Bug#827043: RFS: gdbm/1.12-1 ITP

[Sven Joachim <svenjoac@gmx.de>]

Hi Dmitry,

I am not a DD and thus cannot sponsor your package, but here are a few
remarks.

On 2016-06-11 11:17 -0400, Dmitry Bogatov wrote:

> I am looking for a sponsor for my package "gdbm"

Thanks for taking care of it!

> Changes since last upload:

Your debian/changelog file is actually missing the entry of the last
upload (1.8.3-14), please incorporate it.  That version already included
a few of your changes, plus there is a patch for bug #774394 which your
package seems to be missing.

>   * New maintainer (Closes: #826639)

Please send a message to that bug that you would like to adopt the
package, and retitle the bug accordingly.

>   * New upstream release

Closes: #682239

Probably there are a few other bugs which are fixed by the new upstream
release, e.g. #340479 - since version 1.9, gdbm has large file support.
Please take a look at them, there are not too many.

>   * Update debian/watch
>       - use HTTP over FTP protocol. FTP have some issues with Tor.
>       - check GPG signature
>   * New binary package libgdbm4 over libgdbm3 due upstream changing ABI.

This means that the upload should go to experimental first, and then you
have to start a transition, see
https://wiki.debian.org/Teams/ReleaseTeam/Transitions.

>   * Bump debhelper compat level from 7 to 9

Already done in 1.8.3-14.

>   * Migrate debian/rules to `dh' sequencer

The old package (besides not using debhelper for most things) had a few
idiosyncrasies, such as the one mentioned in #589761.  Hopefully none of
those are needed anymore, but you should at least look at them.  Also,
you get a fix for bug #436944 for free. :-)

>   * Bump standards version to 3.9.8

Already done in 1.8.3-14.

>   * Remove mainainer scripts that called 'ldconfig', which is called by
>     dpkg trigger.

Ditto.

>   * Replace direct pre-dependency on multiarch-support with ${misc:Pre-Depends}

Ditto.


I guess it would be useful to ship the NEWS and NOTE-WARNING files in
the libgdbm4 package.  The latter is probably quite important on 32-bit
architectures.

Thanks for your contribution, and good luck in finding a sponsor.

Cheers,
       Sven