Re: [PATCH] Fix PR C++/28989, post-increment of bool is wrongly considered a lvalue

[Mark Mitchell <mark at codesourcery dot com>]

Andrew Pinski wrote:

> cp/ChangeLog:
> 	* tree.c (lvalue_p_1 <case SAVE_EXPR>): SAVE_EXPRs are never
> 	lvalues.

Are you sure this is safe?  What about something like "++i", which is an 
lvalue, but has side-effects, and might therefore be turned into a 
SAVE_EXPR at some point?  For example, in "new S[++i]", we might make a 
SAVE_EXPR for "++i" before we call "operator new[]", but then need the 
value later.  In this case, I don't think lvalueness would matter, 
though, so maybe we are still OK.

I like the idea of this patch, but I'm nervous that a case like this 
would confuse us.  Please attach your patch to the PR, and add any 
thoughts you might have about these sorts of corner cases.

Thanks,

--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713