Re: [PATCH] Add UNARY_PLUS_EXPR to C++ front-end (take 2)

[Mark Mitchell <mark at codesourcery dot com>]

Gabriel Dos Reis wrote:
> Roger Sayle <roger@eyesopen.com> writes:
>
> | Many thanks to both Mark and Giovanni for suggesting that appropriate
> | fix was to fold the new UNARY_PLUS_EXPR in fold_if_not_in_template.
> | This indeed fixes the g++ regression.  I also took the opportunity
> | to go through and audit all of the uses of CONVERT_EXPR in the cp/
> | directory.  I believe I've now caught all the relevant cases and the
> | 12 remaining uses of CONVERT_EXPR appear to treat it as a synonym
> | for NOP_EXPR.
> | 
> | The following revised patch has been tested on x86_64-unknown-linux-gnu
> | with a full "make bootstrap", all default languages, and regression
> | tested with a top-level "make -k check" with no new failures.
> | 
> | Ok for mainline?
>
> Sorry for not having commented earlier, but I think you need to add
> the pretty printing of UNARY_PLUS_EXPR in cxx-pretty-print.c too.

Yes, this is OK.

There are two other things that we need to add to the C++ front end:

1. A folder for constant-expressions.  This function will be called by 
the front-end when an integral constant-expression appars in some 
situation requiring an actual integer constant, like an expression for 
array bounds.

2. A recursive folder for expressions.  This function will be used by 
the gimplifier.  Once we have it, we will stop calling 
fold/fold_if_not_in_template throughout the rest of the compiler.

I've got a partial patch for (1) that I'll try to finish up soon; then, 
I hope to work on (2) as well.

--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
(916) 791-8304