On 09/30/2011 05:39 AM, Dean, Sandy wrote:
Here is the question I asked IBM:
What is IBM's stance on supporting perl from this point
forward? I am a heavy user of perl in the USS environment, and would
need to redesign my applications if perl were to go unsupported.
The open source group
On 10/01/2011 01:49 AM, Jarkko Hietaniemi wrote:
I don't know about bad blood... irritation, maybe. Amazement at the
pigheadedness of IBM, why they don't see how the current model doesn't
work. Their current model of maintaining an obsolete fork doesn't work.
Well, to be fair, they do every
On 01/12/2012 07:52 AM, Helmut PFOSER wrote:
Hello list,
I've installed the PERL-Modules XML::Parser 2.41 and Expat 2.0.1 on my
z/OS 1.11 USS. PERL Base version installed with z/OS 1.11 is 5.8.7
Running gmake test I've got a lot of errors, I will attach the output
at the end of the mail.
I've
The bottom (actually top :) ) line is that at least some of these errors
were the result of makedepend failing to realize that the C89 standard
allows preprocessor directives to be indented, and failing to account
for continuation lines in preprocessor lines. I am smoking some fixes
for this.
On 02/23/2013 08:09 PM, John Goodyear wrote:
The output looks better. The only conditional declaration error left is
the extra #endif at the end of perly.c. Of course the missing header
errors are there, but that should be OK for now.
I have a suspicion that the inserted #endif was a kludge
On 02/25/2013 05:43 AM, demerphq wrote:
On 25 February 2013 13:36, demerphq demer...@gmail.com wrote:
On 25 February 2013 13:33, demerphq demer...@gmail.com wrote:
On 23 February 2013 19:41, John Goodyear johng...@us.ibm.com wrote:
Finding dependencies for regexec.o.
ERROR CCN3287
On 02/25/2013 07:48 AM, John Goodyear wrote:
Running Perl's make, the first item I ran up against was
ERROR CCN3343 ./perl.h:4494 Redeclaration of PL_fold_latin1 differs
from previous declaration on line 537 of ./utfebcdic.h.
I noticed that Mariusz Stakowski posted this issue last
On 02/25/2013 09:36 AM, John Goodyear wrote:
I'm just getting started with my efforts on z/OS.
It's early, but are my efforts enough to allow you to put away the cleaver ?
I'm optimistic that we can get this working; I'm committed to it, and
John looks to be to.
The problem is keeping
, however
I'm doing some research to verify this.
Has this been experienced on any other platforms?
I don't see what that array is included so many times, and will have to
look at it tomorrow. That array is not used on other platforms.
John Goodyear
Karl Williamson wrote on 02/26
On 02/26/2013 10:21 PM, Karl Williamson wrote:
It looks like the macro expasion split the array pPL_e2utf across 2
lines, and also inserted a couple of parenthesis. It happened right
at offset 2048, so it seems we're hitting another 2K limit, however
I'm doing some research
On 02/27/2013 07:25 AM, Andy Dougherty wrote:
I've pushed this change to the khw/ebcdic topic branch.
(Karl, I hope you don't mind me doing so!)
Thank you for doing that
On 02/28/2013 06:51 PM, John Goodyear wrote:
H.Merijn Brand wrote on 02/28/2013 02:23:33 AM:
Subject: Re: New EBCDIC branch available
On Wed, 27 Feb 2013 16:36:46 -0700, Karl Williamson
pub...@khwilliamson.com wrote:
On 02/26/2013 08:41 PM, John Goodyear wrote:
Karl,
I
On 02/28/2013 10:10 PM, Karl Williamson wrote:
please find out how to specify a quantifier in your system's sed and let
me know)
I googled the error message, and looking at the result, I expect that it
can't deal with a number as large as 1000. You could experiment with
smaller ones to see.
On 03/04/2013 11:35 AM, John Goodyear wrote:
I tried to say in an earlier email that the make after the miniperl gets
linked is likely to fail spectacularly. Instead of continuing the make
at this point, use the already-linked miniperl and do the Hello world
test:
./miniperl -le
On 03/04/2013 12:40 PM, John Goodyear wrote:
On 03/04/2013 11:35 AM, John Goodyear wrote:
I tried to say in an earlier email that the make after the
miniperl gets
linked is likely to fail spectacularly. Instead of continuing
the make
at this point, use the already-linked
On 03/04/2013 02:26 PM, John Goodyear wrote:
On 03/04/2013 12:40 PM, John Goodyear wrote:
On 03/04/2013 11:35 AM, John Goodyear wrote:
I tried to say in an earlier email that the make after the
miniperl gets
linked is likely to fail spectacularly. Instead of
. Hopefully, I was
wrong, and this will work.
