On Fri, Nov 2, 2012 at 8:52 PM, Brian Brooks brooks.br...@gmail.com wrote:
I'm not sure I follow how this comment relates to the patch.
I see that a CXXRecordDecl is created during template instantiation via
RequireCompleteType(). The abort is triggered by dyn_cast to a
CXXRecordDecl before
Ping!
On Fri, Nov 2, 2012 at 11:52 PM, Brian Brooks brooks.br...@gmail.comwrote:
I'm not sure I follow how this comment relates to the patch.
I see that a CXXRecordDecl is created during template instantiation via
RequireCompleteType(). The abort is triggered by dyn_cast to a
Adding cfe-dev...
On Wed, Oct 31, 2012 at 5:26 PM, Brian Brooks brooks.br...@gmail.comwrote:
The attached patch needs review please!
I suspect that we may need to call RequireCompleteType() to do
instantiation *even if* the VDecl is a dependent type. I am not yet able
to reason about this
On Thu, Nov 1, 2012 at 5:36 AM, Brian Brooks brooks.br...@gmail.com wrote:
Adding cfe-dev...
Dropping cfe-dev via bcc, you were right first time.
On Wed, Oct 31, 2012 at 5:26 PM, Brian Brooks brooks.br...@gmail.com
wrote:
The attached patch needs review please!
I suspect that we may need
The attached patch needs review please!
I suspect that we may need to call RequireCompleteType() to do
instantiation *even if* the VDecl is a dependent type. I am not yet able
to reason about this though.
The patch prevents the abort, and the test case (extended with code to
print initialized