The nomalloccheck USE flag was removed from glibc-2.3.5.ebuild (at CVS revision 1.27) and newer without a reason provided. While the default of glibc may be good to catch a limited number of bugs in programs (valgrind detection is far superior) and have them reported by users, the fact it kills processes by default (apart from grabbing a users attention) results in the loss of time and data. Anybody who plays commercial closed source games (which includes me as an occassional player) for which a fix is unlikely, is likely to be using nomalloccheck already. I am aware of using the MALLOC_CHECK_ environment variable and I can create a local glibc ebuild with the patch enabled. If the nomalloccheck USE flag is not going to be reintroduced to provide the user with a choice then the reason for the removal would be appreciated (especially as it is a simple one line patch). I would presume you are trying to force users to report errors?
sure, that's the reason
