* Re: master 1ca436a: Make compilation mode work with warnings from compiled buffer functions
[not found] ` <20181128132418.851A9204F1@vcs0.savannah.gnu.org>
@ 2018-11-28 16:37 ` Glenn Morris
2018-11-28 20:16 ` Alan Mackenzie
0 siblings, 1 reply; 4+ messages in thread
From: Glenn Morris @ 2018-11-28 16:37 UTC (permalink / raw)
To: emacs-devel; +Cc: Alan Mackenzie
Hi,
Alan Mackenzie wrote:
> --- a/lisp/emacs-lisp/bytecomp.el
> +++ b/lisp/emacs-lisp/bytecomp.el
> @@ -124,6 +124,7 @@
> (require 'backquote)
> (require 'macroexp)
> (require 'cconv)
> +(require 'compile)
I would expect this extra require to slow down bootstrap, for an issue
that is not relevant to bootstrap (bug#33475)? Can this be avoided?
Indeed, my automatic overnight build today was about 13 mins, whereas
every previous day this month was < 11.5 mins.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: master 1ca436a: Make compilation mode work with warnings from compiled buffer functions
2018-11-28 16:37 ` master 1ca436a: Make compilation mode work with warnings from compiled buffer functions Glenn Morris
@ 2018-11-28 20:16 ` Alan Mackenzie
2018-11-29 3:51 ` Glenn Morris
0 siblings, 1 reply; 4+ messages in thread
From: Alan Mackenzie @ 2018-11-28 20:16 UTC (permalink / raw)
To: Glenn Morris; +Cc: emacs-devel
Hello, Glenn.
On Wed, Nov 28, 2018 at 11:37:55 -0500, Glenn Morris wrote:
> Hi,
> Alan Mackenzie wrote:
> > --- a/lisp/emacs-lisp/bytecomp.el
> > +++ b/lisp/emacs-lisp/bytecomp.el
> > @@ -124,6 +124,7 @@
> > (require 'backquote)
> > (require 'macroexp)
> > (require 'cconv)
> > +(require 'compile)
> I would expect this extra require to slow down bootstrap, for an issue
> that is not relevant to bootstrap (bug#33475)?
I've looked at this, and I don't think there would be other than a
microscopic slowdown. The previous version of bytecomp.el used at least
one function in compile.el (namely compilation-mode), suggesting that
compile.elc? is already loaded when bytecomp starts. However, for some
fiddly reason a macro in compile.el was no longer defined, hence the
require.
> Can this be avoided?
Yes it can, by moving the declaration of emacs-lisp-compilation-mode
out of bytecomp.el, probably into compile.el itself.
> Indeed, my automatic overnight build today was about 13 mins, whereas
> every previous day this month was < 11.5 mins.
I've timed several builds, and seen no differences between before and
after, appart from normal measurement error. I see 2min 29sec +- 1
second for "cold starts", and 2min 3sec += 1 second for a subsequent
build, when all the source files are in filestore cache.
Are you sure that bug #33475 is the cause of your slowdown?
--
Alan Mackenzie (Nuremberg, Germany).
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: master 1ca436a: Make compilation mode work with warnings from compiled buffer functions
2018-11-28 20:16 ` Alan Mackenzie
@ 2018-11-29 3:51 ` Glenn Morris
2018-11-29 10:26 ` Alan Mackenzie
0 siblings, 1 reply; 4+ messages in thread
From: Glenn Morris @ 2018-11-29 3:51 UTC (permalink / raw)
To: Alan Mackenzie; +Cc: emacs-devel
Alan Mackenzie wrote:
> The previous version of bytecomp.el used at least one function in
> compile.el (namely compilation-mode), suggesting that compile.elc? is
> already loaded when bytecomp starts.
It only uses compilation-mode in interactive use, but 1ca436a33c loads
compile unconditionally.
> However, for some fiddly reason a macro in compile.el was no longer
> defined, hence the require.
Since AFAICS you only need the define-compilation-mode macro,
eval-when-compile should solve the slowdown.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: master 1ca436a: Make compilation mode work with warnings from compiled buffer functions
2018-11-29 3:51 ` Glenn Morris
@ 2018-11-29 10:26 ` Alan Mackenzie
0 siblings, 0 replies; 4+ messages in thread
From: Alan Mackenzie @ 2018-11-29 10:26 UTC (permalink / raw)
To: Glenn Morris; +Cc: emacs-devel
Hello, Glenn.
On Wed, Nov 28, 2018 at 22:51:02 -0500, Glenn Morris wrote:
> Alan Mackenzie wrote:
> > The previous version of bytecomp.el used at least one function in
> > compile.el (namely compilation-mode), suggesting that compile.elc? is
> > already loaded when bytecomp starts.
> It only uses compilation-mode in interactive use, but 1ca436a33c loads
> compile unconditionally.
> > However, for some fiddly reason a macro in compile.el was no longer
> > defined, hence the require.
> Since AFAICS you only need the define-compilation-mode macro,
> eval-when-compile should solve the slowdown.
Ah, I've got you now! Yes, that does indeed save a few seconds on the
bootstrap time. Thanks for making the change in the repository.
--
Alan Mackenzie (Nuremberg, Germany).
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2018-11-29 10:26 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <20181128132417.30869.1170@vcs0.savannah.gnu.org>
[not found] ` <20181128132418.851A9204F1@vcs0.savannah.gnu.org>
2018-11-28 16:37 ` master 1ca436a: Make compilation mode work with warnings from compiled buffer functions Glenn Morris
2018-11-28 20:16 ` Alan Mackenzie
2018-11-29 3:51 ` Glenn Morris
2018-11-29 10:26 ` Alan Mackenzie
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).