httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Hyde <>
Subject Re: [ANNOUNCE] Apache Module Contrib
Date Tue, 04 Aug 1998 20:17:37 GMT
Marc Slemko writes:
 > On Tue, 4 Aug 1998, Ben Hyde wrote:
 > > Marc Slemko writes:
 > >  > On Tue, 4 Aug 1998, Dean Gaudet wrote:
 > > ...
 > >  > > +1 for splitting non-free stuff into a non-free hierarchy (see
 > >  > > for example, they have a non-free hierarchy) 
 > >  > I do _not_ like this splitting out.  
 > > +0 agree with Marc.
 > And, of course, just to reinforce: this doesn't mean things can't be
 > labelled as free, non-free, idiot-of-an-author, etc.
 > > The FreeBSD system has slightly larger stubs on the end user's machine
 > The FreeBSD ports system is far easier in a lot of ways because it has to
 > work on one platform, and if something isn't there on that platform you
 > can just put it there.

This is the nub of what would be fresh design in this system.

 > I'm thinking we will eitherh ave to go for doing the whole thing in C, or
 > requiring perl5.  Requiring perl5 doesn't sound too bad to me.

Agreed, or possibly a little custom something might be more fun
for who ever does the work.

 > > than we would need.  We can, and ought to, presume a
 > > connection to the web.
 > No.  I see no need for this. ...  That is the way the FreeBSD ports
 > system works; if it can't grab a file, it says "hey buddy, I can't get
 > this!  If you can, put it in directory xxx for me."

I attracted to the thought that the distributed part might consist of
little more than a table of [contribution, checksum, sources] plus
harness that handles it.  Nice to keep the distribution small.  I
guess that doesn't preclude the "hey buddy" exception handler.

 - ben

View raw message