PBORI_HAVE_M4RI, PBORI_HAVE_NTL and bad API about brial HOT 8 CLOSED

I will likely drop PBORI_HAVE_NTL. It is likely a remnant, ntl is not a dependency of brial and hasn't been for quite a while as far as I remember.

Will the change require bumping LT_CURRENT or LT_REVISION in your opinion.

from brial.

Hmmm... I would say no... But it's a good question : I'm unsure...

from brial.

Removing PBORI_HAVE_NTL leads to some code simplifications that are not obvious at first glance. I don't think we are changing any API technically, so no bumps. But some code blocks will go.

from brial.

I hate the fact that the code can be different formatting style in the same file. Make things harder to follow properly. But standardizing the style would basically destroy the history.

from brial.

Because m4ri is a requirement at configure time, I think I will also eliminate all conditionals on PBORI_HAVE_M4RI in the code. I should still add it to config.h.in. That will completely get rid of the problems with faugereStepDense.

from brial.

OK, master should have stuff for all these now. Let me know if your problems are still happening. And don't hesitate to signal any other issue.

from brial.

I am preparing 1.2.7 which should have a sdist on github in the next 24 hours. For the record, following libtool guideline I incremented LT_REVISION. You should do that for any code change. I didn't bump it in 1.2.6 because the release only touched the sage python code part of the tarball.

from brial.

OK. Closing. Any new stuff will be for a new issue.

from brial.