[phobos] core.demangle replacing std.demangle?

[Denis]

On Fri, Sep 10, 2010 at 8:37 PM, David Simcha <dsimcha at gmail.com> wrote:
> I tried to use core.demangle to print human-readable profiling reports and
> it still seems to fail horribly on some symbols.  The un-demanglable symbols
> seem to  be really complicated template instantiations, like
> randomShuffle!(chain!(... .  IIRC there's some fundamental limitation where
> once a symbol gets huge DMD relies on hashing, so the mangling no longer has
> an inverse.  Is this true, or is it worth spending some time trying to
> create a decent test case?
>

The following issue I reported over a year ago might help:
http://d.puremagic.com/issues/show_bug.cgi?id=2238

There is also a small test case.