Bug#1028315: lintian: anticipated-repack-count triggered even if repack-count is in actual use

[IOhannes m zmölnig (Debian/GNU) <umlaeute@debian.org>]

Package: lintian
Version: 2.115.3
Severity: normal

Dear Maintainer,

thanks for 'lintian' I love and use it a lot.

Today I'm taking issue with the "anticipated-repack-count" tag, as
lintian complained that i was using a version "0.9.20~dfsg1-1"
(nevermind the tilde for now).

i take offense, because the previous upload of the package in question
was "0.9.20~dfsg0-6", so incrementing the repack-count to "1" is correct
(imho).
it definitely is not an "anticipated" repack count, but a counter in
use.

i guess (not having checked), that the lintian tag simply checks if the
upstream version ends with "dfsg1".

i think a better strategy would be to check whether the repack count
incremented with regard to the previous upload.

that is:

| old version| new version| reaction | reason                           |
|------------|------------|----------|----------------------------------|
| 0.19…      | 0.20+dfsg0 | WARN     | new upstream (no reason to count)|
| 0.19…      | 0.20+dfsg1 | WARN     | new upstream (no reason to count)|
| 0.20       | 0.20+dfsg1 | WARN     | first repack (no reason to count)|
| 0.20+dfsg0 | 0.20+dfsg1 | ok       | consecutive repack |

thanks for making Debian a better place.