[Android-tools-devel] Bug#935308: messed up regex

[Hans-Christoph Steiner]

Adam D. Barratt:
> On Thu, 2019-08-29 at 22:45 +0100, Adam D. Barratt wrote:
>> On Thu, 2019-08-29 at 23:20 +0200, Hans-Christoph Steiner wrote:
>>> Ok, this is embarrasing: after discussing that Debian version regex
>>> and
>>> the version scheme for proposed-updates, I failed to see that the
>>> regex
>>> was wrong (see https://bugs.debian.org/935938).  Its fixed now in
>>> 25.0.0+11+deb10u1.  And hopefully I got the version right.
>>
>> No, that version is higher than unstable.
>>
>> I'm a little confused by what you've uploaded now:
>>
>> android-sdk-meta | 25.0.0+10         | stable           | source
>> android-sdk-meta | 25.0.0+11~deb10u1 | proposed-updates | source
>> android-sdk-meta | 25.0.0+11~deb10u2 | stable-new       | source
>> android-sdk-meta | 25.0.0+11         | testing          | source
>> android-sdk-meta | 25.0.0+11         | unstable         | source
>> android-sdk-meta | 25.0.0+11+deb10u1 | stable-new       | source
>>
>> ~deb10u1 is the source I originally accepted, which generated the
>> duplicate binary package versions. What's ~deb10u2?
> 
> OK, I've now seen the diffs, and the -12 upload to unstable. That
> answers some of my questions, but...
> 
> +android-sdk-meta (25.0.0+11+deb10u1) buster; urgency=medium
> +
> +  * fix version: this adds on top of package from sid
> +
> + -- Hans-Christoph Steiner <hans at eds.org>  Thu, 29 Aug 2019 23:05:54
> +0200
> +
> +android-sdk-meta (25.0.0+11~deb10u2) buster; urgency=medium
> 
> That version would be correct iff the remainder of the changelog was
> from sid.
> 
> Will 25.0.0+11~deb10u2 build correctly-versioned binary packages if I
> accept that one?

Yes, both 25.0.0+11~deb10u2 and 25.0.0+11+deb10u1 work, I just uploaded
the second because I thought that was the proper version number.  Sorry
for the confusion, this proposed-updates process is new to me.  Future
work will be smoother, I promise!


> I noticed that +12 that just hit sid has
> 
>    * Revert "remove broken screenshot2 symlink (Closes: #924175)
> 
> although I can't see why from looking at the bug report. That change
> was also included in the p-u uploads - should it be reverted there?

The buster packages include "remove broken screenshot2 symlink".  The
bullseye/sid package reverts that change because we plan on fixing the
package that provides screenshot2.  That package didn't make it into buster.

.hc