From ef3470822f27436a5919b55baa64e3c4d51a2ef5 Mon Sep 17 00:00:00 2001
From: Karl Williamson pub...@khwilliamson.com
Date: Mon, 4 Mar 2013 19:16:31 -0700
Subject: [PATCH 88/88] XXX: perly.c: Reinstate some ebcdic code
This is an experiment to see
On 03/05/2013 08:00 AM, John Goodyear wrote:
JCG: yylex returns 301: 'ýè'
I am trying to learn more about this area of the code. Where did you
put your print statement that prints this?
Last night, John Goodyear successfully did a Hello World, and today he
ran a full script that read, re-ordered, and wrote a file. So there is
progress being made
On 02/20/2014 08:29 AM, Brad Gilbert wrote:
On Wed, Feb 19, 2014 at 11:43 PM, Konovalov, Vadim
vadim.konova...@emc.com wrote:
From: Karl Williamson [mailto:pub...@khwilliamson.com]
But as it turns out, I agree with him. Unless we can get a regular smoke
facility, there's no point in trying
On 04/14/2014 07:08 AM, Yaroslav Kuzmin wrote:
so file must be created to work correctly in ASCII and EBCDIC system.
I'm having trouble following who said what in this discussion and what
if anything I'm being asked. So I hope the following helps.
charclass_invlists.h is one of 4 files
Thanks, I have replaced the branch
http://perl5.git.perl.org/perl.git/shortlog/refs/heads/khw/ebcdic
with a new one that has this fixed.
On 04/25/2014 02:44 AM, Yaroslav Kuzmin wrote:
perl-5.19
In file regcomp.c in function S_regclass in define EBCDIC
string
14002 if
On 04/29/2014 01:40 AM, Yaroslav Kuzmin wrote:
I have a in vim function to remove spaces, tab end of the line.
But there were no trailing spaces in the original.
The new branch at
http://perl5.git.perl.org/perl.git/shortlog/refs/heads/khw/ebcdic
hopefully has this fixed.
On 04/30/2014 05:22 AM, Yaroslav Kuzmin wrote:
perl-5.19
In file utf8.h in #ifndef EBCDIC.
there's
597 #define is_UTF8_CHAR_utf8_safe(s,e)
and used in define
629 #define
http://perl5.git.perl.org/perl.git/shortlog/refs/heads/khw-ebcdic
has a corrected toCTRL() macro
On 06/06/2014 11:43 AM, Ricardo Signes wrote:
- Forwarded message from Ricardo Signes perl@rjbs.manxome.org -
Date: Fri, 6 Jun 2014 10:39:50 -0400
From: Ricardo Signes perl@rjbs.manxome.org
To: perl5-port...@perl.org
Subject: platform support
List-Id: perl5-porters.perl.org
driven from a Linux server with linux
and git access
It should work
Sandra
The new branch, rebased off of blead, at
http://perl5.git.perl.org/perl.git/shortlog/refs/heads/khw/ebcdic
should fix the build problem your log shows.
From: Yaroslav Kuzmin ykuz...@rocketsoftware.com
To: Karl Williamson
On 06/10/2014 01:23 AM, Yaroslav Kuzmin wrote:
P.S. Why such an attachment to the system smoke ?
test logs and so I can provide.
I don't understand this question. Hopefully it was addressed to someone
else.
On 06/19/2014 04:38 AM, Yaroslav Kuzmin wrote:
On z/OS v1r13 using the compiler XLC .
Perl-5.20 build without module mro.
P.S. The company I work not only with perl, but also entry for other projects.
Yaroslav Kuzmin
On 06/19/2014 11:14 AM, Karl Williamson wrote:
On 06/19/2014 04:38 AM, Yaroslav Kuzmin wrote:
On z/OS v1r13 using the compiler XLC .
Perl-5.20 build without module mro.
