[Bug 458054] Review Request: arm4 - Application Response Measurement (ARM) agent

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=458054





--- Comment #14 from Jon Ciesla <limb at jcomserv.net>  2008-11-04 12:02:37 EDT ---
(In reply to comment #13)
> Updated versions:
> 
> Spec URL: http://arm4.org/Downloads/0.8-0.5/arm4.spec
> SRPM URL: http://arm4.org/Downloads/0.8-0.5/arm4-0.8-0.5.fc9.src.rpm
> 
> I must say though, it's kind of annoying to hit an rpmlint update in the few
> short hour between submission and review :)

Fortunately that isn't THAT common. :)

> Some comments on these reported errors though. First of all, despite the
> assertion of rpmlint, there are many cases where this is a valid design choice,
> such as when a common error handler is put in a library. In this case, it was
> used in panic conditions only, and I still think that's valid. At this point,
> the fix is largely cosmetic, although I will address this more completely for a
> future release.
> 
> Which brings me to the larger issue. Fixing this will always require an
> architectural rethink. At the very least, it will require a change to the
> library version number, potentially putting it out of sync with the upstream
> version. Ultimately this isn't the responsibility of the packager, but of the
> original developer, and as pointed out already the developer may just say deal
> with it. Honestly, if I weren't also the upstream developer, I wouldn't have
> changed anything. As it is, I'm fixing internal libraries that provide no
> public API for problems that I don't think exist.
> 
> I guess my point is that this warning should largely be ignored during reviews.
> The dangers of not doing so are too great.
> 
> OK, rant done, the new code removes these warnings.

If you really feel that strongly about it, why not make that case in a bug
against rpmlint?  It's always being changed to adapt to different use cases. 
And remember, rpmlint silence is not a prerequisite for approval, especially
where warnings are concerned, it's more there to encourage discussion,
documentation, and to make sure that both packager and reviewer have a better
idea what's really going on.

> With regard to your comments on the practice review, I did use the checklist at
> http://fedoraproject.org/wiki/Packaging/ReviewGuidelines. I'll do a better job
> of noting the pass/fail states next time. I'll do a couple of additional
> reviews later this week when I get some of my work out of the way.

Sounds good.  One or two more, plus this package's approval, and we're in
business.  I'll post a formal review of this package soon.

> Thanks!
> Dave

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.