On 12/22/2011 04:33 PM, Jon Peatfield wrote:
> A quick glance in the 5x/SRPMS/ shows no matching srpm to the binary 
> rpms in 57/{x86_64,i386}/SL/ though there is one in 
> 5rolling/SRPMS/vendor/ which seems to match in NEVR but appears that 
> it might be the wrong place unless TUV added it between 5.6 and 5.7...
>
> The reason that I'm looking at this is that for many years we have 
> been using alpine with a bunch of extra (Chappa) patches, and while 
> this new rpm is called alpine (with a larger version number!) it is 
> based on re-alpine which for various reasons does not include all 
> these existing patches - so from the point of view of our users would 
> be inferior.
>
> After talking to Eduardo Chappa I get the impression that he 
> originally offered the patches to the re-alpine team but they didn't 
> want them.  On the other hand they seem to claim that he refused them 
> permission. (sigh!)
>
> Meanwhile re-alpine had an initial burst of activity and now seems 
> somewhat less active (last release was in October 2010), though there 
> might well have been commits since then...
>
> Anyway is the alpine-2.02-2.el5.src.rpm from 5rolling/ the right one 
> to for me to look at?  I've build a test rpm based on that with the 
> extra stuff we add in, and in trivial testing it *seems* to work for me.
>
> (btw the specfile in that srpm sets alpine to use hunspell as the 
> spell checker, but none of my sl5 machines have that... though it is a 
> trivial change to aspell instead...)
>

You are correct, the alpine source from 5rolling is the right one.  Not 
sure why that didn't get pushed into the 57 source.  I'll have to look 
into that.

I'll have to check on the hunspell usage as well.

Pat

-- 
Pat Riehecky
Scientific Linux Developer