Subject: Re: Alt-Insert
From: Sam TH (sam@uchicago.edu)
Date: Tue May 01 2001 - 16:05:46 CDT
On Tue, May 01, 2001 at 09:53:25PM +0100, Rui M Silva wrote:
> On Tue, May 01, 2001 at 03:17:19PM -0500, Sam TH wrote:
> > Well, theoretically the GNOME 2.0 libraries are frozen in July, and
> > they need to be ported to GTK 2 by then.
> > More importantly for us, we need to port AbiWord to GTK/GNOME 2 by the
> > end of the year, in order to be part of GNOME 2.0
> > This is another reason to do 1.0 soon, so we can break the tree
> > horribly for that.
>
> That is no reason at all.
> If so, please explain ABI_OPT_LIBXML2, ABI_OPT_PSPELL, ABI_OPT_*
ABI_OPT_LIBXML2 has 5 #ifdefs.
ABI_OPT_JS has 33 ifdefs.
ABI_OPT_GNOME has 5 ifdefs.
ABI_OPT_GTK2 would have hundreds. That is simply not ok.
>
> Do not argue without arguments. Here are mine: 1) any distro not
> running gtk compatibility libs for at least an year is too dumb to
> even be considered (just look at the transition from gtk+ 1.0 to
> 1.2, there was a gtk10-libs). This way, there will surely be a gtk12
> for more than long enough time for us to adapt to the new toolkit
> version using something like ABI_OPT_GTK2 and ABI_OPT_GNOME2
Right, the libs will exist. That's why Abi version 1.0.x will
continue to function perfectly on these systems. However, that's not
a good reason to hack up the code horribly, or to keep using them when
better libraries are available.
>
> 2) getting a 1.0 (specially with such a huge change in the horizon)
> because of a major upgrade in libs is like saying: in the other
> point of the world it's raining, so I better take my umbrella with
> me. It has nothing whatsoever to do with it.
That's simply not true.
Migrating to GNOME/GTK 2.0 will be a lot of work.
It will involve significant tree instability.
Therefore, it should come at a time when there is already going to be
tree instability, and not when we are trying to make a release.
>
> 3) to be in gnome 2.0 we better have gtk2 and gnome2 support ready
> by that time. Well... considering gtk+ api is considered mostly
> stable, we can start doing that already. bothering ourselves for a
> 1.0 iron curtain will only make us loose time, because we are
> getting a god damned 1.0 out. I think this is as dumb as it can
> be.
Having GTK 2 and GTK 1 code in the tree at once is just a bad
idea. Such abuse of the preprocessor will not be tolerated. I hate
the preprocessor anyway.
>
> Oh... when will you ever learn [ Nick Cave, 1st track of No More
> Shall We Part ] to an end user it doesn't matter whether AbiWord is
> 1.0 or 0.20, its rather more worrysome for them what I believe I
> have no need to repeat myself with, and it's much more of a loss of
> time of development.
You are seriously delusional if you don't think the version number
matters to users. As evidence, from the LinuxWeeklyNews report on
AbiWord 0.7.14:
<quote src="http://lwn.net/2001/0426/desktop.php3">
The AbiWord team released a new version of the AbiWord word
processor. It's not a 1.0-release but it's getting pretty solid.
</quote>
Users care. Users want a stable, bug-free AbiWord. Users don't want
to wait forever.
>
> Call it AbiWord 2001 (masking the not 1.0 if you *REALLY* think that
> end users do care about that -- except when they think it's really a
> final product from 1.0 and it doesn't have the features most would
> like) or what ever, but please, do not send a most surely unthought
> argument like that.
How much software released as 1.0 has all the features everyone wants?
Or even all the big features everyone wants?
That's what I thought.
AbiWord needs a period to stabilize, and release a totally solid 1.0.
We need that much more than we need any particular feature in 1.0.
sam th --- sam@uchicago.edu --- http://www.abisource.com/~sam/
OpenPGP Key: CABD33FC --- http://samth.dyndns.org/key
DeCSS: http://samth.dynds.org/decss
This archive was generated by hypermail 2b25 : Sat May 26 2001 - 03:50:59 CDT