Status of os390 port

We are down to 7 failing tests in the Perl 5 core.

One of them, stat.t, should pass with the new version I just pushed. 
Its failure appears due to a bug in os390 libc; it fails in a vanilla C 
program (thanks to Jarkko Hietaniemi for suggesting to try that, and 
Bram for reminding me) and has been reported to IBM support.

dist/Data-Dumper/t/dumper.t is a failure in the .t.  I just need to fix 
that up.

op/pack.t may have bugs in the core.  But if so, it is in little-used 
areas.  I need to get some time to look at this in depth.

porting/readme.t
lib/DBM_Filter/t/encode.t
lib/open.t

all appear to be from Encode bugs.  I will work on that as time permits.

That leaves ext/POSIX/t/sigaction.t as the one remaining real 
problematic core test.  The code that is failing is this:
     sub hiphup {
         is($_[1]->{signo}, SIGHUP, "SA_SIGINFO got right signal");
     }
     my $act = POSIX::SigAction->new('hiphup', 0, SA_SIGINFO);
     sigaction(SIGHUP, $act);
     kill 'HUP', $$;
and it leads to this:
CEE3206S The system detected a specification exception (System 
Completion Code=0C6).
          From entry point Perl_apply at compile unit offset +00001460 
at entry offset +00001460 at address 1F9B4A50.
ext/POSIX/t/sigaction ......................................... 
FAILED--expected 33 tests, saw 30

My guess is that it is a bug (or feature) of os390, and should be 
skipped.  Suspiciously, another IBM product, AIX 4.2 (though fixed in 
later versions), is known to be buggy for this test, and so if I don't 
hear otherwise soon, I'll skip this test, like it is for AIX 4.2.

The last time we tried the non-core tests, the pass rate was about 91%. 
  Since then Pod::Simple has been fixed.  But the next step is to get 
Encode working, as it is depended on by a lot of code, including some in 
core, mentioned above.

All but one of the code fixes for os390 are in blead.  I wuill get that 
one, dealing with handling the \N{...} construct in single quotish 
context in blead before the code freeze for v5.22, which is in a month.

There are many .t changes to get EBCDIC to work that aren't in blead. 
Again I'll get those in by the code freeze.

The big remaining issue is how to deal with mro.  Yaroslav had loader 
problems with what is shipped in blead, and came up with a solution that 
is controversial.  This needs to be looked at.

In addition, these messages show up in the logs, but don't lead to 
actual failures:

dist/Exporter/t/warn .......................................... ok
FSUM3212 c99: Command option -x is incorrect for z/OS platform - ignored.
FSUM3228 c99: Input file c++ not found.
CCN0461(W) "EXTC89" is not a valid sub-option for "LANGLVL". Option is 
ignored.
CCN0461(W) "EXTC99" is not a valid sub-option for "LANGLVL". Option is 
ignored.
CCN8101(W) The numeric value "3296" specified in 
"report(treatAsError,3296)" is out of bounds. The option is ignored.
CCN8101(W) The numeric value "4108" specified in 
"report(treatAsError,4108)" is out of bounds. The option is ignored.
dist/ExtUtils-CBuilder/t/00-have-compiler ..................... ok
  IEW2457E 9208 SYMBOL CEESG003 UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2457E 9208 SYMBOL CEESTART UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2665S 40FF MODULE *NULL*  IS NON-EXECUTABLE AND WAS NOT SAVED BECAUSE
           STORENX=NEVER.
IEW5033 The binder ended with return code 12.
dist/ExtUtils-CBuilder/t/01-basic ............................. skipped
  IEW2457E 9208 SYMBOL CEESG003 UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2457E 9208 SYMBOL CEESTART UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2665S 40FF MODULE *NULL*  IS NON-EXECUTABLE AND WAS NOT SAVED BECAUSE
           STORENX=NEVER.
IEW5033 The binder ended with return code 12.
dist/ExtUtils-CBuilder/t/02-link .............................. skipped
FSUM3212 c99: Command option -x is incorrect for z/OS platform - ignored.
FSUM3228 c99: Input file c++ not found.
CCN0461(W) "EXTC89" is not a valid sub-option for "LANGLVL". Option is 
ignored.
CCN0461(W) "EXTC99" is not a valid sub-option for "LANGLVL". Option is 
ignored.
CCN8101(W) The numeric value "3296" specified in 
"report(treatAsError,3296)" is out of bounds. The option is ignored.
CCN8101(W) The numeric value "4108" specified in 
"report(treatAsError,4108)" is out of bounds. The option is ignored.
dist/ExtUtils-CBuilder/t/03-cplusplus ......................... skipped
t/04-base.t: 270: 
rv='/RS12/tmp/HIizIhBo4z/alpha/beta/gamma/delta/epsilon/zeta/eta/theta/iota/kappa/lambda'; 
rp='/tmp/HIizIhBo4z/alpha/beta/gamma/delta/epsilon/zeta/eta/theta/iota/kappa/lambda'; 
subdir='/tmp/HIizIhBo4z/alpha/beta/gamma/delta/epsilon/zeta/eta/theta/iota/kappa/lambda'
dist/ExtUtils-CBuilder/t/04-base .............................. ok
  IEW2457E 9208 SYMBOL CEESG003 UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2457E 9208 SYMBOL CEESTART UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2665S 40FF MODULE *NULL*  IS NON-EXECUTABLE AND WAS NOT SAVED BECAUSE
           STORENX=NEVER.
IEW5033 The binder ended with return code 12.
dist/ExtUtils-ParseXS/t/001-basic ............................. ok
  IEW2457E 9208 SYMBOL CEESG003 UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2457E 9208 SYMBOL CEESTART UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2665S 40FF MODULE *NULL*  IS NON-EXECUTABLE AND WAS NOT SAVED BECAUSE
           STORENX=NEVER.
IEW5033 The binder ended with return code 12.
dist/ExtUtils-ParseXS/t/002-more .............................. ok
  IEW2457E 9208 SYMBOL CEESG003 UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2457E 9208 SYMBOL CEESTART UNRESOLVED.  NO CALL LIBRARY SPECIFIED.
  IEW2665S 40FF MODULE *NULL*  IS NON-EXECUTABLE AND WAS NOT SAVED BECAUSE
           STORENX=NEVER.