Bug#1000859: transition: rtmidi

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

Package: release.debian.org
Severity: normal
User: release.debian.org@packages.debian.org
Usertags: transition


The latest upstream version of RtMidi includes a soname bump of
librtmidi. The Debian packages have updated their names accordingly.

The new version of rtmidi has been lingering in experimental for a while
and I've done test-builds of all reverse-dependencies (using 'ratts')
which includes the list of immediate rdeps found at
  https://release.debian.org/transitions/html/auto-rtmidi.html.
The list of immediate reverse dependencies is:
- bambootracker
- din
- giada
- midisnoop
- milkytracker
- octave-audio
- pianobooster
- stk
- polyphone

Of these 'milkytracker' FTBFS with the new librtmidi-dev, due the latter
now requiring C++11 (and 'milkytracker's buildsystem enforcing C++98).
I've filed #1000855 with a suggestion on how to solve this.
All other immediate rdeps build without problems.
I don't expect any runtime problems.

As for the intermediate rdeps, I found two more problems:
- pd-py (FTBFS is most likely unrelated; this package depends on python2
  and has been removed from 'testing' a while ago)
- fasttracker2 (this is in non-free, and i only tested main; the package
  nominally depends on librtmidi-dev, however checking the build logs it
  seems that it is really using vendored libraries instead)

Thanks for your consideration.
mfgad
IOhannes


Ben file:

title = "rtmidi";
is_affected = .depends ~ "librtmidi5" | .depends ~ "librtmidi6";
is_good = .depends ~ "librtmidi6";
is_bad = .depends ~ "librtmidi5";