Re: Proposition: 'NMU' upload of wxwidgets 2.8
On Tue, 12 Feb 2008 19:07:10 -0600 William Pitcock <firstname.lastname@example.org> wrote:
WP> On Wed, 2008-02-13 at 00:55 +0100, Vadim Zeitlin wrote:
WP> > And, seeing from your signature that you're both a Debian and Ubuntu
WP> > developer, I'd like to notice that Ubuntu doesn't seem to find
WP> > anything
WP> > catastrophic with shipping wx2.8 which it does since quite some time.
WP> So Ubuntu ships wx2.8. It doesn't matter to us. The fact that Ubuntu
WP> does ______________ is not generally considered a valid argument for
WP> justifying that Debian does the same.
Yes, I know this and it was never my intention to imply that this alone
was sufficient. But it did, and still does, sound strange to me that Ubuntu
people didn't find any of those apparently numerous and unavoidable fatal
bugs which wx2.8 is so full of.
WP> What you should be telling us is why we should be shipping wx2.8 over
WP> wx2.6 which is considered by many to be more proven than wx2.8.
Could we please bring any facts in this discussion? I replied to a message
stating, without any supporting arguments, that wx2.8 was unsuitable to be
used. You make a less strong but still fairly significant claim that wx2.6
is considered by many to be better than wx2.8. Could you please tell who
those "many" are and why do "they" consider this?
It's very difficult to prove that you're innocent when you don't know what
do you stand accused of.
WP> I am sure if you can come up with valid reasons to do so (especially
WP> identifying critical apps which require wx2.8 features is useful here),
WP> that we will be happy to provide wx2.8.
I don't know how critical these apps are but several of them have been
mentioned previously by different people. In particular, if you appreciate
using Debian as a development platform, the fact that CodeBlocks can't be
built on it is IMHO a pretty critical problem.
And even if a program doesn't require wx2.8 it will still work better with
it than with wx2.6. Moreover, wx2.6 is officially unmaintained since 1.5
years (and in practice for longer) and any future bug fixes will be done
only in wx2.8.
But more generally I thought that it was in the order of things for old
versions of programs and libraries to be replaced with newer ones in newer
Debian releases. I didn't realize there was a need to provide a special
argument for the upgrade, I rather thought that the problem was that wx2.8
was (erroneously and, AFAICS, due to the efforts of one and single person)
deemed to be too broken to be upgraded to in spite of numerous requests
here and in the BTS to do it. If this is not the case I don't think I can
provide an argument more compelling than the ones already expressed before.
So, once again, I can only propose to help to bring wx2.8 to Debian. If
this is deemed to be unnecessary -- so be it. I'd just appreciate if the
decision not to include wx2.8 in Debian could be formulated as being due to
lack of reasons to upgrade and not as being due to "wx2.8 being totally
unsuitable for application development" which is completely slanderous but
unfortunately carries a lot of weight when it comes from the official wx