Otherwise we would get this error in C++11 mode (because of a recent change):
error: non-type template argument of type 'const _GUID *' is not a constant expression
For code like:
template <const GUID* g = &__uuidof(struct_with_uuid)>
class COM_CLASS { };
llvm-svn: 154790
includes a patch from Matthias Kleine with a regression testcase!
Adds a new iterator 'data_iterator' to OnDiskHashTable which doesn't try to
reconstruct the external_key from the internal_key, which is useful for traits
that don't store enough information to do that mapping in their key. Also
deletes the 'item_iterator' from OnDiskHashTable as dead code.
llvm-svn: 154784
This is a complex change that resulted from a great deal of
experimentation with several different benchmarks. The one which proved
the most useful is included as a test case, but I don't know that it
captures all of the relevant changes, as I didn't have specific
regression tests for each, they were more the result of reasoning about
what the old algorithm would possibly do wrong. I'm also failing at the
moment to craft more targeted regression tests for these changes, if
anyone has ideas, it would be welcome.
The first big thing broken with the old algorithm is the idea that we
can take a basic block which has a loop-exiting successor and a looping
successor and use the looping successor as the layout top in order to
get that particular block to be the bottom of the loop after layout.
This happens to work in many cases, but not in all.
The second big thing broken was that we didn't try to select the exit
which fell into the nearest enclosing loop (to which we exit at all). As
a consequence, even if the rotation worked perfectly, it would result in
one of two bad layouts. Either the bottom of the loop would get
fallthrough, skipping across a nearer enclosing loop and thereby making
it discontiguous, or it would be forced to take an explicit jump over
the nearest enclosing loop to earch its successor. The point of the
rotation is to get fallthrough, so we need it to fallthrough to the
nearest loop it can.
The fix to the first issue is to actually layout the loop from the loop
header, and then rotate the loop such that the correct exiting edge can
be a fallthrough edge. This is actually much easier than I anticipated
because we can handle all the hard parts of finding a viable rotation
before we do the layout. We just store that, and then rotate after
layout is finished. No inner loops get split across the post-rotation
backedge because we check for them when selecting the rotation.
That fix exposed a latent problem with our exitting block selection --
we should allow the backedge to point into the middle of some inner-loop
chain as there is no real penalty to it, the whole point is that it
*won't* be a fallthrough edge. This may have blocked the rotation at all
in some cases, I have no idea and no test case as I've never seen it in
practice, it was just noticed by inspection.
Finally, all of these fixes, and studying the loops they produce,
highlighted another problem: in rotating loops like this, we sometimes
fail to align the destination of these backwards jumping edges. Fix this
by actually walking the backwards edges rather than relying on loopinfo.
This fixes regressions on heapsort if block placement is enabled as well
as lots of other cases where the previous logic would introduce an
abundance of unnecessary branches into the execution.
llvm-svn: 154783
Passing -verify to clang without -cc1 or -Xclang silently passes (with a
printed warning, but lit doesn't care about that). This change adds -cc1 or,
as is necessary in one case, -Xclang to fix this so that these tests are
actually verifying as intended.
I'd like to change the driver so this kind of mistake could not be made, but
I'm not entirely sure how. Further, since the driver only warns about unknown
flags in general, we could have similar bugs with a misspellings of arguments
that would be nice to find.
llvm-svn: 154776
As an example, attach range info to the "invalid instruction" message:
$ clang -arch arm -c asm.c
asm.c:2:11: error: invalid instruction
__asm__("foo r0");
^
<inline asm>:1:2: note: instantiated into assembly here
foo r0
^~~
llvm-svn: 154765
initialize an array of unsigned char. Outside C++11 mode, this bug was benign,
and just resulted in us emitting a constant which was double the required
length, padded with 0s. In C++11, it resulted in us generating an array whose
first element was something like i8 ptrtoint ([n x i8]* @str to i8).
llvm-svn: 154756
thinking of generalizing it to be able to specify other freedoms beyond accuracy
(such as that NaN's don't have to be respected). I'd like the 3.1 release (the
first one with this metadata) to have the more generic name already rather than
having to auto-upgrade it in 3.2.
llvm-svn: 154745
thinking of generalizing it to be able to specify other freedoms beyond accuracy
(such as that NaN's don't have to be respected). I'd like the 3.1 release (the
first one with this metadata) to have the more generic name already rather than
having to auto-upgrade it in 3.2.
llvm-svn: 154744
Instead, make it the allocation function's responsibility to add them
to a list and clear it when a top-level decl is finished.
This plugs leakage of TemplateAnnotationIds. DelayedCleanupPool is
ugly and unused, remove it.
llvm-svn: 154743