2.6.0 tag is set on...

[Inge Wallin]

On Monday, February 04, 2013 18:29:45 C. Boemann wrote:
> On Monday 04 February 2013 15:33:05 Inge Wallin wrote:
> > On Monday, February 04, 2013 10:00:05 Cyrille Berger wrote:
> > > The conclusion on IRC was to reenable it for 2.6.1 (in two weeks).
> > 
> > That sucks badly for Calligra Author since the only thing that we can
> > actually say that we bring to the table is ebook export.
> > 
> > The filter was arbitrarily disabled due to a crash, then *not* reenabled
> > when the crash was fixed. The disabling was done by people who are not
> > involved in Author and I'm pretty sure the same is true for the IRC
> > discussion.
> > 
> > What I am - actually - a little bitter about is that the filter was just
> > disabled without any discussion with me the very day before the tagging.
> 
> There is nothing to be bitter about - it had to be disabled or fixed. Crash
> on save is the worst kind of data loss right preventing any kind of
> recovery. You were not around and it affects Words too, so I feel in my
> right to make such a decision.
> 
> I worked with moji until 3 am trying to fix it, whereas you went to bed.
> I'm sorry, but I think I went beyond the call of duty here, and I'm sorry
> I didn't wake up before Cyrille had started the tagging, so I could have
> backported and re-enable the filter. But at 3 am at night I didn't feel
> clear minded enough to check that no other problems were lingering. The
> worst thing I could have done was re-enable something that is a grave
> dataloss issue.

Don't get me wrong - I am very grateful for the help to fix the error. And 
yes, I felt bad going to bed then but I had a responsibility to my child the 
next day so I felt that I had to.

But there is something you are missing: If there was such an important bug and 
you didn't feel comfortable reenabling at that time we could instead have 
requested a move of the tagging one day forward. That would have given time to 
test the fix better and even re-fix it if more problems had been discovered.

This is what would have happened if you had discovered a similarly important 
bug in an area that you were responsible for. And in fact almost anywhere in 
Calligra, I'm willing to bet.

> This bug is so grave that it should be tested better. I've asked moji to
> make unit tests and Friedrich suggested you should throw our document
> collection at the calligra-converter. I've not seen result of those things
> yet, and with demonstrated crashy behavior I think it's a good thing to
> give it another 14 days.
> 
> So I would very much agree with Boudewijn that the alternative here is not
> retagging but rather another RC



> > And now the IRC discussion, also without me, came to the conclusion that
> > it's not so important and that it can just wait two weeks.
> 
> No such discussion took place - I contacted Cyrille trying to help by
> asking if it would be at least possible to add the feature in 2.6.1. This
> alone is already out of normal scope - we don't add new features in bugfix
> releases.
> 
> I'll try to not be so helpful the next time around :(

Then I apologize.  I got the information that there was an IRC discussion from 
Cyrille's very short mail. I couldn't really interpret what he wrote any other 
way.

> > If you ask me,
> > it can't wait two weeks because half of the reason for Author 2.6 has
> > just disappeared.
> > 
> > Can we please reconsider this?
> > 
> > 	-Inge
> 
> I'm against re tagging - but personally would have no problem with
> postponing the release 4 weeks for an RC and then the release.
> 
> Also regarding Author, I'm with Boudewijn to be honest. I don't really see
> what Author brings at this point. It doesn't bring anything that isn't in
> Words too and doesn't even have an icon yet. I'm sorry but I think we
> sometimes do things a bit too laissez faire. Having an idea and start on a
> new feature is nice and cool, but without follow-up to make it work it
> just reflects badly on all of us. And by doing Author I also expect you
> guys to do your share of library and Words bugfixing, and not just lofty
> new features.

Here I would like to point out that I deliberatly chose features to work on 
that were mutually beneficial to Words as well as Author. I could have worked 
on Author specific features already for 2.6 but I thought that Calligra at 
large would benefit from features that are more universally useful.

I do think that Author and Words in the long run will have less and less 
overlap in the userbase. The reason is the different focus we will have when 
choosing features to implement and also how they are implemented UI wise. I do 
also think that it will share an awful lot of technology, of course, which 
makes it reasonable that we do our share of the library work. And I do think 
we have done our share there.

But Author is not only a binary. It's also a promise for the future. By 
creating a new application and a announcing it, we have given a voice to a new 
category of users who previously weren't interested in Calligra at all. That 
can only be good. 

I agree that Words and Author is currently almost exactly the same if you just 
look at the UI and feature set. But the focus is very different and the 
promise for the future to those users is also very different. We want people 
to run Author early so that they become familiar with the Calligra way of 
doing things and start to influence the development in the future. That's what 
will be lost if we don't release Author even if it's very similar to Words 
right now and also (to a lesser degree, perhaps) if we start disabling 
features the last day before the tagging.

Now, regarding the situation at hand, I don't think that a new RC is the way 
to go. That would hurt the rest of Calligra more than it's worth. So if I have 
to choose between releasing yet another RC or wait 2 weeks to get the mobi 
filter I choose the former. So let's do that.

	-Inge

> _______________________________________________
> calligra-devel mailing list
> calligra-devel at kde.org
> https://mail.kde.org/mailman/listinfo/calligra-devel