[PATCH] Fix warnings found by gcc-4.6

Most of these were found by the new -Wunused-but-set-variable warning. Several of these unused variables look like they ought to have been used, so someone who knows this code better than I do should check whether the new warnings found some bugs.

I also fixed some spurious -Wint-to-pointer-cast warnings.

I checked that everything was fixed with `make CFLAGS="-Werror" CXXFLAGS="-Werror"`

[Jeffrey Yasskin, 2011-01-07 00:53:36]

The full patch is at http://codereview.appspot.com/download/issue3888041_1.diff.
Please take a look.

[DougGregor, 2011-01-17 19:32:54]

Looks great, please commit!

http://codereview.appspot.com/3888041/diff/1/lib/Lex/PPMacroExpansion.cpp
File lib/Lex/PPMacroExpansion.cpp (right):

http://codereview.appspot.com/3888041/diff/1/lib/Lex/PPMacroExpansion.cpp#new...
lib/Lex/PPMacroExpansion.cpp:865: assert(IsValid && "invalid __has_include");
(void)IsValid;
FWIW, Chris has already cleaned up this bit.

http://codereview.appspot.com/3888041/diff/1/lib/Parse/ParseObjc.cpp
File lib/Parse/ParseObjc.cpp (left):

http://codereview.appspot.com/3888041/diff/1/lib/Parse/ParseObjc.cpp#oldcode149
lib/Parse/ParseObjc.cpp:149: SourceLocation lparenLoc = ConsumeParen();
The fact that we're not using lparenLoc means that we're providing poor recovery
here (e.g., by not using MatchRHSPunctuation). It's not your problem to fix, but
thanks for pointing it out!

http://codereview.appspot.com/3888041/diff/1/lib/Sema/SemaDeclCXX.cpp
File lib/Sema/SemaDeclCXX.cpp (right):

http://codereview.appspot.com/3888041/diff/1/lib/Sema/SemaDeclCXX.cpp#newcode...
lib/Sema/SemaDeclCXX.cpp:4930: if (!BaseType->getAs<RecordType>()) {
How about 

  if (!BaseType->isRecordType())

which is slightly more efficient?

[Jeffrey Yasskin (google), 2011-01-18 02:14:35]

Thanks, committed. I switched to isRecordType and gave you a TODO over
the lparenLoc.

On Mon, Jan 17, 2011 at 11:32 AM,  <doug.gregor@gmail.com> wrote:
> Looks great, please commit!
>
>
> http://codereview.appspot.com/3888041/diff/1/lib/Lex/PPMacroExpansion.cpp
> File lib/Lex/PPMacroExpansion.cpp (right):
>
>
http://codereview.appspot.com/3888041/diff/1/lib/Lex/PPMacroExpansion.cpp#new...
> lib/Lex/PPMacroExpansion.cpp:865: assert(IsValid && "invalid
> __has_include"); (void)IsValid;
> FWIW, Chris has already cleaned up this bit.
>
> http://codereview.appspot.com/3888041/diff/1/lib/Parse/ParseObjc.cpp
> File lib/Parse/ParseObjc.cpp (left):
>
>
http://codereview.appspot.com/3888041/diff/1/lib/Parse/ParseObjc.cpp#oldcode149
> lib/Parse/ParseObjc.cpp:149: SourceLocation lparenLoc = ConsumeParen();
> The fact that we're not using lparenLoc means that we're providing poor
> recovery here (e.g., by not using MatchRHSPunctuation). It's not your
> problem to fix, but thanks for pointing it out!
>
> http://codereview.appspot.com/3888041/diff/1/lib/Sema/SemaDeclCXX.cpp
> File lib/Sema/SemaDeclCXX.cpp (right):
>
>
http://codereview.appspot.com/3888041/diff/1/lib/Sema/SemaDeclCXX.cpp#newcode...
> lib/Sema/SemaDeclCXX.cpp:4930: if (!BaseType->getAs<RecordType>()) {
> How about
>
>   if (!BaseType->isRecordType())
>
> which is slightly more efficient?
>
> http://codereview.appspot.com/3888041/
> _______________________________________________
> cfe-commits mailing list
> cfe-commits@cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>