[gentoo-dev] Bugzilla things

[C. Brewer]

I'd like to see more user submitted ebuilds make it into the tree..but 
whatever. My latest gripe is the test-request option, let's kill that bs 
right off. Kick it to wontfix or upstream or change it to wtf?, but I think 
packages are moving too fast in general for test-request. By the time users 
and devs figure out what happened with package x-1.0 and straighten it out, 
we can be up to package x-1.1-r3 and then any fix that went into the former 
package tends to wasted time on someone's part.

An example of the can be seen at http://bugs.gentoo.org/show_bug.cgi?id=23111
where a bug was filed because usbutils wouldn't compile against 2.6 headers 
( no surprise there ) and a suggestion was made and asked for test-request, 
with a few flaws to the whole process-

1)   The devs asked said user to re-emerge linux-headers, which with no 
official 2.6 headers in the tree at the time, would have kicked him back to 
2.4 headers,solving the problem with the merge but not actually touching the 
real problem reported. This is probably acceptable by dev standards.
2)   The user didn't respond back (bad user!!!)
3)    I had a similar problem, saw I couldn't completely solve it but saw a 
way around it without requiring anyone to regress their 2.6 headers, and 
submitted my patch...where it still sits under the guise of test-request and 
is now pretty much useless as usbutils somehow worked itself out with regards 
to the 2.6 headers. 

   So QA on new ebuilds is a bitch, and responsibility for them may preclude 
many of them from submission, but how about those patches to existing 
maintained packages? I'm sure that mine isn't the only such case...but that 
was 30 mins I could've spent wondering where those topless college chicks 
were at when I was college:)

Please pay more attn to patch submissions, pretty please?
---
Chuck Brewer
Registered Linux User #284015
Get my gpg public key at pgp.mit.edu!! Encrypted e-mail preferred.




--
gentoo-dev@g.o mailing list