Hey,
Nathan Scott <nathans@xxxxxxxxxx> writes:
> ----- Original Message -----
>> [...]
>> I've focused the changes within pcp.spec.in to start, so to try out the
>> changes, simply ./Makepkgs on a fedora or centos box.
>
> Its looking pretty good I reckon, and the meta-packages are very helpful.
> Couple of small nits in the attached patch, mostly small stuff.
>
> The one real issue there is the pcp-libs Requires: pcp-compat - I don't
> really follow that, and it'd seem like it might cause problems. For the
> case where someone has *only* pcp-libs installed before an upgrade (e.g.
> maybe they have built a monitor tool that depends on libpcp.so) - would
> this not pull in everything, instead of only the new -libs?
Thanks for taking a look. Your changes look good to me, and good catch
on the pcp vs pcp-libs dependency on the pcp-compat package.
For your comment regarding the systemd pmda being commented out, the
regex being used was causing some issues including files that did not
belong with the pmda, I've fixed that at this point, including your
patch on my lberk/dev branch. Still at sourceware.org/git/pcpfans.git
[...]
>> I've been careful to test qa after splitting off pmdas of each language
>> (python, perl, and C, respectively), and there is various updated qa
>> bits to match that. As it currently stands, I have no regressions on my
>> local runs.
>
> Nice. There may be some tests that warrant new _notrun() checks, going
> forward, since more components (esp PMDAs) are now becoming optional.
Yes, there are a handful that I've added for each flavour of pmda.
> We might need to mark %ghost on .NeedRebuild? (I think rpmlint complains
> otherwise, from memory)
I'll have to take a further look into this.
Cheers,
Lukas
|