"Paul Marquess" <Paul.Marquess ntlworld.com> wrote 2007-11-05 18:26:04:
> From: Chun Bing Ge [mailto:gecbcn.ibm.com]
>
>
> > "Paul Marquess" <Paul.Marquessntlworld.com> wrote 2007-11-05
17:08:42:
> >
> > > From: Chun Bing Ge [mailto:gecbcn.ibm.com]
> > >
> > > > "Paul Marquess"
<Paul.Marquessntlworld.com> wrote 2007-11-02
> > 17:40:01:
> > > >
> > > > > From: Chun Bing Ge [mailto:gecbcn.ibm.com]
> > > > >
> > > ...
> > > > >
> > > > > Regarding the other patches you
sent, the one for 004gziphdr.t.
has
> > > > already
> > > > > been done by a previous patch. The
rest of your patches are all
for
> > the
> > > > > IO_Compress_Zlib/t/101truncate-*.t
series of tests. These are
all of
> > the
> > > > > form
> > > > >
> > > > > +if (ord("A") == 193) {
> > > > > + plan tests => 744 +
$extra;
> > > > > +}
> > > > > +else {
> > > > > plan tests => 734 +
$extra;
> > > > > +}
> > > > >
> > > > > Given that you are patching against
5.9.5, I don't know whether
> > these
> > > > are
> > > > > needed for the most recent
development version. I don't see
anything
> > in
> > > > > those test harness that would
account for the extra 10 tests, so
> > I'll
> > > > hold
> > > > > off on applying these for now.
> > > > >
> > > >
> > > > I think there is a better way which
could give you the answer!
> > > > You could change the source data in
those test cases into the
EBCDIC
> > > > values on linux,
> > > > like "abc" -->
"x81x82x83". Then you could get how many extra
tests
> > run
> > > > out!
> > >
> > > For this particular series of tests that
would be fine - the nature
of
> > the
> > > data being compressed doesn't really matter,
so I'll go with that.
> > >
> > > So I'll put a patch together that hopefully
sorts out the
compression
> > issues
> > > on EBCDIC, but is there anyone running the
latest development
sources to
> > > test them out?
> > >
> > > Paul
> > >
> > I'd like to give you the patches for these
tests(perl-current, 5.9.5).
> > I do not have time to test the lastest development
sources, so below
> > patches are FYI.
> > Any question, please let me know, thanks.
>
> Thanks, but hopefully I won't need those patches. I'm
going to make this
> change to truncate.pl instead to remove the ASCII data.
>
>
> *** ./t/lib/compress/truncate.pl Fri Nov 2
13:59:57 2007
> --- ./t/lib/compress/truncate.pl Fri Nov 2
14:12:04 2007
> ***************
> *** 14,25 ****
> my $Error =
getErrorRef($CompressClass);
> my $UnError =
getErrorRef($UncompressClass);
>
> ! my $hello = <<EOM ;
> ! hello world
> ! this is a test
> ! some more stuff on this line
> ! and finally...
> ! EOM
>
> my $blocksize = 10 ;
>
> --- 14,34 ----
> my $Error =
getErrorRef($CompressClass);
> my $UnError =
getErrorRef($UncompressClass);
>
> ! # my $hello = <<EOM ;
> ! #hello world
> ! #this is a test
> ! #some more stuff on this line
> ! #and finally...
> ! #EOM
> !
> ! # ASCII hex equivalent of the text above. This
makes the test
> ! # harness behave identically on an EBCDIC
platform.
> ! my $hello =
> !
"x68x65x6cx6cx6fx20x77x6fx72x6cx64x0ax74x
68x69x73" .
> !
"x20x69x73x20x61x20x74x65x73x74x0ax73x6fx
6dx65x20" .
> !
"x6dx6fx72x65x20x73x74x75x66x66x20x6fx6ex
20x74x68" .
> !
"x69x73x20x6cx69x6ex65x0ax61x6ex64x20x66x
69x6ex61" .
> ! "x6cx6cx79x2ex2ex2ex0a" ;
>
> my $blocksize = 10 ;
>
>
> The enclosed patch should apply to a fresh 5.9.5. It
updates all the
> compression modules to their latest versions. It should
include all the
> EBCDIC changes needed.
>
> Can you test this please? If it works ok, I'll supply
the patch to the
main
> development source.
>
Yes, it works ok!
Actually, I used this way to pass this tests on z/OS.
Ge, Chun Bing
空山新雨后，天气晚来秋
|