[Nauty-list] readgraph()

William Rummler w.a.rummler at gmail.com
Sat Feb 12 12:57:09 EST 2011


I could not reproduce the long running time you mentioned with the
latest source and input, but I have a feeling that it might have
something to do with a scanf() that was waiting for input. (On the
other hand, that is a total guess and might be completely wrong.)

I did not notice any statements for printing the "testX" output in
this source file (testnau3.c). Is this the right file (or similar
enough to the right one)? If I recall correctly, your testnau4.c had
some print statements for "testX" output.

I quickly went over this one and found only one major issue, which was
that the variable l is not initialized before its use in the malloc()
calls for lab, ptn, and orbits. Those malloc() calls should probably
be moved to just after the sequence of fscanf() calls for anz, h, and
l.

Since there was no corresponding nauty-format input given for
jbc050_5, I used your code to generate the corresponding "jba"
nauty-format file that would be used. I then re-ran your program with
both the given jbc050_5 and the generated jba050_5, and it seemed to
run correctly and produce correct output. (All five isomorphism tests
were positive, and the two files generated by the putgraph() code
sections were identical.)

Since you mention that you traced the problem to the nauty() call, I
think I should ask if you sent the right file(s)?

- William

On Fri, Feb 11, 2011 at 12:02 PM,  <niess at ma.tum.de> wrote:
> Thank you very much for your help. Now the programs work - at least for
> small files. When I tried to run them with the slightly bigger file
> jbc050_5 however, they did neither stop nor produce a result. After adding
> more test outputs, I came to the conclusion that the place where the
> problem occurs (but probably not originates) is again the call to nauty. I
> will attach the modified testnau3.c and jbc050_5. It reached the output
> "test6" but even after running about 24 hours did not come to test7. Now I
> have not the slightest idea what I can do about that (except killing the
> program). Can you please help me again?




More information about the Nauty mailing list