[with patch, needs work] codes over rings

[wdj]

This module constructs codes over rings of the form ZZ/mZZ, that is, submodules of FreeModule?(IntegerModRing?(m), n). The main authors are Cesar Agustin Garcia-Vazquez (who was an undergrad in Mexico when he wrote this) and Carlos A. Lopez-Andrade (his advisor). I made some changes to make it more consistent with LinearCode?. (It still has some hidden differences - the basic problem being that FreeModule? has no submodule or span method analogous that of VectorSpace?.)

It is in Cython, which I confess I don't really understand well. My role is simply to take Cesar's code (which he emailed to me), tweek it a bit, and create a patch. He has explicitly agreed to distributing it under GPLv2+.

[wdj]

applies to 4.1.alpha1

[wdj]

This applies to 4.1.alpha1. This passed sage -testall except for

The following tests failed:                                                                             


        sage -t  "devel/sage/doc/fr/tutorial/programming.rst"
        sage -t  "devel/sage/sage/misc/darwin_utilities.pyx" 
Total time for all tests: 6017.1 seconds                     

Neither of these failures seem related to this ticket.

[AlexGhitza]

The patch doesn't contain the crucial file ring_code.pyx. Maybe you forgot to add this file to the hg repository?

[wdj]

ignore the previous patch; this applies to 4.1.2.rc2

[wdj]

This latest patch fixes some problems prointed out in private emails from Dan Gordon.

[wdj]

I can also say that the module in coding pass sage -t, with this patch applied, in OS 10.6 on an imac. OS10.6 does not pass sage -testall. However, I do have ubuntu 9.04 32bit installed under vmware on this machine and can test it there if desired.

[mhansen]

Is there any way we can get documentation for the cdef methods as well as possibly some more code comments? I'm just worried that there's no one who is able to maintain this code.

Also, it's not clear to me that there are not memory leak issues with this code as there are lots of malloc's and only one free.

[wdj]

If someone can explain to me what those functions will do, I will add comments on them. As I said, I am basically doing a favor by posting a patch of someone else's code. I had to make a lot of additions to make it consistent with other parts of the codes library, but basically it is Cesar's code. Personally, I really don't understand the cython stuff.

Does this mean that the code will not be accepted?

[wdj]

Replying to mhansen:

Is there any way we can get documentation for the cdef methods as well as possibly some more code comments? I'm just worried that there's no one who is able to maintain this code. Also, it's not clear to me that there are not memory leak issues with this code as there are lots of malloc's and only one free.

As I said, I could try to add comments to the code, but I really don't know cython nearly well enough to deal with malloc issues.

I've emailed Cesar and not gotten any response about the memory leak issues.

Given that, should the status change to "resolve as won't fix"? That would be too bad since Sage could really use "linear codes" over rings.