Current situation of amarok, and of latex tools

[Vincenzo Ciancia]

Il giorno mer, 13/05/2009 alle 15.22 -0400, Daniel Chen ha scritto:
> 
> 
> You replied:
> > Actually, Daniel, it is rather poor.  You saw what happened on my
> computer.
> > If I didn't turn on greedy, it would lock constantly.  And seeing as
> it worked
> > without greedy not a month before release, that was a very late
> regression.
> 
> But is your hardware indicative of the common case?

Daniel: you don't know me and you are not supposed to. I sometimes point
out facts on this list, that we all (including me) commonly tend to
underestimate because of our belief that ubuntu is Good. It is. No
doubt.  However, I DO file bug reports in time and DO reply to
developers. I consider that a duty. I can't become an ubuntu developer
because I lack time for that. You have requirements about developers and
that's very good. I can't meet those requirements therefore I limit
myself to a literate tester.

I came on this list during alpha to signal that intel's driver was
extremely slow by default and this IS the common case on integrated
intel cards sold on a huge amount of laptops. I was responded ON this
list that it was known, and there was time.

One of the most active persons in the xorg/ubuntu community, Bryce
Harrington, has this bug and he is actively working with the community
to solve the problem. But "there was time" do you remember? A forward
port of the old driver should have been done, instead of saying that
there was time. Then, in time for release, seeing that the grave bug was
not solved, ubuntu could have shipped the old driver by default and the
new driver for testers. In the future I would be happy to see more
attention not to break things in such a heavy way.

Latex stack is broken. The common use case WAS kdvi+kile for all the
"gui-friendly texers" that I know. 

Regarding cooperation with upstream, I wasted hours in reporting all the
debug information on the intel BTS for the infamous "VGA-OUT broken"
bug. I just don't use my laptop for presentations. I wasted hours, but
when the intel developers had enough information, they stopped taking
care of the bug. Here it is. 

ubuntu

https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/137234

intel

https://bugs.freedesktop.org/show_bug.cgi?id=16702

Daniel: sounds like a rant, behaves like a rant, then it is a rant?? No.
I think that if I waste my time testing and reporting (and I may not be
the better, but am not the only one) prompt reaction to regression is a
payback. If I don't see it, I stop testing. Because my free time,
exactly as yours, is not for free. I don't ask for stopping the
progress. Just keep the working stuff enough time.

When kdvi was added back to jaunty, it was in response to a request
here. There was a request open on the BTS but someone (me) needed to
come here and point out reasons. Then when kdvi was removed from the
archive, nobody needed to ask me or the bug reporter or other users, if
okular was working. There was no time for that, the release was
pressing. It did not cost ANY ***** THING to keep kdvi there. It costs
ubuntu its image as a good texing platform, not having a good latex
stack.

Why not just taking care of regression by forward port by default? Not
for everything of course, but would it really cost so much? Users can
already tag regressions. Just let's introduce a procedure that avoids
regressions not to be cared of. I mean: it is disgusting: we have the
source code, it works, but we ship the broken version. Why on earth?
Let's ship both as it is already done for so many packages and
libraries. Let us do that in response to regression tags, after
verifying the use case, of course.

Vincenzo