Re: [Patch] mt_allocator: spare mem & fix alignment problems

[Paolo Carlini <pcarlini at suse dot de>]

Benjamin Kosnik wrote:

> ...rope certainly seems implicated. It is strange, however, that this
> only fails with __mt_alloc and not with new_allocator or __pool_alloc.
>  
Yes, puzzling. Something I still don't fully understand about 
mt_allocator is
the logic in deallocate: for instance, from the comments you would believe
that remove is always > 0. Well, if one put there a stupid printf and run
allocator/insert.cc learns that for B7-B17-B22 THREAD and for
T7-T17-T22 THREAD it becomes as small as -127000 !?!?!

> We need to deprecate rope. We have no evidence that it's being used, and
> it's the sole user of other crusty files. People who are curious about
> it can just dig through the archives.
>
> I'll propose a patch to do so immediately after 3.4.0.
>  
Too bad, I would say. Because after all I agree that algorithmically 
there is
something very interesting in it. I learned only a few days ago that 
Hans Boehm,
yes *that* Hans Boehm, created it around 10 years ago...

Paolo.