Re: [PATCH] Fixed `scm_i_take_stringbufn ()'

[Ludovic Courtès]

Kevin Ryde <address@hidden> writes:

> I made some memory leak checks for myself by demanding that
> "(assoc-ref (gc-stats) 'bytes-malloced)" was steady across some
> repetitions of a test.  That's actually a different number though (is
> it?), but it should in theory be possible to exercise some of the
> malloc/free counting or balancing (post 1.8 that is :-).

I'm not sure doing so would be that easy because `bytes-malloced'
follows a periodic distribution whose period is unknown (i.e., each call
to `scm_take_locale_stringn ()' makes it slightly increase, and each
mark/sweep sequence -- which is occurs roughly periodically -- changes
it back to a much lower value).

Thanks,
Ludovic.