From: Tomi Ollila <tomi.ollila@iki.fi>
To: Austin Clements <amdragon@MIT.EDU>
Subject: Re: [PATCH 2/2] test: emacs: run list-processes after
	accept-process-output in emacs 23.1
In-Reply-To: <20120813151613.GN11179@mit.edu>
References: <1344165182-29246-1-git-send-email-tomi.ollila@iki.fi>
	<1344165182-29246-2-git-send-email-tomi.ollila@iki.fi>
	<20120813151613.GN11179@mit.edu>
User-Agent: Notmuch/0.13.2+128~g2eb637a (http://notmuchmail.org) Emacs/23.1.1
	(x86_64-redhat-linux-gnu)
Date: Thu, 16 Aug 2012 10:44:05 +0300
Message-ID: <m27gsz2se2.fsf@guru.guru-group.fi>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: notmuch@notmuchmail.org
Precedence: list

On Mon, Aug 13 2012, Austin Clements <amdragon@MIT.EDU> wrote:

> Quoth Tomi Ollila on Aug 05 at  2:13 pm:
>> When running emacs tests using emacs 23.1.1 the tests block (until timeout)
>> when emacs function (notmuch-test-wait) is called.
>> 
>> There is an emacs bug #2930 titled:
>> 23.0.92; `accept-process-output' and `sleep-for' do not run sentinel
>> 
>> It seems this is present in emacs 23.1.
>> 
>> Calling list-processes after accept-process-output seems work around
>> this problem; in case Emacs version is 23.1 a defadvice is activated
>> to do just that.
>
> Should this workaround perhaps go in notmuch-test-wait directly,
> instead of being implemented as advice?  

This way the fact that this is workaround for bug that exists only in
23.1(*) is emphasized; The notmuch-test-wait can exist in a (simpler)
format where it doesn't need to know about this bug.

(*) This bug probably exists in emacs 22 but the MUA has not worked
on emacs 22 at least for a year now.

> If we do want to keep it as advice, should it go in notmuch-lib.el
> along with the few other compatibility functions?

In tests we have fixed environment where we can write workarounds
as "global" advices. If we did this in notmuch-lib.el we'd be changing
users' environment, possibly causing surprises...

... but neither sleep-for or accept-process-output are used in 
any of the notmuch elisp code so such "fixes" are not needed now...

... and this would ever be a problem, we might already support only
emacs 24 or newer >;) ...

Tomi

>> ---
>> 
>> Thanks to Austin for the comments and IRC discussions on the matter.
>> 
>>  test/test-lib.el |    8 ++++++++
>>  1 files changed, 8 insertions(+), 0 deletions(-)
>> 
>> diff --git a/test/test-lib.el b/test/test-lib.el
>> index 52d9936..4330352 100644
>> --- a/test/test-lib.el
>> +++ b/test/test-lib.el
>> @@ -35,6 +35,16 @@
>>      "Disable yes-or-no-p before executing kill-emacs"
>>      (defun yes-or-no-p (prompt) t)))
>>  
>> +;; Emacs bug #2930:
>> +;;	23.0.92; `accept-process-output' and `sleep-for' do not run sentinels
>> +;; seems to be present in Emacs 23.1.
>> +;; Running `list-processes' after `accept-process-output' seems to work
>> +;; around this problem.
>> +(if (and (= emacs-major-version 23) (= emacs-minor-version 1))
>> +  (defadvice accept-process-output (after run-list-processes activate)
>> +    "run list-processes after executing accept-process-output"
>> +    (list-processes)))
>> +
>>  (defun notmuch-test-wait ()
>>    "Wait for process completion."
>>    (while (get-buffer-process (current-buffer))
> _______________________________________________
> notmuch mailing list
> notmuch@notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch