[C++ Patch] PR 42055

[Paolo Carlini]

Jason Merrill wrote:
> If it's ambiguous, surely there are more than 0 candidates.  Perhaps
> we need to assign the return value of chainon to candidates?
You are absolutely right. Maybe I got confused because, as far as I can
see, when we didn't have this regression, we didn't output the
candidates either (neither does EDG, FWIW). Anyway, updated, regtested
patch below. Ok?

Paolo.

//////////////////
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: CL_42055_2
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20091116/1d8c6d04/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: patch_42055_2
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20091116/1d8c6d04/attachment-0001.ksh>