gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Bodewig <bode...@apache.org>
Subject Re: Gump and APR
Date Fri, 26 Nov 2004 10:44:15 GMT
On Fri, 26 Nov 2004, <minfrin@sharp.fm> wrote:

> I'm keen to jump in to get httpd to build, both as source, and as an
> RPM (and other packing formats, as practical).

I know and I'm sorry for letting this sit for so long.

Right now I think we need two things:

* a way to allow Gump to invoke make

  This I'll try to work on.  Given my close-to-non-existant Python skills
  and that you declared you'd be at the same level a while back, there
  probably isn't much of handholding I could offer.  The blind leading
  the one-eyed or so.

* A way to tell the httpd where to look for APR

  This is where I could use your help.  I think we'll need some
  <configure> builder that we can tell where to look for APR's header
  files and the library itself.  Probably more a general way to
  specify command line parameters when running configure.

  Say I have APR compiled but not installed (in a system wide sense),
  how would I invoke configure for httpd - or apr-util, which is
  probably the next ring of the chain?  Would I need to set some env
  variables or pass parameters to make as well?

Something that you could do more or less immediately - and where I can
provide some guidance is starting to declare the modules and projects
we'll need.

If you look into
<http://cvs.apache.org/viewcvs.cgi/gump/repository/apr.xml>, here I
define a "repository" for APR, simply a directory to base common
source trees on in the Subversion sense.

In <http://cvs.apache.org/viewcvs.cgi/gump/project/apr.xml> I define a
<module> (nomenclature is CVS like for historic reasons) which in this
case only means that I'm talking about the apr/trunk directory inside
the apr "repository" defined above.

Inside of that module we have two "projects" which are atomic steps we
have to perform.  First we run buildconf and then configure.  If the
first one failed, we won't run the second.

Those files get hooked into Gump via the "profile"
<http://cvs.apache.org/viewcvs.cgi/gump/profile/gump.xml?r1=1.463&r2=1.464&makepatch=1&diff_format=u>

So we'd need an additional module for apr-util and one for httpd (and
a repository definition to go with it).

I'll be travelling by the start of next week which means I'll have
some hacking time in a train but will also be offline most of the
time.

Cheers

        Stefan

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org


Mime
View raw message