[Coin-discuss] MS STL UMRs, Purify

Matthew Galati Matthew.Galati at sas.com
Sun Apr 8 10:11:40 EDT 2007


Laci and I have been looking at a CoinUtils ticket:
   https://projects.coin-or.org/CoinUtils/ticket/28

"
This is a valid UMR. The PurifyPlus team has investigated this before, so there's no need to provide us with a testcase. There is an error in the MS STL implementation where they are referencing uninitialized memory. Very unfortunate, but true none the less. The best way to deal with this at present is to create a filter(s) for this error.
"

I decided to send this to coin-discuss, because my googling turned up a solid explanation - and, if you use Purify, you'll probably want to read this:
   http://www-128.ibm.com/developerworks/forums/dw_thread.jsp?message=13852497&cat=24&thread=79543&treeDisplayType=threadmode1&forum=344#13852497
   http://www-128.ibm.com/developerworks/forums/dw_thread.jsp?message=13919960&cat=24&thread=79543&treeDisplayType=threadmode1&forum=344#13919960

Matt




More information about the Coin-discuss mailing list