Re: [Gnumed-devel] CVS guidelines

[Karsten Hilbert]

> I feel that most of the rules makes sense - except those suggesting that we
> should have a peer review process.
I think Sebastion means that extensive changes require
extensive discussion as opposed to refusal to explain
stuff even when explicitely requested.

> number of patches. IMHO this review happens already: I for my part always look
> at the changes others made to my code and to code that might be important
> for the parts I coded.
Absolutely ! Way to go. That's why I am REQUESTING meaningful
commit messages on individual files (well, sometimes a few
files can be committed grouped as in "attribute name so-and-so
changed, so go use it"). That's also why I REQUEST not
committing files that haven't changed, really. Yes, I do
understand that some tools in the ilk of Cervisia might decide
for you that some file appears to have changed and commit it
but I don't care what tools you use. I REQUEST that unchanged
files don't get committed by force. Yes, you may still commit
almost-unchanged files but you need to put in a meaningful
commit message such as "whitespace fixup" etc.

> I believe that others do the same (Karsten usually
> cleaned up and fixed the changes I did in his code, I hope he wasn't t
> too embarassed that I changed it without previous review).
Nope. Your patches are of high quality. So are Ian's. Now, the
*concepts* within Syan's code may well be of high quality,
too, but that massive commit was poor.

> IMO the most important rule is that you have to think before commiting.
ACK. Couldn't agree more. Have admittedly violated that rule myself
before.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346