P.S. The company I work not only with perl, but also entry for other
projects
On 06/20/2014 05:41 AM, Yaroslav Kuzmin wrote:
(07:39) RS12 : PDKUZM | ~/ussport/perl/perl-5.20 : perl -v
This is perl 5, version 19, subversion 12 (v5.19.12) built for os390
Ok. That is somewhat behind. There have been several fixes that should
improve things. The latest is at
456318 · Russia
Tel: +7.922.2.38.33.38
Email: ykuz...@rocketsoftware.com
Web: www.rocketsoftware.com
От: Karl Williamson pub...@khwilliamson.com
Отправлено: 23 июня 2014 г. 10:19
Кому: Yaroslav Kuzmin; Ricardo Signes
Копия: carr...@nationwide.com; perl-mvs
От: Karl Williamson pub...@khwilliamson.com
Отправлено: 23 июня 2014 г. 10:19
Кому: Yaroslav Kuzmin; Ricardo Signes
Копия: carr...@nationwide.com; perl-mvs@perl.org
Тема: Re: НА: НА: НА: НА: [perl@rjbs.manxome.org: platform support]
On 06/20/2014 05:41 AM, Yaroslav
On 06/26/2014 12:55 AM, Yaroslav Kuzmin wrote:
perl/perl-5.20.0 : ./perl -v
This is perl 5, version 19, subversion 12 (v5.19.12) built for os390
The version now in
http://perl5.git.perl.org/perl.git/shortlog/refs/heads/khw/ebcdic
has some fixes that should clean up these failures
files in the failure :
comp/*.t cmd/*.t run/*.t io/*.t re/*.t opbasic/*.t op/*.t uni/*.t
On 06/30/2014 03:12 AM, Yaroslav Kuzmin wrote:
git commit 168f82262ef7b8feb200e1c5d09f747414896cf0
Author: Karl Williamson k...@cpan.org
Date: Fri Jun 27 16:31:41 2014 -0600
~/ussport/perl/perl-5.20
On 06/30/2014 01:30 PM, Karl Williamson wrote:
The problem here is probably something simple, easy to fix once we find
it. This is an easy method that should find what the first bad regex
is, anyway:
cd t
./perl -Dr -Ilib TEST base/*.t /dev/tty
In the unlikely event that doesn't generate
On 07/07/2014 04:18 AM, Yaroslav Kuzmin wrote:
git commit 69c0f27f128a54f4769f18fa9e03e4b29f45aff2
~/ussport/perl/perl-5.20 : ./perl -v
This is perl 5, version 21, subversion 2 (v5.21.2) built for os390
Copyright 1987-2014, Larry Wall
MVS (OS390) port by Mortice Kern Systems, 1997-1999
On 07/24/2014 09:39 PM, Yaroslav Kuzmin wrote:
~/ussport/perl/perl-5.20 : ./perl -v
This is perl 5, version 21, subversion 2 (v5.21.2) built for os390
Thanks, this enabled me to find and fix some bugs, including one that
has repercussions in several places, and the next iteration should
On 08/15/2014 03:09 AM, Yaroslav Kuzmin wrote:
git commit 09861a101b2aabfee2ad22840d97fd16e04c03f0
Author: Karl Williamson k...@cpan.org
Date: Tue Jul 29 20:47:15 2014 -0600
Make test
Yaroslav Kuzmin
On 08/15/2014 03:35 AM, Yaroslav Kuzmin wrote:
Hi gues,
A method of converting source project with a script
(translate_to_native.pl) , is not convenient, since it requires first
compile the project without the support of UTF-EBCDIC, and then convert
the source project with the aid of the script
On 10/07/2014 04:29 AM, Peter Rabbitson wrote:
My main concerns is that the logs are written out in EBCDIC, and I am
not sure if you will be able to properly read them if they are gzip-ed
on the very same mainframe hardware. To clear this up I am attaching a
smaller test-log in the form of:
Those logs showed me a fairly basic flaw which could be causing several
of the remaining failures. In the meantime, I have fixed other bugs
based on code reading, and looking at previous logs.
