Ondřej Vašík <[email protected]> wrote: > Jim Meyering wrote: >> Unfortunately, with this patch, "make check" fails the cp/link-heap test. >> Investigating why, I discovered that it introduces a huge leak into >> cp -a, at least on F10. 80MB worth in this case. >> >> I'll let you investigate that one ;-) > > Leak is in libattr - added by 2.4.43 (~8 months ago) - reported as > RHBZ #485473 . Kamil Dudka already proposed patch for it. Leak doesn't > occur for older libattr or libattrs with that patch.
Thanks to both of you for tracking it down. That looks fine, now, except for the F10 test failure due to the leak. When will the libattr fix be in F10? That is what will determine whether I can accept this patch before the upcoming stable release. I cannot accept a change that will make "cp -a" leak so much on F10. If a libattr fix won't make it to F10 this week (early next week might be ok, too), then we'll just defer this change-set until after coreutils-7.1. If you think a libattr fix *will* make it to F10 this week, then please ensure that no other major distribution is affected. _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
