Gerard Beekmans wrote:
Jeremy Huntwork wrote:
The upstream keyword would certainly work. If Matt would like, you
could also add either a new ticket type or a new component to reflect
that the problem lies upstream. There's several ways to skin this cat. :)
And think about a scenario like this one: What do we do then when a
developer denies that the bug is upstream (it happens...sometimes
developers are too proud and will not admit a bug that is blatant in
their software).
Ah, you must have met Joerg Schilling via LKML already then? :-)
Honestly, most upstream developers I've been in contact with have been
only too happy to help. Either they'll quickly acknowledge it's their
problem, or at the very least help to debug exactly where the problem lies.
It wouldn't be LFS-specific but it can't be marked as
upstream if you want to keep on good terms with said developer.
Well, if the developer can't prove that the fault we report is entirely
because of the way we've a) compiled and/or b) configured their package
or our environment then the only thing we can do is mark it as an
upstream defect. Obviously, we can't force upstream to acknowledge it
or apply our patches, but marking bugs in this way would at least store
this IMO important information about the cause/origin of the bug(s).
At this stage, I've a preference to marking this via a new ticket type
of "Upstream Defect". It's obviously important though that we keep
track of when and where such defects were reported upstream and the
resolution of any subsequent discussions.
Regards,
Matt.
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
