[PATCH] D51174: Move Microsoft Demangler to Support

[Zachary Turner via Phabricator via llvm-commits]

zturner added a comment.

In https://reviews.llvm.org/D51174#1211249, @rnk wrote:

> I guess I don't like this change because it makes our demanglers, which should ideally be super standalone, no STL usage code, more entangled with the rest of LLVM. It would be nice if, for example, we could compile this code into sanitizer libraries. Today we have a solution (llvm-symbolizer) that essentially pushes the symbolization work out of process into LLVM code, but you can imagine wanting to have a complete in-library solution.


I don't know, we don't have these restrictions anywhere else in LLVM.  Lots of things would be nice I guess, but unless anyone needs it today, I'd prefer to go with YAGNI and if we ever do, then we can do something like what Richard did with the Itanium demangler.


https://reviews.llvm.org/D51174