[Coin-discuss] BCP error?

Laszlo Ladanyi ladanyi at us.ibm.com
Tue Aug 5 00:53:29 EDT 2003


Good catch!

The correct fix is that if new_indexed_var cuts off any dual rays then the
pointer is pushed onto vars_to_add and new_indexed_var needs to be zeroed out
(this last part was missing). I have committed the fix. Please, check it out,
take a look at it and try it.

Thanks,
--Laci

On Mon, 4 Aug 2003, Jörg Herbers wrote:

> I encountered a problem in BCP_lp_fathom.cpp: If 
> MaintainIndexedVarPricingList is set to true, BCP_restore_feasibility tests 
> the indexed variables (provided by BCP_lp_user::next_indexed_var and 
> BCP_lp_user::create_indexed_var) for a positive product with the dual rays. 
> If any of the products is <0 (line 482), the newly created variable is 
> inserted into vars_to_add. Nevertheless, the new variable (new_indexed_var) 
> will always be deleted afterwards (line 476 or 494) regardless of whether it 
> was inserted or not. In my implementation, this results in a crash. If I 
> understand this part of the code correctly, the variable shouldn't be 
> deleted if it was inserted (since it will be used later on, e.g. in line 
> 118).
> 
> Thanks for any help.
> 
> Jörg
> 
> _________________________________________________________________
> Help STOP SPAM with the new MSN 8 and get 2 months FREE*  
> http://join.msn.com/?page=features/junkmail
> 
> _______________________________________________
> Coin-discuss mailing list
> Coin-discuss at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss
> 




More information about the Coin-discuss mailing list