Re: Beagle hackfest at GNOME Summit
- From: Dan Winship <danw novell com>
- To: desktop-devel-list gnome org
- Subject: Re: Beagle hackfest at GNOME Summit
- Date: Thu, 12 Oct 2006 08:48:53 -0400
Joe Shaw wrote:
> Here's a quick rundown of the things people decided to work on:
> * Dan Winship - As a broader goal desktop goal, decided to look
> into moving Beagle to Cmake and learn its strengths and
> weaknesses.
So, mostly I did this because I was just hanging out in that room, and
didn't actually have any beagle hacking I was planning to do, and I
didn't want Joe to mock me like he did Robert:
> * Robert Love - Hanger on; spell checker; incorrectly
> attributing typos to me.
But also because Joe and Alex Larsson and I had somehow ended up talking
about CMake earlier, although I don't remember why. (Background:
http://cmake.org, http://lwn.net/Articles/188693/ ("Why the KDE project
switched to CMake"))
> It sounds like there are a lot of weaknesses, at
> least within using it in GNOME.
Well, I didn't actually get to look that deeply into it. It has a
definitely Windows-ish feel to it which I found kind of icky, but then,
I find autotools kind of icky too, so that's not that much of a change.
Also, I'd been hoping that KDE's automake-to-cmake conversion script
would be useful, but it turned out not to be (I guess it only works if
your Makefile.am's use certain KDE-ish conventions?). Many of the
existing CMake macros wouldn't work for us as-is; eg, there's a
pkg-config macro, but it doesn't let you check the package version, so
we'd have to write a new one. Also, the fact that you need to have cmake
installed in order to build the package was a point against it relative
to autotools. (cmake generates Makefiles [on UNIX], but you can't just
run cmake and then ship the Makefiles, because the cmake invocation is
the equivalent of running configure in an autotools-based setup.)
So anyway, it turns out that I couldn't migrate a large multi-language
project from autotools to cmake in a 2-hour hacking session (go figure),
and cmake isn't 100% perfect, but maybe it still could turn out to be a
good solution for GNOME. Mostly I'm just posting this so that people
don't rule it out because they think I actually did a detailed study.
(And so that OSNews doesn't post a story entitled "New KDE Build System
Not Good Enough For Us, Say Arrogant GNOME Pricks".)
-- Dan
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]