Re: Memory again

[Carsten Mattner <carstenmattner@googlemail.com>]

On Sat, Nov 26, 2011 at 3:35 PM, Dmitry Antipov <antipov@mvista.com> wrote:
> On 11/26/2011 05:26 PM, Carsten Mattner wrote:
>
>> After having used 4 erlang-mode buffers and having killed all buffers
>> for more than an hour and having run M-x garbage-collect there still
>> seems to be unreclaimable space.
>
> This is an internal heap fragmentation, the most common disadvantage
> of simple mark and sweep GC.

Allow me to be naive and ask this: what about compaction?
Would the allocator in use fragment the overall system memory
space if it frees more aggressively?
Does emacs fear the memory may be used by someone else
and hold onto what it claimed? Half-joking, half-serious asking.

>> It's been sitting there with 0 buffers in *scratch* at around 54megs
>> for a couple hours.
>
> Try to restore your 4 erlang-mode buffers and other stuff which has been
> previously used to reach 54M. If the heap will not grow substantially
> beyond 54M, this means that fragmented space was mostly re-used.
> If you ends up with 2x or more heap, there is a serious memory leak
> somewhere.

Nothing like that happened, but I had to restore some of the files as
it was from a temporary directory. Restoring the directory made emacs
realize the "file has changed on disk". Is this checked via a hash/checksum
or does emacs carry along more than that?

Details: before I opened the same buffers as good as I remembered it
was at 55megs up from 54.2megs (approx.) and it went up .4megs.