[Coin-discuss] Revision of GLPK

[Brady Hunsaker]

On Tue, 2004-07-27 at 18:18, Joseph Young wrote:
> 	I updated OsiGlpkSolverInterface.cpp to be compatible with the new 
> API used in v4.4 and 4.5.  However, there are a few odd quirks:
> 

Joseph,

Thanks for doing this!  If you are interested in sharing your code under
the Common Public License, then we can include it directly in COIN-OR. 
I'll follow up off the list with some more details.

Brady

> In unitTest (OsiGlpkSolverInterfaceTest: 495), assert(eq(ev[0],3.0)) was 
> failing.  In appeared as though getMatrixByRow was reading in the rows in 
> reverse order.  This should not have happened because this function is 
> correctly reading in the elements and indices from GLPK's data structures.  
> So, it looks like there may be a problem with the copy routines elsewhere.  
> Right now, I have a hack to fix this, but someone might want to look at it 
> further.
> 
> In unitTest (OsiSolverInterfaceTest: 3275), is segfaulting for unknown 
> reasons on OsiSolverInterface *s=emptySi->clone().
> 
> 	Anyway, you can play with the new version if you'd like.  Since 
> OsiGlpkSolverInterface.cpp is around 2500 lines, it's probably not prudent 
> to post it here, so would anyone like a copy?
> 
> Sinc, Joseph Young
> 
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss