[MLton-devel] MLton and profiling

[Stephen Weeks]

> I would say that I would definitely want different occurances (due to
> duplication and inlining) of the same function counted together, not
> separately.  As you say, it really is an implementation detail.  I.e.,
> I like the old way better.

I knew somebody would want it.  I will put -coalesce {false|true} in
today.  I think you may change your mind though after playing with
larger programs where there are tens or even hundreds of copies of a
function (e.g. fold). 

In any case, the nice thing about doing the coalescing in the mlprof
postprocessing rather than in the compiler backend, is that things
like -ignore work better.  For example, you can -ignore fold without
it looking like every caller of fold merges together.


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel