C++ PATCH for c++/36963

[Jason Merrill]

PR 36963 points out a problem with the specification of narrowing 
conversion diagnostics in the initializer_list proposal: it affects 
previously well-formed code.  For most cases this is unlikely to affect 
anyone, but for initialization of a float from an undecorated 
floating-point constant (with implicit double type) it seems likely to 
cause problems.

I expect this to be addressed at the next meeting, but for the time 
being I've applied this fix: if the source is a literal, don't complain.

Tested x86_64-pc-linux-gnu, applied to trunk.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 36963.patch
Type: text/x-patch
Size: 2090 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20080804/8e97d700/attachment.bin>