899999  9.526...e-05
    uncompface-pop-grays    83          0.0130459999  0.0001571807

Re: Changes committed gnus/contrib (ChangeLog compface.el)

[Katsumi Yamaoka]

>>>>> In <87pt38um9p.fsf@orebokech.com> Romain Francoise wrote:

> Katsumi Yamaoka <yamaoka@jpl.org> writes:

>> Move the version of ELisp-based uncompface program from the lisp directory
>> because of the copyright problem.

I'm sorry to have not checked the author had assigned the right.
He is still missing.  I want him to have not met misfortune.

> This is problematic since (unless I missed something) files in contrib/
> are not automatically byte-compiled.  I upgraded my No Gnus copy and was
> surprised to see that articles with X-Face headers were rendered _very_
> slowly[1] on my not so slow system...  Byte-compiling compface.el fixed
> the problem, rendering is instantaneous again.

Is there any reason you don't want to use the version of compface.el
which still exists in the lisp directory?  The reason you need to use
the ELisp-based uncompface program may be the following: there's no
uncompface binary; you worry the uncompface binary may be invaded;
you're interested in the program itself.  Otherwise, lisp/compface.el
can be used.

> Would it be a good idea to byte-compile files in contrib as well?

You don't happen to include the contrib directory in load-path, do
you?  I don't think that's a good idea.  Instead, you can move
required modules from the contrib directory to the lisp directory
before performing `configure; make install'.

The best way will be to rewrite the ELisp-based uncompface program by
someone who can assign or has been assigned the right.

Re: Changes committed gnus/contrib (ChangeLog compface.el)

[Romain Francoise]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> Is there any reason you don't want to use the version of compface.el
> which still exists in the lisp directory?

Yes, I don't have the necessary tools installed, and I don't want to
install them: I find the Elisp-based solution much more elegant and it
reduces the number of external tools Emacs depends on in my setup.
Besides, my system is fast enough to do this in Elisp.

> You don't happen to include the contrib directory in load-path, do
> you?  I don't think that's a good idea.  Instead, you can move
> required modules from the contrib directory to the lisp directory
> before performing `configure; make install'.

In fact I do, it makes it easier to track changes in those files: if I
move them elsewhere I'd have to check that my version and the version in
contrib are the same each time I update Gnus.  And I don't do `make
install' at all, I just add the lisp directory to my load-path.

I will modify the Makefile to byte-compile files in contrib in my local
version, then.

> The best way will be to rewrite the ELisp-based uncompface program by
> someone who can assign or has been assigned the right.

That sounds like a waste of time.  :-(  Let's hope the original author
will reply soon.

Thanks for your reply,

-- 
Romain Francoise <romain@orebokech.com> | I know, you love the song but
it's a miracle -- http://orebokech.com/ | not the singer.

Re: Changes committed gnus/contrib (ChangeLog compface.el)

[Katsumi Yamaoka]

>>>>> In <873c03e2vs.fsf@orebokech.com> Romain Francoise wrote:

> Katsumi Yamaoka <yamaoka@jpl.org> writes:

>> Is there any reason you don't want to use the version of compface.el
>> which still exists in the lisp directory?

> Yes, I don't have the necessary tools installed, and I don't want to
> install them: I find the Elisp-based solution much more elegant and it
> reduces the number of external tools Emacs depends on in my setup.
> Besides, my system is fast enough to do this in Elisp.

Indeed.  I think that is good enough too.

>> You don't happen to include the contrib directory in load-path, do
>> you?  I don't think that's a good idea.  Instead, you can move
>> required modules from the contrib directory to the lisp directory
>> before performing `configure; make install'.

> In fact I do, it makes it easier to track changes in those files: if I
> move them elsewhere I'd have to check that my version and the version in
> contrib are the same each time I update Gnus.  And I don't do `make
> install' at all, I just add the lisp directory to my load-path.

(I use it, furthermore, I always make bootstrap Emacs because I
think I need to check whether they work. ;-))

> I will modify the Makefile to byte-compile files in contrib in my local
> version, then.

You might need to do it personally all the time.  Although I'm
not opposed to making it perform to Makefile by default since it
won't take a long time to compile all the contrib modules, I'm
not sure it doesn't bring about something bad.

>> The best way will be to rewrite the ELisp-based uncompface program by
>> someone who can assign or has been assigned the right.

> That sounds like a waste of time.  :-(  Let's hope the original author
> will reply soon.

For the moment, what I can say is only a joke.  Probably he will
be in the way of a solo voyage to cross the Pacific Ocean. :-(

Re: Changes committed gnus/contrib (ChangeLog compface.el)

[Lars Magne Ingebrigtsen]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> The best way will be to rewrite the ELisp-based uncompface program by
> someone who can assign or has been assigned the right.

Did TAKAI Kousuke ever reappear?  It would be kinda nice to be able to
use the elisp-based uncompface...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen