> -----Original Message-----
> From: David Malcolm <dmalc...@redhat.com>
> Sent: Sunday, March 16, 2025 13:00
> To: Robert Dubner <rdub...@symas.com>; gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH]cobol: create new
> gcc/testsuite/cobol.dg/group1/check_88.cob test
>
> On Sat, 2025-03-15 at 16:14 -0500, Robert Dubner wrote:
> > This works on a x86_64-linux machine, although I had to do a complete
> > rebuild to make it take.
> >
> > If this meets with the approval of the global reviewers, please apply
> > it,
> > with a suitable commit message.
> >
> > The main characteristic of my trying to cope with modifying my
> > workflow
> > and coping with the GCC server and our COBOLWORX server, and
> > understanding
> > how ChangeLog entries get created has resulted in my finding new and
> > marvelous ways of tying my shoelaces together.
> >
> > I'll get the hang of it. But I have spent hours and hours sorting
> > out the
> > infrastructure, after having spent about an hour creating this new-
> > ish
> > code.
> >
> > Although I wouldn't want to have to explain just how much of that
> > hour was
> > spent figuring out that ----*---- had to be replaced with ----\[*\]--
> > --.
> >
> >
> > diff --git a/gcc/testsuite/cobol.dg/group1/check_88.cob
> > b/gcc/testsuite/cobol.dg/group1/check_88.cob
> > new file mode 100644
> > index 00000000000..b29fe626af6
> > --- /dev/null
> > +++ b/gcc/testsuite/cobol.dg/group1/check_88.cob
> > @@ -0,0 +1,116 @@
> > +*> { dg-do run }
> > +*> { dg-output "-><-\n" }
> > +*> { dg-output "-> <-\n" }
> > +*> { dg-output "->\"\"\"<-\n" }
> > +*> { dg-output "->000<-\n" }
> > +*> { dg-output "->ÿÿÿ<-\n" }
> > +*> { dg-output " \n" }
> > +*> { dg-output "-><-\n" }
> > +*> { dg-output "-> <-\n" }
> > +*> { dg-output "->\"\"\"\"<-\n" }
> > +*> { dg-output "->0000<-\n" }
> > +*> { dg-output "->ÿÿÿÿ<-\n" }
> > +*> { dg-output " \n" }
> > +*> { dg-output "There should be no garbage after character 32\n" }
> > +*> { dg-output
> > "-------------------------------\[*\]--------------------------------
> > \n" }
> > +*> { dg-output "üüüüüüüüüüüüüüüüüüü Bundesstraße
> > \n" }
> > +*> { dg-output "üüüüüüüüüüüüüüüüüüü Bundesstraße
> > \n" }
> > +*> { dg-output " \n" }
> > +*> { dg-output "There should be no spaces before the final quote\n"
> > }
> > +*> { dg-output "\"üüüüüüüüüüüüüüüüüüü Bundesstraße\"\n" }
> > +*> { dg-output " \n" }
> > +*> { dg-output " IsLow \"\"\n" }
> > +*> { dg-output " IsZero \"000\"\n" }
> > +*> { dg-output " IsHi \"ÿÿÿ\"\n" }
> > +*> { dg-output " IsBob \"bob\"\n" }
> > +*> { dg-output " IsQuote \"\"\"\"\"\n" }
> > +*> { dg-output " IsSpace \" \"\n" }
> > +*> { dg-output " \n" }
> > +*> { dg-output "CheckBinary Properly True\n" }
> > +*> { dg-output "CheckBinary Properly False\n" }
>
> FWIW, I don't think there's any ordering implied by all these dg-output
> directives, so that e.g.:
>
> +*> { dg-output "CheckBinary Properly True\n" }
> +*> { dg-output "CheckBinary Properly False\n" }
>
> is merely verifying that there are instances of those two lines to be
> pruned from the output, but if they appear in either order, the test
> will succeed, which might be too weak for some kinds of testing, i.e.
> as written, I *think* both:
>
> CheckBinary Properly True
> CheckBinary Properly False
>
> and:
>
> CheckBinary Properly False
> CheckBinary Properly True
>
> would pass, whereas I think you'd want the latter to be a failure.
>
> Perhaps the output messages could be modified a bit to disambiguate
> them?
>
> Alternatively, we have dg-{begin,end}-multiline-output; see the example
> in:
> https://gcc.gnu.org/pipermail/gcc-patches/2025-March/677314.html
> but:
>
> * I don't remember if that's only for stderr or if it will work on
> stdout too
> * COBOL doesn't have multiline comment blocks (though I suppose gcobol
> could add them if it really needed them, perhaps guarded by a command-
> line option)
> * the PASS/FAIL results are on each dg-{begin,end}-multiline-output
> pair without much info on what individual lines failed if you get a
> FAIL, so you might have to choose a suitable "granularity" to split the
> output up into, or else tracking down failures would become a big pain.
>
> Hope this is helpful
> Dave
Dave, my experiments showed what Jakub mentions in a later message: The
dg-output regex strings get chained together into one gargantuan regex
string that makes me nervous to look at. But it works.
I will look at multiline support. I don't see it mentioned in
https://gcc.gnu.org/onlinedocs/gccint/Directives.html, although Google did
find https://gcc.gnu.org/legacy-ml/gcc-patches/2015-09/msg00729.html, and
grep shows it appears a lot in the existing testcases.
