httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <>
Subject Re: --enable-dav
Date Thu, 06 Jul 2000 01:13:49 GMT
On Wed, Jul 05, 2000 at 02:20:53PM -0700, wrote:
> > | Our binary builds only do the default modules, and mod_so.  It should be 
> > | very easy to turn DAV on with apxs, which does work in 2.0.  The
> > | instructions for turning it on should be:
> > 
> >   If we are building Apache with DSO support, I think we should build  
> > everything by default, but only enable the 'standard' stuff.  This business  
> > of tying what we build to what we enable is lame.  Build it all, so if the  
> > user wants to use it, all they have to do is turn it on.  They shouldn't  
> > have to download the source and build the parts they were missing.
> As long as we are on this topic, has anybody looked at the size of 2.0
> tree recently?  I am looking at a tree I just checked out of CVS.  The
> tree is 12M.  Compare that to the 1.3 tree which is 8.7M.  Notice, that's
> without doing any compiles at all and ignoring the CVS directories!  I'm
> not saying we need to stop putting stuff in the tree, just thought this
> was interesting to look at.  I'm currently trying to clean some stuff, and
> merge some code together so that this gets a bit smaller, but I have no
> idea how successful I'll be.  :-)

Dude. I hear you. When I cvs add'd the DAV code and it arriving in the
checkin email as 476k, I totally bugged. I had *no* idea there was that much
source. I knew the size of the tarballs (about 160k), but never really
considered what the unpacked size would be.

Well... trimming back is goodness. But I'm also a fan of "batteries
included" and think that much of Apache's market share is due to the fact
that people can unpack, build, install and have what they need. With the
additional packaging that people are doing to drop in PHP and mod_perl, the
"out of the box utility" is growing. But: I wouldn't want to rely solely on
the packagers for that utility.

[ we're facing a similar dilemma in the Python core development and its
  distribution: ship with "all the right stuff" or defer to packagers? seems
  people generally want the standard distro to have it ]


Greg Stein,

View raw message