Re: "If you're still seeing problems, please reopen." [Was: bug#25148:]

[Óscar Fuentes]

Alan Mackenzie <address@hidden> writes:

>> Maybe we should consider some of those which are not "significantly
>> better" at triaging and fixing bugs but are much better at interfacing
>> with end users and occasional contributors?
>
> What could be easier than M-x report-emacs-bug or just sending an email
> to address@hidden?

This topic was discussed several times already on this ml.

First of all, email is not configured or even available on all machines.
This forces the user to copy&paste text from Emacs to another
application (easy) or to another machine (not so easy). Attaching files
is not straightforward even when sending from Emacs itself.

After the initial report, if the user wishes to participate he must do
so by email, taking care of not altering To: and/or CC: headers. Also,
some users might have preferences about their level of partipation on
the ensuing discussion (be notified about every message, only messages
that mention him or only the message that resolves the bug). Actually,
the reporter risks being spammed by dozens of emails after a bug report
without an obvious method for stopping the flood.

> I run Gentoo, and for many days now a package won't build, and I want to
> report it as a bug.  But I just can't face the hassle of opening
> Firefox, looking up Gentoo's bugzilla address, logging on to it with a
> password I've got written down somewhere, filling in numerous fields on
> a screen, many of which are not relevant (but they're compulsory), then
> having to return to it in order to respond to responses.  It's just too
> much work.  I suppose I'll get around to it some time.
>
> That's not to say debbugs is perfect.  But for me, it beats bugzilla and
> other Web browser based trackers handsomely.

Emacs could act as an interface for some of those bug trackers,
automatically filling known fields. Instead of sending an email, it
would send an HTTPS request. It would be report-emacs-bug, but with the
advantage of being fully operational whenever the machine has access to
the WWW.

IIRC all this was discussed months ago and some of the proposed
candidates had an HTTP API that would make possible to implement a
front-end on Emacs.