[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Are mails sent to xxxx <at> buildd.debian.org sent to /dev/null ?

Blars Blarson <blarson@blars.org> writes:

> In article <[🔎] 20041218093530.GD8881@mauritius.dodds.net> vorlon@debian.org write:
>>> What would help would be an email address where any DD can send a
>>> signed mail to request a rebuild or to set a dep-wait or a build
>>> failure.
>>> There could be a webpage where one selects the package(s) and
>>> architecture and it generates a mail template that just needs to be
>>> signed and send in case you fear the syntax is to complex.
>>I wouldn't worry nearly so much about syntax as I would about DDs blindly
>>requeuing packages that are going to fail, and crushing the buildd network.
>>As it is, I've seen far too many botched attempts at fixing FTBFS bugs, by
>>maintainers who didn't understand why their packages were failing; I think
>>giving DDs a button to push any time one of them assumes the buildd is at
>>fault (which is often assumed) is of dubious benefit.

Or create a group of people with the right. In effect create some
redundant buildd admins. They can be second class admins, i.e. they
don't have an actual buildd they control. They might not even need
access to one.

> Make such rebuilds lower priority than new builds, and have a
> mechanism to block DDs who abuse it.  The latter might even be able to
> be automated based on percentage of requested rebuilds failing.
> On sparc, about 1/3 of the "maybe_failed" builds are due to simple
> build dependancy problems.

It would be nice if one could upload say KDE and preseed the
wanna-build with all the Dep-Waits one know will happen. There is
realy no point for the buildds to waste hours on installing
build-depends just to notice one of them is too old and fail.

I think it would be worth a shot to allow DDs to set Dep-Wait for
their packages. Sure there would be some mistakes but even DDs can
sometimes be thought. And if it doesn't work out there isn't realy so
much harm done. The buildds will have a small hickup and then go on


PS: All of this post sarge. No point risking a hickup during release
when it would hurt.

Reply to: