Ticket #1611 (assigned enhancement)

Opened 1 year ago

Last modified 3 months ago

polybori should use the m4ri library from libm4ri spkg

Reported by: burcin Assigned to: burcin (accepted)
Priority: major Milestone: sage-3.2.2
Component: packages Keywords:
Cc: PolyBoRi

Description

PolyBoRi? comes with its own copy of m4ri libraries. After #1505, Sage provides these libraries through a package.

The PolyBoRi? build process should be changed to use the library and headers provided by libm4ri-*.spkg.

Change History

12/27/2007 04:24:54 AM changed by malb

Michael Brickenstein wrote:

If you can wait until the end of my holidays, which will begin tomorrow and end at 6th of January, I will do the switch myself.

01/03/2008 07:07:54 AM changed by burcin

  • status changed from new to assigned.
  • milestone changed from sage-2.9.2 to sage-2.10.

02/15/2008 03:23:43 PM changed by mabshoff

The issue might be resolved during the update to PolyBoRi? 0.2.rcX during Bug Day 10.

Cheers,

Michael

02/17/2008 11:57:01 AM changed by mabshoff

This should be done while updating PolyBori? - see #2060.

Cheers,

Michael

03/08/2008 05:37:53 AM changed by burcin

PolyBoRi requires some changes in the m4ri libraries. They are waiting for changes upstream (in m4ri) to switch over, so newer versions of PolyBoRi don't support this.

03/09/2008 08:27:38 AM changed by malb

I guess I am upstream so I should diff their m4ri against ours (upstream).

05/11/2008 05:12:36 PM changed by mabshoff

Fixing this bug is planned for the PolyBoRi? 0.5 release.

Cheers,

Michael

09/06/2008 05:41:58 PM changed by mabshoff

  • cc set to PolyBoRi.

#3264 does not fix the problem yet. Are there any options to make this happen with PolyBoRi? 0.5?

Cheers,

Michael

09/06/2008 05:59:07 PM changed by malb

There is some movement w.r.t. this issue in PolyBoRi?, but I wouldn't suspect it to be available in 0.5.

09/07/2008 11:23:24 PM changed by PolyBoRi

it will be available in 0.5. However, you should make sure, that malloc work also for allocating zero bytes in M4RI, else it will crash (doesn't work with mm_malloc on Mac OS X 10.5).