[gccsdk] Firefox 2 patches

Theo Markettos theo at markettos.org.uk
Wed Mar 21 05:32:48 PDT 2007


On Tue, Mar 20, 2007 at 09:22:33PM -0700, Peter Naulls wrote:
> Let's hop down from druck's high horse for just a moment.  No doubt he
> thinks he's doing everyone a huge favour by "forcing" me to comply with
> it, but this is again the short-term thinking I'm so opposed to, and
> very much the attitude he predicted some years ago could and did drive
> developers away.  If push comes to shove, then I will supply a tarball
> with everything and no further help, and will have washed my hands of
> the whole affiar.  But instead, if we're actually still interested in
> collaborative development, as I'm certain you are, then perhaps we can
> give me the benefit of the doubt and go about this in a more productive
> manner that helps everyone.

OK... I think part of the problem is not having a reliable idea of how it
works.  Is it, for example, 100% original FF sources with all the extra
changes being in ChoX11 and build system?  In which case the upstream
sources are all that's required.  How does your FF2 build system work: is it
a Firefox source tree that you hack as appropriate, or does it grab a source
tree and patch it?  Could you, for example, release a tarball alongside the
original tarball you downloaded, so that other people can diff it?  That
wouldn't be too much effort on your part and, whilst it wouldn't be ideal
long term would at least give people an idea of what's involved.  What are
the differences as far as building goes between FF1.5 and FF2?

Possibly I think what a few people like druck are getting at is that with
copyleft software you relinquish quite a lot of control as to what people do
with your code.  You can't stop it being used to launch nuclear missiles,
and you can't stop forks.  You can try by educating people, but if they want
to be fools you can't stop them.  The flipside of this is that if they want
to help out they can just get on and do it, without taking up any of your
time.

What you can do instead is have a well-managed project which doesn't accept
patches from the fools.  Looking at open-source projects out there it seems
that those that have enough momentum tend to carry the community with them,
and forks only happen for projects that are dead or almost dead.  If there's
a disagreement, just state the policy and its reasons and let people decide
for themselves.

> > I think it would already a good start and learning step to build FF1.5.
> > The build system and all the necessary patches for libraries are there.
> > The only thing which stands in a smooth build from start to end is that we
> > don't have enough people to regulary build everything and keep up with
> > upstream changes.
> 
> John is absolutely right.  Unless and until you can build 1.5 - which
> after all, John and I went to extrodionary lengths to make easy, there
> is little point in trying to do anything at all with 2.  Having said
> that, 2 is now old news, and if RISC OS Firefox development is at all
> serious, the focus must be on 3.  That is a much harder problem, due to
> considerable changes.

Thanks, that's a useful starting point.  It's the not knowing quite what's
involved that makes it difficult to start getting involved - if the changes
between FF1.5 and FF2 are mostly those of upstream then I see it would be
little changed.

I've added some notes based on these postings to the Firefox wiki page.

> > For FF1.5 building everything is there.  I'm nearly sure that when you
> > have done this, you already have 95% of the work done for an FF2 build.
> > I wouldn't be surprised that an FF2 build (with X11 based interface) nearly
> > builds out-of-the-box with same or simalar patches made for FF1.5.
> 
> That's approximately correct, but again, I really need you to build 1.5
> first.

Thanks.  Am throwing some CPU at it now :)

Theo




More information about the gcc mailing list