So there is a new version to try,
http://perl5.git.perl.org/perl.git/tree/refs/heads/khw/ebcdic
On 10/09/2014 07:18 AM, Yaroslav Kuzmin wrote:
I build perl with option DEBBUGGING.
git commit 5902e2e4
perl -v
This is perl 5, version 21, subversion 2 (v5.21.2) built for os390
Copyright 1987-2014, Larry Wall
MVS (OS390) port by Mortice Kern Systems, 1997-1999
log file in EBCDIC encoding.
.
Yaroslav Kuzmin
Developer C/C++ ,z/OS , Linux
3 Zhukovskiy Street · Miass, Chelyabinsk region 456318 · Russia
Tel: +7.922.2.38.33.38
Email: ykuz...@rocketsoftware.com
Web: www.rocketsoftware.com
От: Karl
...@rocketsoftware.com
Web: www.rocketsoftware.com
От: Karl Williamson pub...@khwilliamson.com
Отправлено: 9 октября 2014 г. 23:56
Кому: Yaroslav Kuzmin; Peter Rabbitson; perl5-port...@perl.org;
perl-mvs@perl.org
Тема: Re: НА: НА: НА: Bug Report z/OS
On 10/09/2014 11:06
On 10/14/2014 06:58 AM, Ricardo Signes wrote:
* Yaroslav Kuzmin ykuz...@rocketsoftware.com [2014-10-14T05:03:10]
P.S. Last week I sent a email to Peter, but he has not written anything in
the mailing list.
Peter has been traveling overseas, so he may be experiencing a bit of delay. I
believe
On 10/14/2014 03:03 AM, Yaroslav Kuzmin wrote:
Number of passed tests increased to 94%.
Note that this result is not from the latest from me that Yaroslav had
tried, but from a previous version. The latest should up the percentage
passed to 95+%.
mro is currently split into mro.c and a module in ext/mro. This
arrangement isn't working in os390, EBCDIC, because of a loader issue.
The EBCDIC porter, Yaroslav Kuzmin, can get it to work in os390 by
adding the contents of ext/mro.c to ./mro.c, and making ext/mro.xs into
a dummy.
There are just a few problems left in the port of the Perl core to
os390. Most of them are test issues.
This turns out to be the case with the -DDEBUGGING option causing a huge
make test output to be generated. This was caused by a test turning on
tracing by setting $^D, which it avoids on
On 10/21/2014 05:51 AM, Yaroslav Kuzmin wrote:
I build perl 5.21.5 (git commit e833c9b8, branch smoke-me/khw-ebcdic, but I
took a file hints/os390.sh from branch khw/ebcdic)
~/ussport/perl/perl-5.21.5 : ./Configure -DDEBUGGING ( without module mro)
- OK.
That branch was just to make
On 10/22/2014 12:02 PM, Karl Williamson wrote:
On 10/21/2014 05:51 AM, Yaroslav Kuzmin wrote:
I build perl 5.21.5 (git commit e833c9b8, branch smoke-me/khw-ebcdic,
but I took a file hints/os390.sh from branch khw/ebcdic)
~/ussport/perl/perl-5.21.5 : ./Configure -DDEBUGGING ( without
module
On 10/24/2014 01:59 AM, Yaroslav Kuzmin wrote:
I build perl-5.21.6 (git commit 3e2d781a)
~/ussport/perl/perl-5.21.5 : ./perl -v
This is perl 5, version 21, subversion 6 (v5.21.6) built for os390
But to successfully compile I had to change two files. Makefile.SH and
ext/mro/mro.xs
On 11/09/2014 11:00 AM, Karl Williamson wrote:
On 11/06/2014 05:41 AM, Yaroslav Kuzmin wrote:
I build project perl-5.21.6 (git commit 70291a84)
: perl -v
This is perl 5, version 21, subversion 6 (v5.21.6) built for os390
https://drive.google.com/file/d/0B5PTttxwo7qANjkxdl8zREZfNTQ
...@rocketsoftware.com
Web: www.rocketsoftware.com
От: Karl Williamson pub...@khwilliamson.com
Отправлено: 9 ноября 2014 г. 23:00
Кому: Yaroslav Kuzmin
Тема: Re: НА: updated EBCDIC porting status
On 11/06/2014 05:41 AM, Yaroslav Kuzmin wrote:
I build project perl
: +7.922.2.38.33.38
Email: ykuz...@rocketsoftware.com
Web: www.rocketsoftware.com
От: Karl Williamson pub...@khwilliamson.com
Отправлено: 13 ноября 2014 г. 10:19
Кому: Yaroslav Kuzmin
Копия: perl-mvs; Perl5 Porters
Тема: Re: НА: НА: updated EBCDIC porting status
On 11/19/2014 05:45 AM, Yaroslav Kuzmin wrote:
On z/OS in perl processing of script lib/sigtrap.pm it hangs
on string
107
s/([\N{U+80}-\N{U+FF}])/sprintf(M-%c,utf8::unicode_to_native(utf8::native_to_unicode(ord($1))0177))/eg;
in function S_regclass overflow occurs in a loop variable.
