On Mon, 2006-01-09 at 22:01 +0100, Georg Baum wrote:
> Am Montag, 9. Januar 2006 09:02 schrieb Lars Gullik Bjønnes:
> > Martin Vermeer <[EMAIL PROTECTED]> writes:
> > | Does this mean the patch can go in for 1.4.0? Or not? It fixes a
> > | regression.
> > 
> > Now I am a bit lost...
> > What regresssion?
> 
> The on-screen fonts of nested environments are correct in 1.3, but can be 
> wrong in 1.4. See the test case at 
> http://bugzilla.lyx.org/show_bug.cgi?id=2015
> 
> I don't understand at all why this bug creates such a huge discussion. 
> Martin found the reason for the bug, the problem is well understood, 
> there are several possibilities to fix it (exploit the ParagrphList == 
> std::vector<Paragraph> identity or not), now only one solution has to be 
> chosen. IMHO it does not matter at all if the fix makes LyX slower. 
> Correctness is more important than speed (of course speed can be a reason 
> to choose among the possible fixes that were presented).
> 
> > (I obviously have to fire up bugzilla.)
> 
> Yes.
> 
> 
> Georg


The problem seems to be that neither Lars nor Jean-Marc have the
personal resources to spare to keep track of everything that is going on
around patches and bug fixing.

This is not a problem of the freeze, though it tends to make things
worse as they become the bottleneck.

In any free software project, there has to be a group of "senior
developers" taking these critical decisions. For LyX that is Lars and
Jean-Marc. Requirements are fluency in C++ and the half-dozen "little
languages" (LaTeX, Python, bash, gettext, boost...), but more
importantly, ability to manage a largish project, and to see the big
picture. You don't get that if you haven't been along for a number of
years.

Unfortunately we lost Angus for this. Unfortunately also, senior
developers tend to be older people (as in: not students anymore) with
lots of other obligations and responsibilities. Being competent does
that to people ;-)

As we are all doing this in our spare time, we should use all the help
we can get, especially from automating things. We are using Bugzilla to
keep track of bugs (and of their fixes, though it isn't very good at
that). Often when somebody reports a bug, or a fix, he is told "put in
in Bugzilla, so it isn't forgotten". What happens of course is that
rather often it *is* forgotten there, and falls by the wayside when
after three re-posts and two reminders on the list, still nobody gives
the go-ahead, or even sensible critique. 

I am sure we lost promising contributors due to this "wet cloth"
experience. If it doesn't put me off, it is because I am a bonehead...

After this negativity, a proposal to improve things: 

1) Use the WHINE feature of Bugzilla. You can make it whine
periodically, e.g., once a week, to specified users. I propose we create
a user account for this, and make this account .forward to
[EMAIL PROTECTED]

2) The whine mail sent out once a week to the list should contain a
shortlist of unfixed bugs for the next release, displaying HTML links
and, most importantly, the KEYWORDS field.

3) Users placing their bugs and fixes in Bugzilla should remember to add
the word PATCH to the Keywords field, if a patch is available.

(In other words, more or less what Georg did in another post, but
automatically. And what Michael did earlier.)

I don't think that a single digest mail every week to the list is too
much, and it helps us keep track of things and focus the mind.

Does this sound like an idea?

- Martin

PS. Bugzilla's disk seems to be full.

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to