After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 611407 - GLPK solver error message is uninformative
GLPK solver error message is uninformative
Status: RESOLVED FIXED
Product: Gnumeric
Classification: Applications
Component: Analytics
git master
Other Linux
: Normal enhancement
: ---
Assigned To: Morten Welinder
Jody Goldberg
Depends on:
Blocks:
 
 
Reported: 2010-02-28 15:10 UTC by Konstantin Tokarev
Modified: 2010-03-03 00:07 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Sample optimization problem (2.10 KB, application/x-gnumeric)
2010-03-01 21:38 UTC, Konstantin Tokarev
Details

Description Konstantin Tokarev 2010-02-28 15:10:26 UTC
Error is shown if I try to run solver using GLPK. GLPK (glpsolve, libgplk0, development packages) is installed
Comment 1 Morten Welinder 2010-02-28 16:03:21 UTC
What is the precise error?

Do you have a glpsol binary on your system?
Comment 2 Konstantin Tokarev 2010-02-28 17:11:52 UTC
Yes,I have
Comment 3 Morten Welinder 2010-02-28 20:53:04 UTC
> What is the precise error?
Comment 4 Konstantin Tokarev 2010-03-01 16:05:00 UTC
Sorry, I've found that GLPK works. Probably error was caused by non-linearity of problem. But error message was quite uninformative: "Time 0:00. Error". Maybe you can show output of glpsol or interpret it's return code?
Comment 5 Morten Welinder 2010-03-01 19:13:08 UTC
Closing.

If you find a sample where we are that uninformative, please refile.
Comment 6 Konstantin Tokarev 2010-03-01 19:23:13 UTC
Is "Time 0:00. Error" informative?
It would be great if some information was given, e.g. "Problem is non-linear"
Comment 7 Andreas J. Guelzow 2010-03-01 20:37:02 UTC
Is suspect that that may be the only info we get from GLPK. (And anybody will agree that it is quite uninformative.) Would you be able to show us that problem?
Comment 8 Konstantin Tokarev 2010-03-01 21:38:03 UTC
Created attachment 154977 [details]
Sample optimization problem
Comment 9 Konstantin Tokarev 2010-03-02 10:48:16 UTC
Here it is
Comment 10 Morten Welinder 2010-03-02 23:21:09 UTC
welinder@anemone:~/gnome/gnumeric/src> GNM_DEBUG=solver ./gnumeric ~/Download/problem.gnumeric 
Registering glpk
Registering lpsolve
Spawning glpsol --noscale --write /tmp/program-IQEH8U.out --cpxlp /tmp/program-27QH8U.cplex
glp_read_lp: reading problem data from `/tmp/program-27QH8U.cplex'...
glp_read_lp: 4 rows, 4 columns, 4 non-zeros
glp_read_lp: 12 lines were read
glp_simplex: original LP has 4 rows, 4 columns, 4 non-zeros
PROBLEM HAS NO DUAL FEASIBLE SOLUTION
Time used:   0.0 secs
Memory used: 0.0 Mb (43380 bytes)
glp_write_sol: writing basic solution to `/tmp/program-IQEH8U.out'...
glp_write_sol: 10 lines were written
Solver process exited with code 0
Unregistering glpk
Unregistering lpsolve
Comment 11 Morten Welinder 2010-03-02 23:58:28 UTC
And here's the solution file written:

4 4
1 1 0
1 0 0
1 0 0
1 0 0
1 0 0
2 0 0
2 0 0
2 0 0
2 0 0

The second line is key here: "1" and "1" mean that both the primal and the
dual have *undefined* status.  This is wrong, but I guess we should work
around it.
Comment 12 Morten Welinder 2010-03-03 00:07:41 UTC
This problem has been fixed in our software repository. The fix will go into the next software release. Thank you for your bug report.

(You now get "infeasible")