[PATCH] D49095: [COFF] Store a pointer to the parent instead of just the parent name in InputFile

[Zachary Turner via Phabricator via llvm-commits]

zturner added a comment.

In https://reviews.llvm.org/D49095#1156542, @ruiu wrote:

> The problem of having an archive file pointer instead of an archive file name is that we don't (and don't have to) instantiate an archive file object for an archive given as an argument for /wholearchive. How do you deal with that?


It seems like I don't deal with that currently, however all the tests pass.  So either that means when an argument is passed to `/wholearchive`, we never end up needing the archive name anyway, or it means that we're lacking test coverage somewhere.  Do you know which it is?  If it's the former, then I guess I don't need to do anything.  If it's the latter, then is there any reason we *can't* instantiate an `ArchiveFile` object?  The expensive thing is parsing the archive, which we have to do regardless.


https://reviews.llvm.org/D49095