[Bug 609728] Review Request: sparsehash - Extremely memory-efficient C++ hash_map implementation

[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=609728

--- Comment #3 from Kalev Lember <kalev at smartlink.ee> 2010-07-02 17:36:44 EDT ---
(In reply to comment #1)
> Two suggestions:
> 1. move all files too -devel subpackage

When I was creating the package, I couldn't make up my mind whether to put
everything in -devel or not. If you think it's better that way, I'll move all
files over to -devel.


> 2. It'll better to leave this package as noarch, I think test in not very
> useful for development libs. You can add %check to packages which require
> sparsehash.    

I respectfully disagree here.

Building the package as noarch means that it's built only once for every
primary architecture we have. For F-12, it's only one build for i686, x86_64,
ppc, and ppc64. That would mean not running self tests on 3/4 architectures.

If we have a package which comes with self tests, we should try to run them if
possible. You are saying these tests should be moved to packages which require
sparsehash; I disagree. Every component should test itself if it is possible.
And besides that, a large part of sparsehash package consists of self tests; if
they aren't necessary why do you think upstream wrote them in the first place?

The package is also less than 0.5 MB, which wouldn't be much of a saving on
mirrors either.

-- 
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.