Gregory Sharp wrote:
This is very helpful. Exactly the advice I was seeking! Slowly the process is starting to make sense. I have requested the binNMU for insighttoolkit. (Strangely I didn't get a return email from reportbug. Maybe it is in the queue. Or maybe I made a mistake. We'll see...) Once the binNMU is successful, it seems a "give-back" on plastimatch is needed.
Seems like the binNMU has finally worked, the build log [1] looks much better now. :-)
There are still problems on 2 platforms, but they seem unrelated:1. On arm64 it's "BD-Uninstallable", which is because of a chain of uninstallables: gccxml -> cableswig -> libinsighttoolkit3-dev -> plastimatch. There's probably nothing we can/have to do about it. 2. On hurd-i386 there's a plastimatch-related compilation problem with a constant "MAX_PATH" that is missing on that platform. Looks like plastimatch has been failing on that platform all along because of this problem.
I'm not sure, but since those problems seem to be unrelated, I think that bug #769975 should probably be closed now. What do you think?
Cheers, Martin [1] https://buildd.debian.org/status/package.php?p=plastimatch