PR16752[PATCH]: 'mode' attribute for unusual targets doesn't work properly.

[Stepan Dyatkovskiy]

Hi Eli,
Tried to commit the second patch, but some buildbots did like it. It 
only works if we add one more branch to getRealTypeByWidth:

diff --git a/lib/Basic/TargetInfo.cpp b/lib/Basic/TargetInfo.cpp
index 38cb411..bf6acd2 100644
--- a/lib/Basic/TargetInfo.cpp
+++ b/lib/Basic/TargetInfo.cpp
@@ -174,7 +174,10 @@ TargetInfo::RealType 
TargetInfo::getRealTypeByWidth(unsigned BitWidth) const {

    switch (BitWidth) {
    case 96:
-    if (&getLongDoubleFormat() == &llvm::APFloat::x87DoubleExtended)
+    if (&getLongDoubleFormat() == &llvm::APFloat::x87DoubleExtended ||
+        // Due to test/Sema/attr-mode.c test it seems that
+        // PPC machines should use PPCDoubleDouble type for XC mode.
+        &getLongDoubleFormat() == &llvm::APFloat::PPCDoubleDouble)
        return LongDouble;
      break;
    case 128:

If its OK, I would like to commit patch and this change.

-Stepan.

Eli Friedman wrote:
> LGTM.
>
> -Eli
>
>
> On Mon, Sep 9, 2013 at 5:04 AM, Stepan Dyatkovskiy <stpworld at narod.ru
> <mailto:stpworld at narod.ru>> wrote:
>
>     Since I had split the patch and committed only ASTContext and
>     TargetInfo getIntTypeByWidth and friends, there is the second patch,
>     that fixes PR16752 itself.
>
>     Please, find it in attachment for review.
>
>     -Stepan.
>     Eli Friedman wrote:
>
>         On Tue, Sep 3, 2013 at 2:16 AM, Stepan Dyatkovskiy
>         <stpworld at narod.ru <mailto:stpworld at narod.ru>
>         <mailto:stpworld at narod.ru <mailto:stpworld at narod.ru>>> wrote:
>
>              Hi Eli,
>              Sorry for latency.
>              As you remember this patch should correct 'mode' attr
>              implementation. You proposed to use generic way of type
>         detection
>              for each target: just scan for target types and select one with
>              suitable width.
>
>              Unfortunately I can't commit patch with generic
>         implementation. I
>              got test failure for clang. And I expect more failures for
>         another
>              targets. The reason is next.
>
>              The target could still keep mode unsupported even if it has
>         type of
>              suitable width. I mean the next.
>              For example this string should cause error for i686
>         machines (128
>              bit float):
>              typedef          float f128ibm __attribute__ ((mode (TF)));
>              But in case of generic approach for all targets it would be
>         passed.
>              In this case virtual functions allows to implement exceptions.
>
>
>         The generic implementation is still essentially correct; the the
>         issue
>         is that getLongDoubleWidth() isn't actually the right value to
>         check.
>            It returns "sizeof(long double) * 8", not the actual width of the
>         underlying float format.  You should be able to work around this by
>         checking getLongDoubleFormat() instead of getLongDoubleWidth().
>
>              In this case 'getIntTypeByWidth' may be a bit confusing name.
>              Instead it is supposed to return type for some particular
>         mode, not
>              by width. Something like getInt/RealTypeForMode(width, sign).
>              Though, currently I kept the original name.
>
>
>         If you want to change it, that's fine.
>
>         -Eli
>
>
>