[cfe-dev] __is_empty type trait patch

[Eli Friedman]

On Sun, Aug 2, 2009 at 7:45 PM, Sean Hunt<rideau3 at gmail.com> wrote:
> Hello,
>
> I'm a new coder hoping to help out with the C++ implementation on clang.
> After fiddling around with several things I discovered were all being
> worked on already, someone recommended I implement the __is_empty type
> trait, so that is what I did. I was told to send a message here now;
> I've attached a patch.

Cool.

> A few notes about this:
>
> I put tests in test/SemaCXX/type-traits.cpp. There are a bunch of tests
> I added to __is_pod and __is_empty, mainly for testing purposes. I was
> going to remove them, but then I figured there's no sense removing tests
> - more is better.

Yeah, having a bunch of tests is good.  It's generally not a good idea
to put in commented-out tests, though; if it gives an error, use an
expected-error marking, and put a comment that the error isn't
correct.  (If it crashes, of course, it's better to leave it commented
out.)

> I also encountered, while doing checks for zero-width bit-fields, an
> error "bit-field i has zero width". I added the word "named" to the
> diagnostic so that it's clear what the error really is. This change is
> unrelated but is minor so I included it anyways.

I committed this separately.

 bool Sema::VerifyBitField(SourceLocation FieldLoc, IdentifierInfo *FieldName,
-                          QualType FieldTy, const Expr *BitWidth) {
-
+                          QualType FieldTy, const Expr *BitWidth,
bool *ZeroWidth) {
+  // Default to true; that shouldn't confuse checks for emptiness
+  if (ZeroWidth)
+    *ZeroWidth = true;

This looks suspicious; is it really supposed to default to true?  (At
first glance, it looks like it will misclassify a struct with only
bitfields.)

Otherwise, looks fine.

-Eli