[Bug libc/24271] tdestroy() doesn't release resources

[polazhinets.a at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=24271

--- Comment #8 from polazhinets.a at gmail dot com ---
(In reply to Florian Weimer from comment #7)
> (In reply to polazhinets.a from comment #6)
> > It doesn't matter. The reduced reproducer itself is enough to show the bug
> > in action. Let me make some changes so that there are even less reasons to
> > assert, that I'm blind. I've appended an extra zero to constant 0x100000 and
> > freed that unhappy arr (which is noticeable in system monitor), but it still
> > consumes more than 700mb until the whole process terminates immediately
> > after the return of getchar(). Why not just give up on trying to prove, that
> > I'm stupid? Well, here's a screenshot...
> 
> I'm sorry we are explaining this poorly.
> 
> You failed this issue as a tsearch/tdestroy bug.  We do not dispute that you
> are experiencing an issue.  However, as Andreas pointed out in comment 3,
> this is likely a duplicate of malloc bug 18504.  tsearch/tdestroy does not
> use its own allocator, it is layered on top of malloc/free internally.  This
> means that allocator problems can be visible via tsearch/tdestroy, as
> appears to be the case here.

It makes sense... tsearch/tdestroy tend to use malloc/free in somewhat tricky
way, revealing its buggy nature, which personally I didn't experience, when
using the latter directly in some trivial cases. Hadn't it been fixed yet?
Report, pointed by Andreas, seems like left unsatisfied.

-- 
You are receiving this mail because:
You are on the CC list for the bug.