[Xorg] X.org CVS branches need to change

Egbert Eich eich at pdx.freedesktop.org
Wed Mar 17 07:27:36 PST 2004

Keith Packard writes:
 > Around 18 o'clock on Mar 16, Egbert Eich wrote:
 > > One thing we have never talked about is code review.
 > > I'd feel a little nervous if everybody is allowed
 > > to work directly in the trunk adding unreviewed code.
 > It would be nice to see a culture of peer review inculcated in the
 > community.  I'm not sure I see us able to instantiate a policy on the
 > matter though.  As with branching, I'd like to encourage people to ask 
 > questions about practice and procedures for making non-trivial changes in 
 > the code base.

Looks like other projects do extensive peer review. I've talked
to gcc people: there almost everything seems to get discussed on
mailing lists before it gets committed - at least if these changes
go into a very central piece of code.

 > > If this will not produce security concnerns it may 
 > > introduce inconsistency as people may not have the 
 > > experience on how things are handled and what functions 
 > > exist.
 > That's one of the main reasons I will be doing development only in the
 > modular environment; each developer there gets to see a stable world aside
 > from their own piece, updates to external packages happen only when they
 > have reached a stable point and have been tested to ensure some modest
 > level of compatibility.

Some of your work requires changes on many modules even if you don't
want to fully integrate things but just want to test it. Therefore
the world around you may not be all that stable.

 > In either case, it's hard to find people interested and able to evaluate 
 > every patch that goes into a system as large as X, especially if that 
 > process is seen as burdensome on overall development.  Having developers 
 > identify patches to critical parts of the system, or which are non-obvious 
 > in some way and getting those looked at by someone else would be a good 
 > start.

I agree, however the non-critical and non-central pieces are the ones
that get nasty and pop up when you least expect it. I had this happen 
to me with Xnest. I came across a problem in Xnest half a year after
it has been checked in. Appearantly nobody ever bothered to test Xnest
during all that time. On the other hand Xnest seems to be importand 
enough to some people so that bugs in there get escalated to me...


More information about the xorg mailing list