This is great news! We finally have a smoke report for z/OS after an
absence of many years. Getting these regularly is the criteria for
keeping support for this platform going forward, and it appears it's
going to happen.
I will need access to logs in order to debug the failures.
On
On 11/21/2014 03:03 AM, Yaroslav Kuzmin wrote:
I have no way of sending mails with mainframe. I take the report and log files
from the directory
logs/smokecurrent/
https://drive.google.com/file/d/0B5PTttxwo7qAM2Fuel84bzU0VXM
On mainframe no git tools. Therefore, the revision is substituted
On 11/24/2014 05:26 AM, Yaroslav Kuzmin wrote:
Smoke logs available at -
https://drive.google.com/file/d/0B5PTttxwo7qATVQybXJkRW1sa2c
Automated smoke report for branch ? 5.21.6 patch
RS12: 2817 (2817/)
onos/390 - 23.00
using c99 version
smoketime 1 hour 21
We're getting this failure on os390:
# Failed test 'S_IFMT()'
# at Fcntl/t/mode.t line 88.
# got: '4278190080'
# expected: '-16777216'
If you add those two numbers up it becomes clear that this bug is merely
one of signed vs unsigned. Likely a constant is missing its trailing
This fixes (I hope) about 20% of the remaining failures. Some tests are
cut down so that later you can run them by hand without getting
extraneous stuff. But for now, just smoke it as usual
Thanks
Karl Williamson
On 11/27/2014 05:09 AM, Yaroslav Kuzmin wrote:
Smoke logs available at -
https://drive.google.com/file/d/0B5PTttxwo7qAUnhnZHE5aUFFa2M
Automated smoke report for branch ? 5.21.7 patch
RS12: 2817 (2817/)
onos/390 - 23.00
using c99 version
smoketime 1 hour 17 minutes
On 11/26/2014 03:32 PM, Tony Cook wrote:
On Wed, Nov 26, 2014 at 02:39:52PM -0700, Karl Williamson wrote:
We're getting this failure on os390:
# Failed test 'S_IFMT()'
# at Fcntl/t/mode.t line 88.
# got: '4278190080'
# expected: '-16777216'
If you add those two numbers up
On 11/26/2014 03:32 PM, Tony Cook wrote:
I suspect the other S_I* constants should be unsigned too.
Me too, but I'm leaving that to someone else.
I hopefully have fixed some errors, and added debugging output to help
find some of the others.
This is rebased to the latest blead. It likely won't merge automatically.
I hopefully have fixed a couple more tests, and added even more
debugging output to help find some of the others.
This is rebased to the latest blead. It likely won't merge automatically.
I hopefully have fixed a 3 or 4 more tests, and added more debugging
output for another one.
This is rebased to the latest blead. It likely won't merge automatically.
As of today, we are down to 19 failures of the tests we run on os390,
EBCDIC. This is a 98% pass rate. I'll need help in squashing most of
the rest. I'm annotating them all here. Suggestions welcome. They are
grouped in categories.
The following Deparse failure is the most worrisome to
On 12/03/2014 01:29 PM, Father Chrysostomos wrote:
Karl Williamson wrote:
The following Deparse failure is the most worrisome to me of the
remaining issues This is because it's returning malformed UTF-8, and it
looks to me like there is a wild read or write. If the system had a
tool like
There are several fixes.
This is rebased to the latest blead. It likely won't merge automatically.
It would be really good to run this on a DEBUGGING perl, manually if
necessary, so we can look at that output. The smoker code you've been
using (as I showed in my message 6 hours ago) is not
I have uploaded an html file to
https://www.dropbox.com/s/qii0stvgledw1yn/diff.out?dl=0
It's too large to send to this list.
