[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: Playing the KDE CVS roulette
From:       Bernhard Rosenkraenzer <bero () redhat ! de>
Date:       2001-12-26 15:05:57
[Download RAW message or body]

On Wed, 26 Dec 2001, Bryce Nesbitt wrote:

> But this is a nightmare when trying to get a stable version to compile.  

If you need a stable version, go to ftp.kde.org. ;)

> I often have to make clean and recompile from scratch.  It sometimes
> takes days to get back to a working KDE3 after starting to update.

By any chance, is your setup something along the lines of
"Hm, kdelibs didn't compile, so I'll stick with kdelibs-as-of-2-weeks-ago 
and compile kdebase"? That won't work all the time - things in kde* often 
depend on recent changes in kdelibs.

I have a cronjob that compiles KDE from CVS every day, and mails me if 
there are any errors. Most of the time, it finishes without errors.

Another thing to watch out for is your compiler. If you're using something 
ancient (e.g. gcc 2.{7,8}.* or egcs), some things will break because the 
compiler doesn't speak proper C++.
If you're using a very recent gcc (3.0.x or pre-3.1), things will 
occasionally break because the compiler enforces standards very strictly, 
and developers using an older compiler won't notice there's anything wrong 
with their code.

> Is there a better way?  Could there be a quiet period each 24 hours
> where developers are expected to check in only compile fixes?

That would be a rather boring day, because the tree usually compiles. Your 
best bet is to use a Thursday tree, right before BIC Friday starts in the 
earliest timezone.

> An extension to CVS to make multi-file checkins atomic?

That could break quite a few things. Using a new foo.cpp file with the old 
foo.h file is usually a horrible idea.

LLaP
bero

-- 
This message is provided to you under the terms outlined at
http://www.bero.org/terms.html


 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]