FTBFS gcc8

[Andreas Hasenack]

https://launchpadlibrarian.net/379507226/buildlog_ubuntu-cosmic-armhf.pam-p11_0.2.0-1_BUILDING.txt.gz

pam_p11.c: In function ‘randomize’:
pam_p11.c:556:36: error: comparison of integer expressions of different signedness: ‘ssize_t’ {aka ‘int’} and ‘unsigned int’ [-Werror=sign-compare]
  if (0 <= fd && read(fd, r, r_len) == r_len) {
                                    ^~

It fails in i386 and armhf (32bits only then?).

[Christian Ehrhardt  (paelzer)]

This was long lost (60 days) as it came in via an auto sync so nobody kept track.
Thanks to a new report it is now on our radar (and future issues will be picked up sooner).

For this case we have a few options:
1. fix the issue in the code, but realize it is not a new issue just an old one exposed by a new warning
2. disable this particular new compiler warning for now and expect that upstream will have improved e.g. until next release
3. try to build the older version and do a +really upload to revert to that for now.

I'll do some checks to decide between those three and then report back here.

[Christian Ehrhardt  (paelzer)]

Debian (in the meantime) is affected as well (new compiler -> FTBFS).
It is available as an update in -unstable but did not migrate to testing in Debian yet.
(As it did not migrate to -release in Cosmic).

[Christian Ehrhardt  (paelzer)]

Results of my checks:
- 1.6 would build fine (see [1]
- there is an upstream fix which would fix it [2] built in [3]

The question now is, do we "want" 2.0 or 1.6 in Cosmic considering the time in the release.
I'll discuss that with some peers and then prep it one of those ways for review.

[1]: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3421/+packages
[2]: https://github.com/OpenSC/pam_p11/commit/55b3888d5e9208948a40725350ca2f50020c4838
[3]: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3422/+packages

[Christian Ehrhardt  (paelzer)]

After considering the upstream release notes [1] not containing anything that seems breaking (but fixes we would want).
Also no issues reported for it being in Debian-unstable at least for 60 days as well as being release upstream since may.

We might hit issues either way, currently I think picking the 0.2 version is the better choice.
Due to that this is what I'll prepare.

[1]: https://github.com/OpenSC/pam_p11/releases

[Christian Ehrhardt  (paelzer)]

Also pinged Debian to know of the fix so they can get it right as well.

[Christian Ehrhardt  (paelzer)]

Uploaded to cosmic, tracking migration ...

[Launchpad Janitor (janitor)]

This bug was fixed in the package pam-p11 - 0.2.0-1ubuntu1

---------------
pam-p11 (0.2.0-1ubuntu1) cosmic; urgency=medium

  * d/p/fix-gcc8-build-unsigned-comparison.patch fix FTBFS with
    gcc8 (LP: #1793135)

 -- Christian Ehrhardt <email address hidden> Wed, 19 Sep 2018 09:59:58 +0200