[PATCH] [dwarfdump] Prettyprint DW_AT_APPLE_property_attribute bitfield values.

David Blaikie dblaikie at gmail.com
Wed Oct 8 17:27:13 PDT 2014

Probably sensible enough then to just make it a dumping test - I think
that's generally 'better'/purist when adding dumper features, I just
understand it's sometimes a bit more fuss. So I'll never complain about you
doing it that way (checking a small sample source, a binary, etc - you can
see the other pure dwarfdump tests).

& since you guys are planning to make this more of a production tool, I
think it's a good idea to have more tests for dwarfdump itself - so that we
don't accidentally make holes in the dwarfdump test coverage when we
refactor tests.

On Wed, Oct 8, 2014 at 5:23 PM, Frederic Riss <friss at apple.com> wrote:

> >>! In D5689#4, @dblaikie wrote:
> > Looks reasonable (apart from testing) - depends how you want to do it. I
> assume there are some existing tests that test properties that could be
> updated to use this feature? (does this not regress any existing tests?)
> >
> > If there aren't any such tests, I imagine we want some - and they could
> just test this incidentally.
> >
> > (I'm usually on the fence about writing tool tests separately from unit
> tests... *shrug* sometimes it's nice, sometimes it seems like overkill)
> There aren't any. I discovered the discrepancy between the enum values
> yesterday by trying to produce some tests and the results didn't actually
> make sense. I can add an llvm test, sure, but my ObjC foo being what it is,
> I fear that the code won't make any sense :-) Thus it won't test much
> except the actual dumping.
> http://reviews.llvm.org/D5689
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20141008/0e154807/attachment.html>

More information about the llvm-commits mailing list