Re: UI Freeze Breakage Request: Nautilus

[Christian Neumair <chris gnome-de org>]

Am Dienstag, den 22.02.2005, 10:41 -0700 schrieb Elijah Newren:
> On Tue, 22 Feb 2005 16:45:26 +0100, Christian Neumair
> <chris gnome-de org> wrote:
> 
> > How do you mean that? I wrote patch [1] and [2] myself, where [2] is a
> > heavily improved version of [1]. Both are been tested extensively by me.
> 
> We naturally expect that you've heavily tested your own patches.  That
> should always be done at any point in the release cycle.  I believe
> Murray was asking more about how widely others have tested.  (Although
> I heavily test my patches before committing, I frequently find that
> someone else has a different use case that didn't occur to me and that
> causes problems)

I heavily rely on existing and extremely tested code (the nautilus drag
data interface interface, that is). 

> > By mentioning [3], I just wanted to point out that a little bit of the
> > code (a flag) will be changed in the final version.
> > Maybe the term "to be applied" wasn't clear enough. I should have said
> > "to be checked in".
> > 
> > I'd also like to point out that I was asked by Alex Larsson to handle
> > this process [3].
> 
> "It also needs GDK_BUTTON2_MASK as martin pointed out. And I agree with
> martin that we need to figure out some of the corner case behaviour.
> However, it might be nice to get this into 2.10 as is, so we can at
> least work in most cases. We can figure out the rest later.
> 
> Do you mind handling the release team freeze break request?"
> 
> Mention of corner cases sounds scary.  Especially when we're trying to
> enforce stability and we're so close to hard code freeze (it may be
> that the patch only makes things better while still not fixing
> everything, but that's not totally clear and even if that's true it's
> still a little scary).  While this patch does look /really nice/, I'd
> rather not risk problems from this which could detract from other
> showstopper bugs that may be reported or that already exist.  I'm
> therefore leaning against approving it, but perhaps two others on the
> team will approve it.

You don't seem to have read the wole discussion. I can't clarify each
quote/cross-reference, which would be very exhausting.
By corner cases he was referring to another post by Martin Wehner who
pointed out that in some cases the association between activation URI
and activated URI gets lost [1]. But this problem has absolutely nothing
to do with the changes my patch.

Again, Nautilus people already accepted the patch and as you all know
Alex is experienced enough to know what he is doing.

Why is this process that bureaucratic and exhausting for a simple dumb
hacker like me? :/

[1]
http://mail.gnome.org/archives/nautilus-list/2005-February/msg00048.html

-- 
Christian Neumair <chris gnome-de org>