Bug 29924

Summary:	Is Maxima broken? Some functions don't run any more
Product:	Mageia	Reporter:	Adelson Oliveira <adelson.oliveira>
Component:	RPM Packages	Assignee:	Mageia Bug Squad <bugsquad>
Status:	RESOLVED WONTFIX	QA Contact:
Severity:	normal
Priority:	Normal	CC:	fri
Version:	8	Keywords:	UPSTREAM
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:	maxima and wxMaxima	CVE:
Status comment:

Description Adelson Oliveira 2022-01-21 23:05:43 CET

One of the mathematical functions of Maxima doesn't go any more.

Install wxMaxima or only maxima package

1 - launch wxMaxima
2 - issue h:matrix([2.0,0.2+%i*0.1],[0.2-%i*0.1,1.8]);
3 - issue cholesky(h);

  Unable to find the Cholesky factorization
   -- an error. To debug this try: debugmode(true);

I've checked the files in /usr/share/maxima/5.44.0/share and it seems there is no reason to the function not to be seen.

There is a bug (28352) where the maintenance of wxMaxima is said to be a concern...

Comment 1 sturmvogel 2022-01-22 05:51:29 CET

That seems not to be a problem with any symbolic linking or the program itself but with your matrix formula
https://www.google.com/search?q=Unable+to+find+the+Cholesky+factorization

Comment 2 sturmvogel 2022-01-22 06:47:50 CET

I tried the upstream version as appimage. https://sourceforge.net/projects/maxima/files/Maxima-Linux/5.44.0-Linux/

There the same error occurs. So it is either an upstream issue in version 5.44 or your formula/matrix is wrong as already mentioned.

Comment 3 sturmvogel 2022-01-22 07:24:18 CET

Same error occurs also on Opensuse Tumbleweed with wxMaxima 21.11.0 and maxima 5.45.1

So no Mageia problem. Either upstream or wrong matrix type as the cholesky function needs spezial types of matrix.


Bugreport against Mageia version is invalid!

Comment 4 Morgan Leijström 2022-01-22 12:48:01 CET

Closing as such then.

Thank you sturmvogel for the analysis and tests!

Thank you Adelson for reporting what could have been a problem.

Bug 28352 mentioned in comment 1 is about it needing to be rebuilt for shipping another version of wxgtk3.1, now handled in Bug 29848, tested verified in https://bugs.mageia.org/show_bug.cgi?id=29848#c65

CC: (none) => fri
Resolution: (none) => INVALID
Status: NEW => RESOLVED

Comment 5 sturmvogel 2022-01-22 14:28:38 CET

Additional proof that Adelson's matrix type is not the right one for a Cholesky decomposition: If you try a proper matrix type from the example page you will find that it works as intended with our Maxima package:
https://rosettacode.org/wiki/Cholesky_decomposition#Maxima

Comment 6 Adelson Oliveira 2022-01-22 14:36:24 CET

(In reply to sturmvogel from comment #3)
> Same error occurs also on Opensuse Tumbleweed with wxMaxima 21.11.0 and
> maxima 5.45.1
> 
> So no Mageia problem. Either upstream or wrong matrix type as the cholesky
> function needs spezial types of matrix.
> 
> 
> Bugreport against Mageia version is invalid!

Testing matrix is OK as can be seen if one test it on softwares like scilab, octave, etc. Also in maxima 5.23, in CENTOS 6, runs OK. My best bet is that some of the packages maxima depends on has been changed and it caused the problem. In fact, I haven't had this problem up to a couple of weeks ago. It seems I have to report this to maxima developers. 

Thanks a lot for the return

Comment 7 Adelson Oliveira 2022-01-22 14:41:31 CET

The requirement for using cholesky is that the matrix is hermitian positive definite. This matrix fill this requirements.

Comment 8 Morgan Leijström 2022-01-22 18:35:44 CET

You seem to know what you are talking about :)

( I don't... ;) )

Yes please file a bug upstream and link to it from here.

Setting to wontfix for now.

Please reopen when a fixed version upstream is released, or there is another fix (whatever it depends on)

Keywords: (none) => UPSTREAM
Resolution: INVALID => WONTFIX