[Asterisk-Dev] build failure of cvs head cvsup'd ~1200 UTC 06-06-05

[Steve Kann]

Kevin P. Fleming wrote:

> Jerris, Michael MI wrote:
>
>> Is it time to set up a little compile farm for testing at least compile
>> success on commits?  This has bit us several times and has been an issue
>> with things from different shell versions, make versions, gcc versions,
>> ect.  How hard would it be to run everything through an automated
>> compile test before commit?
>
>
> Actually, that's on the agenda to be discussed this week at Digium. I 
> want a 'build farm' to test commits, on various different operating 
> systems and CPU architectures, before the commits go into the public 
> tree. Even the best of us will make mistakes like this, and I can't 
> predict when something will fail to build on FreeBSD or Darwin, so I 
> think it's important that we get this set up to try to ensure that CVS 
> commits are as well-tested as possible.


Check out "tinderbox" from Mozilla..

The mozilla tinderboxen sit in a loop, checking out CVS, compiling, and 
doing some basic tests, over and over..

http://tinderbox.mozilla.org/showbuilds.cgi?tree=SeaMonkey

It doesn't test patches before they get into CVS, but it gives you 
feedback shortly after you commit, and then keeps track of who is "on 
the hook" -- i.e. whose commit broke the build, and what that commit was..

-SteveK