On 04/27/2017 07:35 PM, Martin Sebor wrote:
On 04/25/2017 09:55 PM, Martin Sebor wrote:
On 04/25/2017 04:05 PM, Jeff Law wrote:
On 04/21/2017 03:33 PM, Martin Sebor wrote:
I think I may have included the (partial) the fix for 79062 to
get some tests to pass but I'm not 100% sure.
The following test fails without either this change or the
apparently equivalent change in fold-const.c:
gcc.c-torture/execute/builtins/sprintf.c
Do you have a preference for one or the other?
Let me first submit the fix for the -fexec-charset limitation
(bug 80523), see if I can separate out the partial fix for 79062,
and then resubmit this patch.
This patch was submitted here:
https://gcc.gnu.org/ml/gcc-patches/2017-04/msg01325.html
FWIW, my fix for bug 79062 is only partial (it gets the pass
to run but the warnings are still not issued). I don't quite
understand what prevents the warning flag(s) from getting set
when -flto is used. This seems to be a bigger problem than
just the sprintf pass not doing something just right.
Clearly the charset bug fix is a prerequisite for this one, as
is the two line change above to get the pass to run with LTO.
The rest (issuing warnings with LTO) seems like an independent
issue that this fix shouldn't have to be blocked on.
Do you agree?
Given we don't have a good handle on the root cause for 79062, it seems
we need to figure that out. Once we know the root cause that should
guide us WRT what to do with the 2 line hunk (ie, independent or not).
jeff