It contains a 'diff' listing between running a test (with the -DpTv
option) that is failing on ebcdic and passing on linux
To avoid any ambiguity, I've placed the
On 12/05/2014 03:41 PM, Father Chrysostomos wrote:
Where things really start to go wrong is at line #8696. The linux is
looking at my ... and os390 is looking at \nmy ...
I'm doing this blindly. Is this the 'my' in the here-doc?
$WANT = 'EOT';
my $VAR1 = '42
';
EOT
Yes.
The test is attached in its entirety. I see no real difference betweeen
running it on Linux vs os390 with -DpTv. It's just that the on Linux,
it outputs EXECUTING... and succeeds, whereas it doesn't do either on
os390, and the test harness says
FAILED--expected 1 tests, saw 0.
I can't
The test is attached in its entirety. I see no real difference betweeen
running it on Linux vs os390 with -DpTv. The file has been pared down
to 2 tests. In both, Linux says it is at EOF, and EBCDIC says line 2
Can't declare do block in my at - line 2, at EOF
---
Can't declare do block
We have found and changed a potential bug that could be causing some or
all of the parsing problems. I've cleaned up a couple of other tests.
This is rebased to the latest blead. It likely won't merge automatically.
Again, this should be run on a DEBUGGING perl.
the smoking code.
On 12/04/2014 04:57 PM, Karl Williamson wrote:
I had never looked at the end of the smoke reports we're getting until
today, but this is the error:
==
Configuration: -Dusedevel -DDEBUGGING
On 12/07/2014 11:12 PM, Karl Williamson wrote:
We have found and changed a potential bug that could be causing some or
all of the parsing problems. I've cleaned up a couple of other tests.
This is rebased to the latest blead. It likely won't merge automatically.
Again, this should be run
On 12/08/2014 08:13 AM, Karl Williamson wrote:
On 12/07/2014 11:12 PM, Karl Williamson wrote:
We have found and changed a potential bug that could be causing some or
all of the parsing problems. I've cleaned up a couple of other tests.
This is rebased to the latest blead. It likely won't
We sometimes get this message:
dist/Data-Dumper/t/bugs ... #
Failed test 'GVs with UTF8 names (or not, depending on perl version)'
# at t/bugs.t line 116.
# got: 'GLOB(0x21185f60)'
# expected: 'GLOB(0x21201838)'
CEE0802C Heap storage control
On 12/06/2014 11:16 PM, Father Chrysostomos wrote:
Karl Williamson wrote:
The test is attached in its entirety. I see no real difference betweeen
running it on Linux vs os390 with -DpTv. The file has been pared down
to 2 tests. In both, Linux says it is at EOF, and EBCDIC says line 2
Can't
On 12/07/2014 11:12 PM, Karl Williamson wrote:
The previous version fixed 6 of the 18 failing tests. This new version
adds some more debugging output in places, and adds a couple of tests
that had been skipped for a long time because the core code wasn't ready
to test them. Hopefully
We still have the problem of the smoke not running DEBUGGING. Below is
extracted from the log file Yaroslav posted. I searched for the FSUM
error codes online, and found no extra detail beyond the text in the
accompanying messages.
Configuration: -Dusedevel -DDEBUGGING
On 12/07/2014 11:12 PM, Karl Williamson wrote:
I have pushed a new branch to be tested. I may have fixed just one more
test, but have added more debugging to track down some of the others.
The previous version had added a couple of tests that had been skipped
for a long time because
Thanks for the latest smoke result. I see that you got DEBUGGING to work
in it. I have pushed a new branch to be tested. It likely won't merge
automatically.
But it includes a temporary patch to Configure to get your bash. That
should keep you from having to remove Policy.sh.
There is
On 12/16/2014 07:42 PM, Karl Williamson wrote:
Thanks for the latest smoke result. I see that you got DEBUGGING to work
in it. I have pushed a new branch to be tested. It likely won't merge
automatically.
