On 12/5/05, Jeremy Huntwork <[EMAIL PROTECTED]> wrote:
> Randy McMurchy wrote:
>
> > Because no matter how you execute the configuration stage, you end
> > up with the *same exact* thing. I would have thought you knew that.
>
> This is not intended to start an argument, but I really am just curious.
> If in fact both methods do produce the same results, what is mozilla's
> reason behind their suggested method and why was it ever considered a
> good thing to break from it in the first place?

I'm kind of speculating, but these are the major reasons, I think:
1.  They have a lot of custom autoconf type scripts to hash out the
myriad options and arches that it's built for.  This is kind of what I
pointed out: instead of passing a 30 line configure statement, you can
use a simple text file.  This method allows to also use the default
options a lot easier.  It's pretty straightforward to get stock
firefox with . $topsrcdir/browser/config/mozconfig.

2.  There's a lot of custom CVS pulling stuff in there to handle the
tons of branches and projects they have.  client.mk makes it sane to
maintain a single source tree and try to build multiple projects from
it.  Have you ever taken a look at their CVS tree?  It's insane:
http://lxr.mozilla.org/  And, yes, there are people who successfully
build multiple projects (suite, browser, etc.) from a single source
tree.

I've wondered this myself a few times, especially for #1 and for why
they don't go fully autotooled.  In the end, though, they decided to
make a customization of the CMMI system to support their huge source
tree, and now it's pretty mature.  It certainly doesn't take a wizard
to figure out how to use it (especially with the docs), and it's
backwards compatible to standard CMMI, anyway.

--
Dan
--
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to