Re: Elisp native profiler

[Eli Zaretskii]

> From: Stefan Monnier <address@hidden>
> Cc: address@hidden
> Date: Tue, 25 Sep 2012 21:00:59 -0400
> 
> >>>>> "Eli" == Eli Zaretskii <address@hidden> writes:
> > If you want the profiler to actually work in the w32 build, additional
> > code needs to be written, to replace setitimer and the SIGPROF signal,
> > which aren't supported on Windows.
> 
> Actually, even without the timer you can still use the memory profiler
> where the samples are taken when malloc is called (more or less).

Yes, I was talking about the "cpu" profiler.

> It would be nice to add support for the cpu-profiler, but for now I'm
> focusing on getting the branch in a good enough state to be merged into
> trunk, so "fully working on Windows" is not nearly as important as
> "doesn't break the Window build".

Will the merge happen before the feature freeze?  If not, the
adaptation of itimers to MS-Windows can wait.  Otherwise, it doesn't
make sense to me to have an interim release with an important new
feature that is not supported on MS-Windows.