Re: [Bug-glpk] [Fwd: Bug in GLPK Graph]

[Chris Matrakidis]

Andrew,

 

Please note that there is a bug in the original RELAX IV Fortran code,
which I tried to fix with a brute force changes; see
https://lists.gnu.org/archive/html/bug-glpk/2013-04/msg00009.html

Some time ago I came across a different fix for the same bug:

https://github.com/jdleesmiller/relax4_ruby/blob/master/ext/relax4/relax4.c#L3811

This suggests a change in mcfrelax.c like this:

if (n > na)

   csa.prdcsr = xcalloc(1+n, sizeof(int));

else

   csa.prdcsr = xcalloc(1+na, sizeof(int)); 

or maybe a more conservative:

csa.prdcsr = xcalloc(1+n+na, sizeof(int));

Best Regards,

Chris Matrakidis