Remove stray ellipsis treatment entirely

In D33339#759797, @rsmith wrote:

In D33339#759146, @hubert.reinterpretcast wrote:

The check-all target passes even if the ellipsis-after-declarator-id disambiguation as a declarator is removed entirely.

[...]

So, on the whole, the stray ellipsis treatment is both too complicated and not complicated enough.

I think if we want to keep it, the way to do that would be to carry on through the disambiguation process and treat it as a tiebreaker (that's what we do, for instance, if we see an undeclared identifier in a position where we're looking for a type). I'm not convinced that's worthwhile, especially since our existing testcases do not need this disambiguation rule, but perhaps we could remove the stray ellipsis treatment entirely for now and reconsider adding it back if we find poor diagnostics result from it later?

In D33339#759125, @rsmith wrote:

Should I assume our "misplaced ellipsis" diagnostic requires that we disambiguate the ill-formed ellipsis-after-declarator-id as a declarator in some cases? If so, do we have tests for that somewhere?

Tests for the diagnostics are in clang/test/FixIt/fixit-cxx0x.cpp and clang/test/Parser/cxx11-templates.cpp.
The check-all target passes even if the ellipsis-after-declarator-id disambiguation as a declarator is removed entirely. The disambiguation process is not needed involved in many cases, and when it is, can choose the declarative interpretation for other reasons.

Has it been discussed whether this is something to be addressed in the optimizer as opposed to the front-end?

Address review comments; update to revision 294580

Ping 2?

Ping!

Address review comments; update to revision 292996

Friendly ping...

In D25480#567552, @ahatanak wrote:

__builtin_fpclassify takes five int arguments followed by one last argument that is of floating point type. Do you know if there is a way to specify the last one argument is a floating point rather than using '.'?

There is already code in clang/lib/Sema/SemaChecking.cpp to generate diag::err_typecheck_call_invalid_unary_fp.

A test should probably be added as well.

In D24011#529508, @timshen wrote:

If there seems to be few applications are depending on this optimization, I wonder if we can turn off the constant folding for now, and fix APFloat later if someone really cares about it?

Just wanted to note that the front-end would still encounter the issue with APFloat, and I would be concerned if the (front-end) constant expression evaluation for ppc_fp128 is disabled entirely.

LGTM.

LGTM.

Make FixedSizeStorage accessible to test code

Make FixedSizeStorageOwner accessible to test code

Replace call to LLVM_CONSTEXPR function in constant expression context

Avoid MSVC C4099: replace 'struct' with 'class'

Address review comments: add tests, access control, Doxygen

In D22668#499164, @aaron.ballman wrote:

I don't suppose there's a way to test these changes, is there?

It's a utility class (which is not even used yet). I am not aware of testing for the ADTs, etc. aside from using them internally. Perhaps I should mark the change as NFC?

This looks straightforward. @mclow.lists, are you okay with me committing this for Jason?

Minor concerns noted. I'd be okay with this patch, but would like someone else to give the okay (or indicate what issues need to be addressed).

I believe tests are necessary.

This is superseded by D19322.

Reapply r273664 with workaround for MSVC

Update to r273650 with the requested const accessor

Ping.

Ping.

Ping (#2).

Set requires-clause when creating TemplateParameterLists; NFC

Ping.

@rsmith; I've addressed Faisal's comment. Please let me know if this patch (and D19770) is good to go. If it isn't ready yet, I'd like your opinion on D19770+D19771.

Address Faisal's comment; supercedes D19771

Adjust for changes to D19770

Responded to all comments.

Replace alias template with struct and typedef member; add usage example

The alias template hit a build-compiler bug: the substitution of the pack expansion of Counts does not work properly. I am investigating a replacement with a templated struct with a typedef member.

Use AlignOf<BaseTy> instead of AlignmentCalcHelper

An example of using FixedSizeStorage can be found in D19771.