httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Laurie <>
Subject Re: let's see the patches
Date Wed, 13 Mar 1996 11:58:11 GMT
> > Seeing a patch mailed to the list or browsing them on
> > is very convient, simple and fool-proof for some of us. If we have 
> > that *with* cvs for those who want to play with it 
> > then everyone is happy.
> You can browse them on now. cvs really is no effort at all if
> you can login to hyperreal. It's as simple as cvs diff -r 1.x -r 1.y where
> x and y are given in the mail message. That's about as straightforward as
> ftp'ing to the site and downloading a patch.
> I don't see that diffs will really be that much use to people because they'll
> be against the last revision, so unless you're keeping your local sources
> up to date anyway then the diff won't be much use to you.
> Learning cvs is a 1/2 hour job max to do the basics necessary for using it
> with Apache. If you're really so busy that you can't spend that much time on
> it then please don't take it upon yourself to review code because that's
> soemthing that should not be rushed!
> Installing and using it remotely is a different issue altogether but up
> until a month ago people had to connect to hyperreal anyway. If you don't
> want to bother with all the remote cvs hassle then just keep doing what
> you were doing before i.e. logging in to hyperreal, just spend a few mins
> learning the one or two commands you need to use cvs instead of ftp!
> A really basic tutorial for Rob and Andy to get diffs:
> Put this in .cshrc or whatever
> setenv CVSROOT /export/home/cvs
> Do this once the first time you ever use it.
> mkdir cvswork
> cd cvswork
> cvs co apache
> Now whenever you want a diff of all the latest changes do
> cd cvswork
> cvs diff
> I mean, is that really so much work?
> If you want an actual patch of the mail message you were looking at then do
> cvs diff -r 1.x -r 1.y
> where x and y were given in the mail message.
> There's an implicit requirement in using cvs that some people are having
> a hard time with. If you give someone cvs access then you trust their
> abilities to produce good code, if you don't think they're technically
> capable then don't give them access. 
> The old method worked fine when 5 or 6 people made small patches and it was
> easy to communicate with each other. Apache is now much much bigger and
> growing all the time and no single person can read every single line that
> gets submitted and people have to accept that and trust the individuals
> making the changes. Any real problems will come out in the wash, so to speak,
> as long as we make sure the code is getting used and is fully tested before
> release.

Just one snag - many people don't have logins on hyperreal. Until CVS there
was no requirement for a login to participate. I think we should keep it that

I'll try to find some time to automate some patch extraction for non-CVS users,
unless someone else wants to volunteer?



Ben Laurie                  Phone: +44 (181) 994 6435
Freelance Consultant and    Fax:   +44 (181) 994 6472
Technical Director          Email:
A.L. Digital Ltd,           URL:
London, England.

View raw message