Setting test suite not to try debug output cases

Iain Sandoe idsandoe@googlemail.com
Mon Sep 5 12:29:30 GMT 2022



> On 5 Sep 2022, at 13:23, Koning, Paul via Gcc <gcc@gcc.gnu.org> wrote:
> 
> 
> 
>> On Sep 5, 2022, at 5:29 AM, Richard Biener <richard.guenther@gmail.com> wrote:
>> 
>> 
>> [EXTERNAL EMAIL] 
>> 
>> On Fri, Sep 2, 2022 at 8:57 PM Koning, Paul via Gcc <gcc@gcc.gnu.org> wrote:
>>> 
>>> Given that pdp11 no longer supports debug output, I get a lot more test failures, like this:
>>> 
>>> spawn -ignore SIGHUP /Users/pkoning/Documents/svn/build/pdp/gcc/xgcc -B/Users/pkoning/Documents/svn/build/pdp/gcc/ -mlra -fdiagnostics-plain-output -Og -g -w -c -o 20000105-1.o /Users/pkoning/Documents/svn/gcc/gcc/testsuite/gcc.c-torture/compile/20000105-1.c
>>> xgcc: warning: target system does not support debug output
>>> cc1: warning: target system does not support debug output
>>> FAIL: gcc.c-torture/compile/20000105-1.c   -Og -g  (test for excess errors)
>>> 
>>> I assume there is some way in the test suite machinery to globally skip all "debug output" cases.  How would I do that?
>> 
>> Hmm.  In testsuite/lib/prune.exp there's
>> 
>>   # Ignore stabs obsoletion warnings
>>   regsub -all "(^|\n)\[^\n\]*\[Ww\]arning: STABS debugging
>> information is obsolete and not supported anymore\[^\n\]*" $text ""
>> text
>> 
>> maybe you can (selectively for pdp11) add similar pruning of the
>> 'target system does not support debug output' message?
>> I think you should be able to use
>> 
>>  if { [istarget pdp11-*-*] } then {
>>    regsub -all " ... " ...
>>  }
> 
> Thanks, I'll look into that.  Should it be more general to cover any other targets that don't have debug output?  I think pdp11 isn't the only one but I'm not sure.

For the parts of the testsuite that have a specific  “debug.exp”  you could also make the .exp do an early return like

if { [istarget pdp11-*-*] } {
  return
}

That assumes that there is no point in actually compiling the testcases for “improved coverage” in some way.

Iain



More information about the Gcc mailing list