[Debian-med-packaging] [bambamc] tag debian/0.0.50-1 created (now e9f14c4)

[Andreas Tille]

On Thu, Jun 05, 2014 at 12:36:23PM +0200, rf at q-leap.de wrote:
> 
>     Andr> Just UNRELEASED is fine and the final uploader will set the
>     Andr> tag.  The rationale is that different Sponsors could pull the
>     Andr> status of the repository and it is good to know whether
>     Andr> something is just uploaded or not (without checking the Debian
>     Andr> mirror).  It could always be that I'm tempted to do some last
>     Andr> minute minor changes which would require fiddling around with
>     Andr> tags.  Or it might happen that the upload is not done in a
>     Andr> timely manner and some tag is hanging around for a long time
>     Andr> with no corresponding upload.  So this method is at least this
>     Andr> what IMHO makes the most sense.
> 
> I see your point. In the (hopefully) rare case the uploader needs to
> make some changes, he could simply reset the tag using 
> 'git tag -f <tagname>' and tell the submitter he should update his tags
> from remote (like you just did with bambamc).

Yes, that's for sure possible.
 
>     Andr> I have some other sponsee in Debian GIS and he prefers to
>     Andr> finalise the package with "unstable" + tag - that's also a way
>     Andr> to do things since the drawbacks I explained above are usually
>     Andr> quite rare in practice.  We just need to agree upon a common
>     Andr> procedure.
> 
> I'd opt for this as well, since we already need a tag to be set for our
> internal build scripts to work.

I can cope with this as I said and since you have obviously a good
reason that's fine for me.  In this case its your duty to make really
sure that the package will be uploaded soon and bother the DDs of the
team to quickly sponsor the package.  I'm pretty sure nobody will call
you impatient if you send a ping every second day (since in most
practical cases this will not be needed anyway).
 
Kind regards

       Andreas.

-- 
http://fam-tille.de