[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#9074: 24.0.50; problems trying to mail bug report, again

From: Lawrence Mitchell
Subject: bug#9074: 24.0.50; problems trying to mail bug report, again
Date: Thu, 14 Jul 2011 09:29:21 +0100
User-agent: Gnus/5.110018 (No Gnus v0.18) Emacs/24.0.50 (gnu/linux)

Drew Adams wrote:


>> This is a way to get you to show how
>> to reproduce the problem. Send the actual report from your normal,
>> configured Emacs.

> No.  For many people, including me, their normal, configured
> Emacs is far too complex to be the basis of most bug reports.
> We do not want to pick all of a user's extra customizations, if
> we can help it.

The customisations in your normal emacs have /nothing to do/ with
the bug report.  The bug report has a /self-contained/ recipe
starting from emacs -Q, but for all one cares it might have been
written on parchment and sent by pigeon.

I don't know how you produce bug reports but for me the process
is something like:

* notice bug in my (customised) emacs
* emacs -Q
* try and reproduce bug
* if successful go back to the normal emacs, hit report-emacs-bug
* type the commands I used from emacs -Q to trigger the bug
* send

At no point do I try and send email from an emacs -Q instance.

> I don't know where this mailer stuff is saved, but if it is
> saved in the user's `custom-file' or init file then in order to
> start from emacs -Q and get to a reasonable bug-report state
> s?he needs to answer this mail nonsense EACH time.

Yes, that's what -Q does, ignore /all/ user settings.  How do you
propose ignoring all but one user setting in a way that doesn't
lead to more pain.

> This is (should be considered) totally unacceptable.  It works
> against what _we_ want in terms of bug reporting.  Not to
> mention that it is simply not nice to users.

> At the very least we should be able to get rid of the
> backwardness: making users who will ultimately hit `n' for SMTP
> configuration jump through lots of irrelevant obstacles.  Ask
> the SMTP question first, if you absolutely must ask it each
> time, and then DTRT for a `n' response.

This I agree with.  If the option exists to send mail through
emacs or not it should go something like:

M-x report-emacs-bug RET

Do you want to configure SMTP in Emacs?

Write bug report.

Hit send.

If mailclient-send-it is being used, you get to fill in the
address (maybe) in your normal mail client.  If you're passing
off to a local sendmail (sendmail-send-it) you'll need to specify
a valid address at this point.

So effectively I feel there are two bug reports here:

1) Change order of questioning for one-time SMTP configuration to
avoid pointless questions for people who don't want to send mail
in Emacs.

2) Add a method of saving a user customisation that is picked up
when running emacs -Q.

The former seems like a sensible thing to do.  I (FWIW) think the
latter is a terrible idea.

Lawrence Mitchell <address@hidden>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]