Subject: Responsability (Was: autoconf at last)
From: Thomas Fletcher (thomasf@qnx.com)
Date: Wed Aug 02 2000 - 07:34:31 CDT
On Wed, 2 Aug 2000, Goran Thyni wrote:
> On Tue, Aug 01, 2000 at 05:43:50PM -0700, Aaron Lehmann wrote:
> > I would vote for this going into the tree some once it is a bit more
> > complete (optionally building a gnome version when using
> > --enable-gnome with configure, more widely tested, etc). People have
> > said to come up with code if they wanted autoconf and Goran has made
> > a great start with that. I hope gettext is next! :)
>
> It will not be tested unless it is in CVS.
> My suggestion is:
> add a README.autoconf and say that
> "it is developers only for now"
> and commit it as is unless someone
> think it has serious design flaws.
>
> I will test it on Solaris,
> hope other will test it on FreeBSD,
> CygWin etc.
>
> plans:
> * --with-gnome or --enable-gnome
> * --with-gnome-xml or --enable-gnome-xml
> * try out gettext support
Now I hope that the autoconf that you have built is
flexible enough to be used for other platforms outside
of the unix tree. I know that you could use autoconf
for BeOS builds (though the current tree works very
well in that domain) and the same thing for QNX/Photon
builds. I don't know enough about the windows cygnus
environment, but presumably it could be used there as
well. Now of course the big question is, can you automate
the maintenance of the Makefiles so that we don't have
to maintain two seperate sets of Makefiles .. the
Abi ones and the autoconf ones. I already think that
it is a pain to have to go and add files to my Makefiles
when we add a dialog.
I want things to improve in the build environment so
that life gets easier (for _all_ platforms) not harder.
If you think that I'm being ornery about this, all you
need to do is take a look at the last week to see how
many times the build was broken for one reason or another.
I know that this is an open source project and I am
just one voice of many, but people need to be VERY AWARE
of the ramifications of their changes. If you aren't
sure, then ask the list for advice. I'm quite certain
that we have developer representatives from all platforms
here. It is fine and dandy to submit a patch, but as I've
said before, unless specifically asked (such as by Martin
when he was doing early list development) I'm not going to
patch my code. If you aren't sure that all platforms
have: usleep, wcrtomb, some_other_cool_function ... then
ask and the list will come to a solution of how to best
implement it (usually a cover we can then put into platform
dependant code rather than #ifdefing ...).
If we were to really be living by the Tinderbox rules,
no one of the platforms would be able to have _any_
new features put in until we were green across the
board. This is the reason we have Tinderbox and the
reason we have a mailing list, and people who have
write access should take this commitment to keeping the
build clean very seriously (this isn't to accuse anyone
of not taking it seriously ... just a reminder to us all).
Incidentally, is there a list of who has write cvs access?
In reality there should be at least one major contributor
per platform who can do some moderate QA on patches before
submitting them.
I don't want to come off sounding holier than thou ...
the only reason you don't see the mistakes on the QNX/Photon
platform is that I don't have a Tinderbox entry (yet) ;^)
and 90% of my contributions have been in working solely on
platform dependant code.
In any case don't want to crimp anybodies enthusiasm, just
wanted a bit of a reality check in how we are working.
Code on!
Thomas
-------------------------------------------------------------
Thomas (toe-mah) Fletcher QNX Software Systems
thomasf@qnx.com Neutrino Development Group
(613)-591-0931 http://www.qnx.com/~thomasf
This archive was generated by hypermail 2b25 : Wed Aug 02 2000 - 07:33:55 CDT