Source: petsc Followup-For: Bug #983892 Control: tags -1 help I'm looking into this problem. It's not immediately obvious where the problem ultimately lies for this reason: the petsc64 scripts have been constructed by analogy (copy&adapt) with the correspond petsc (petsc32) scripts.
The prerm scripts have been configured to remove these alternatives. At first scrutiny, libpetsc64-complex3.14-dev.prerm seems to be addressing the alternatives in the same way as libpetsc-complex3.14-dev.prerm (apart from handling some old legacy alternatives which shouldn't be affecting libpetsc64-complex3.14-dev) If libpetsc-complex3.14-dev is uninstalling cleanly as intended then why would libpetsc64-complex3.14-dev not be doing the same? Likewise, libpetsc64-complex3.14-dev should be behaving the same as libpetsc64-real3.14-dev. Why is this problem only occuring with libpetsc64-complex3.14-dev ? If anyone can spot the difference then I'll be happy for the help. Drew
