[patch] More fallout from Fix PR c++/28250: ICE with invalid catch

[Volker Reichelt]

On 21 Jul, Mark Mitchell wrote:
> Volker Reichelt wrote:
> 
>> The patch fixes 4 ICEs, namely the remaining part from PR 28250,
>> and the PRs 28257, 28259, and 28267.
>> 
>> Shall I apply this to mainline?
> 
> Once we resolve the issues, below, yes, I think you should.
> 
>> What about the 4 regressions?
> 
> Remove the expected warnings.

OK.

> And see my previous invective on warnings issued from the middle end. :-)

[snip]

>> @@ -3257,6 +3258,9 @@ cp_finish_file (void)
>>      }
>>    while (reconsider);
>>  
>> +  if (errorcount || sorrycount)
>> +    return;
>> +
> 
> Why do you want to exit early here?

Because later in cp_finish_file we have

  cgraph_finalize_compilation_unit ();
  cgraph_optimize ();

If I don't skip these, I get ICEs from the middle-end.
I could move the check for (errorcount || sorrycount) right before
those calls. I chose the earliest position for the check that causes
no further regressions instead, but I don't know which is really
better (the testresults are the same).

Regards,
Volker