Should we eliminate make.inc.example?

[rbabich]

There seems to be increasing danger of make.in and make.inc.example getting out of sync. If there's consensus, I'd be in favor of going over completely to Balint's autoconf approach.

[maddyscientist]

You've got my vote.

[gshi]

Out of sync is not big of deal in the devel branch. We only need to make sure it is synced when we make a release.

[bjoo]

Syncing is easy in principle, just do a build to test something and then copy the make inc over. However, it is a step that we're always likely to forget to do :(

Do we want to make a 'release checklist' that we go through prior to a release?
(even if we don't enforce it rigorously it may be useful to aid our release process)

[rbabich]

I have a release checklist that I'll update. Maybe I should post it to the wiki (on github)?

[bjoo]

It can also be kept under revision control :)

[maddyscientist]

I have decided to make a unilateral decision and remove make.in.example from the library. With the increasing number of compilation options which have dependencies, e.g., I've just added QIO support which requires QMP, I think maintaining make.inc.example and make.inc is both a waste of resources, and can lead to compilation errors.