But it includes a temporary patch to Configure to get your bash. That
should keep you
This Perl code in t/op/stat.t:
ok(opendir(DIR, .), 'Can open . dir') || diag Can't open '.':
$!;
use warnings;
ok(stat(DIR), stat() on dirhandle works);
ends up calling this code in pp_sys.c:
PL_laststatval = PerlLIO_fstat(my_dirfd(IoDIRP(io)), PL_statcache);
I have pushed a new branch to be tested, incorporating today's blead. It
likely won't merge automatically.
It looks to me like you are getting close to being able to do this
completely automatically. I got the sense that the latest result was
emailed automatically.
Feel free to try out
On 12/17/2014 06:17 PM, Craig A. Berry wrote:
On Wed, Dec 17, 2014 at 5:56 PM, M. Ray Mullins m.ray.mull...@gmail.com wrote:
Finding out what the C runtime put in errno would be helpful.
Printing $! in the test script would likely do the trick, though I
don't see Perl_my_dirfd setting errno
On 12/25/2014 01:10 AM, Yaroslav Kuzmin wrote:
Smoke logs available at -
https://drive.google.com/file/d/0B5PTttxwo7qAakdfbkRDRklfWDA
Automated smoke report for branch ebcdic 5.21.7 patch
8f7949da3760661252f10651cd69851f92dc739f v5.21.6-755-g8f7949d
RS12: 2817 (2817/)
onos/390 -
, Karl Williamson wrote:
This Perl code in t/op/stat.t:
ok(opendir(DIR, .), 'Can open . dir') || diag Can't open '.':
$!;
use warnings;
ok(stat(DIR), stat() on dirhandle works);
ends up calling this code in pp_sys.c:
PL_laststatval = PerlLIO_fstat(my_dirfd(IoDIRP(io)), PL_statcache
On 12/30/2014 12:10 AM, Yaroslav Kuzmin wrote:
P.S. I am on vacation from 01.01.2015 to 12.01.2015
Have a good vacation
This would be a good time to update the current status. There are
currently 10 failing core tests, 99% pass rate. These last few are
proving somewhat intractable. I'll
On 01/23/2015 12:08 AM, Yaroslav Kuzmin wrote:
It's not clear to me if you've tried changing the verbosity
configuration file and that's what you've sent us, and that the files
that are supposed to be generated by that option aren't getting
generated, or if you haven't gotten to that yet.
In
There was progress in this smoke. I've added code to Cwd to work around
the $SYSNAME self-pointing symbolic link. But we get things like
'/RS12/TMP/_WVLFEWRL4/ALPHA/BETA/GAMMA/DELTA/EPSILON'
Can you explain what that /RS12 is?
The run/locale.t failure is due to getting an environment
This fixed some things, but we have a new failure due to a new blead
feature, in bop.t.
There's a new branch with some more development, and debugging info to test.
On 02/02/2015 06:23 AM, Yaroslav Kuzmin wrote:
Smoke logs available at
This failure was caused by an unrelated issue on blead; it also broke
win32 builds. The latest version to test now has that fixed, and
rebased to the latest blead
On 02/03/2015 12:35 AM, Yaroslav Kuzmin wrote:
Smoke logs available at
This smoke gave the results so that I think we now know how to fix the
locale test failure. I have pushed another branch which should tell.
I hope to have fixed dist/ExtUtils-CBuilder/t/04-base.t in this branch,
and got closer to the problem in dist/Data-Dumper/t/dumper.t.
There were no
On 01/16/2015 12:45 AM, Yaroslav Kuzmin wrote:
Smoke logs available
athttps://drive.google.com/file/d/0B5PTttxwo7qAZk5faE1ic0w3OWs
(set verbose mode -v 2 Test::smoke, but file *.log *.out *.rpt not copy in dir
smoke/logs )
I think this didn't set the verbose mode properly. Further
On 01/15/2015 06:44 AM, Yaroslav Kuzmin wrote:
Concerning LC_ALL.
Do you have some initialization code for your
shell that sets the others? The point of LC_ALL is you can use it,
setting one thing instead of having to many.
In system z/OS USS, I do not do additional settings LC_ALL
But at a
On 02/12/2015 12:42 AM, Yaroslav Kuzmin wrote:
Smoke logs available at
https://drive.google.com/file/d/0B5PTttxwo7qANm5DUzVCRjlIQmc
This didn't give some output I thought it would. So could you run this
test manually and email the output?
cd t
./perl -I../lib
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
1 - 100 of 226 matches
Mail list logo