I’ll raise the point at our Tues telecon and see who among them might have the cycles to address this as I’ll be on the road with limited time over the next week, and then we’ll be hitting the holidays.
> On Dec 13, 2014, at 11:12 AM, Eric Chamberland > <eric.chamberl...@giref.ulaval.ca> wrote: > > On 12/12/2014 01:12 PM, Ralph Castain wrote: >> I just checked it with —enable-memchecker —with-valgrind and found that many >> of these are legitimate leaks. We can take a look at them, though as I said, >> perhaps may wait for 1.8.5 as I wouldn’t hold up 1.8.4 for it. > > wait! > > When end-developpers of other software valgrind their code, they find leaks > from openmpi and then they ask themself: "Did I made a misuse of MPI?" So > they have to look around, into the FAQ, and find this: > > http://www.open-mpi.org/faq/?category=debugging#valgrind_clean > > and tell theme self: "Fine, now with this suppression file, I am sure the > leaks are my fault!" and try to find why theses leaks remains in their code... > > then, not understanding what is wrong... they ask the list to see if it is > normal or not... ;-) > > May I suggest to give suppression name like > "real_leak_to_be_fixed_in_next_release_#" so at least, you guys won't forget > to fix it, and all of us won't be upset about misuse of the library? > > Or maybe put them into another suppression file? But list them down > somewhere: that would really help us! > > Thanks, > > Eric > > ps: we valgrind our code each night to be able to detect asap new leaks or > defects... > _______________________________________________ > devel mailing list > de...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel > Link to this post: > http://www.open-mpi.org/community/lists/devel/2014/12/16585.php
