On 14.09.2019 13:54, Martin Husemann wrote: > On Sat, Sep 14, 2019 at 01:45:08PM +0200, Kamil Rytarowski wrote: >> OK, thank you. So to sum it up. Plugging resource leaks costs time and >> human effort and this is what is not wanted by some people. For those >> who find it useless are not forced to do it, so can ignore it. For >> others it can be useful activity. >> >> Thanks! I will go for __NO_LEAKS ifdef. > > But it is not a good idea to clutter perfectly fine sources with such > #ifdefs and unused/untested/likely broken code (even if it does not affect > the default binaries). > > Martin > > This is right. I don't intend to do this now/soon. I don't intend to do it before passing all the tests in LSan/ASan and before getting valgrind aboard. The most interesting target here is rumpkernel.
Attachment:
signature.asc
Description: OpenPGP digital signature