List Info

Thread: Status of VMS with Symlinks on blead@31806




Status of VMS with Symlinks on blead@31806
user name
2007-09-08 12:37:06
With my submitted patches in place, and with one tentative
bugfix to 
stat() in VMS.C, here is a status of where things are for me
on VMS.

As rsync://ftp.linux.activestate.com is stuck at 31806, I
have not been 
able to easily test patches after that point.

Failed 21 tests out of 1332, 98.42% okay.

         [-.ext.Cwd.t]cwd.t

This needs more investigation.  The readlink() routine that
worked under 
PCP modes with the RMS SYMLINK kit is returning an error
condition for 
every thing I have tried with it.  I may need to add a mount
point for 
the disk on it self in order for this to work, or may need
to do 
something else.

I want to make sure that I understand the other failures
before I make 
changes related to this test.

         [-.ext.Safe.t]safe2.t

This was just discussed in another thread.

        
[-.lib.CPANPLUS.Dist.Build.t]02_CPANPLUS-Dist-Build.t

This passes 36 tests when run interactively.  There is a lot
of verbose 
output that is not preceded by # characters.

Destructively modifies path and perl5lib logical names in
the process 
table.  All the CPANPLUS tests are doing this.

         [-.lib.CPANPLUS.t]00_CPANPLUS-Internals-Utils.t

New failure:
not ok 19 -    Looks like a proper version
#   Failed test '   Looks like a proper version'
#   at [-.lib.cpanplus.t]00_cpanplus-internals-utils.t line
78.
#                   '5_10_0'
#     doesn't match '(?-xism:d.d.d)'

My guess is that somewhere a directory named '5.10.0' was
converted to 
'5_10_0' and that this regex can not deal with that.

If someone fixes this, please do not always expect VMS to
make this 
conversion.  Patches are in progress for VMS to optionally
not do this 
on ODS-5 volumes.

         [-.lib.CPANPLUS.t]08_CPANPLUS-Backend.t

not ok 230 - Found list of installed modules
#   Failed test 'Found list of installed modules'
#   at [-.lib.cpanplus.t]08_cpanplus-backend.t line 213.

I have not determined exactly what this test is looking for,
so I do not 
know how to fix.

         [-.lib.CPANPLUS.t]20_CPANPLUS-Dist-MM.t

This is now failing for me.  It was working before.  Lots of
failures 
starting at:

not ok 45 -    Makefile.PL deleted
#   Failed test '   Makefile.PL deleted'
#   at [-.lib.cpanplus.t]20_cpanplus-dist-mm.t line 261.

# Looks like you failed 7 tests of 81.

This will need further investigation.

         [-.lib.File.Find.t]find.t
         [-.lib.File.Find.t]taint.t

Find.t and taint.t are failing because a previous test is
leaving a file 
named '[.t].;' behind.  With that file removed, the tests
pass.  I have 
not found which test is leaving this behind.

         [-.lib.Module.Build.t]basic.t

This passes when run interactively.  Lots of output not
prefixed by #

         [-.lib.Module.Build.t]ext.t

This passes when run interactively.

         [-.lib.Module.Build.t]extend.t

not ok 7
#   Failed test at [-.lib.module.build.t]extend.t line 55.

Looking for a VMS directory as a subdirectory of a VMS
directory, and 
the directory delimiters are causing the test to fail.

         [-.lib.Module.Build.t]install.t

Manifying [.blib.lib]simple.pm ->
[.blib.libdoc]__simple.rno
Use of uninitialized value within dirs in pattern match (m//)
at 
/PROJECT_ROOT/PERL-BLEAD/lib/Pod/Man.pm line 795.

Use of uninitialized value $fixedpath in substitution (s///)
at 
../lib/File/Spec/VMS.pm line 503.
Use of uninitialized value $path in pattern match (m//) at 
../lib/File/Spec/VMS.pm line 39.
Use of uninitialized value in transliteration (tr///) at 
../lib/File/Spec/VMS.pm line 46.
Use of uninitialized value $path in substitution (s///) at 
../lib/File/Spec/VMS.pm line 47

Can't open perl script "Build": no such file or
directory
.
# Looks like you failed 6 tests of 34.
%SYSTEM-F-ABORT, abort


         [-.lib.Module.Build.t]manifypods.t

# Looks like you failed 3 tests of 21.

         [-.lib.Module.Build.t]metadata.t

Several cases of expecting the file system to preserve case,
and when 
VMS is put in that mode, it still fails as something appears
to go into 
a recursive loop in the test script.

         [-.lib.Module.Build.t]ppm.t

Use of uninitialized value within dirs in pattern match (m//)
at 
/PROJECT_ROOT/PERL-BLEAD/lib/Pod/Man.pm line 795.

: checksum error at [-.lib.module.build.t]ppm.t line 122
Invalid header block at offset unknown at
[-.lib.module.build.t]ppm.t 
line 122

not ok 11 - blib/html/site/lib/Simple.html
#   Failed test 'blib/html/site/lib/Simple.html'
#   at [-.lib.module.build.t]ppm.t line 178.
not ok 12 - blib/html/bin/hello.html
#   Failed test 'blib/html/bin/hello.html'
#   at [-.lib.module.build.t]ppm.t line 179.
# Looks like you failed 9 tests of 12.

         [-.lib.Module.Build.t]runthrough.t

not ok 7
#   Failed test at [-.lib.module.build.t]runthrough.t line
86.
#          got: 'Simple-0_01'
#     expected: 'Simple-0.01'

# Looks like you failed 3 tests of 32.

A number of other errors.

         [-.lib.Module.Build.t]test_type.t

This passes when run interactively, a lot of output not
prefixed by #

         [-.lib.Module.Build.t]tilde.t

tilde support is not in VMS right now.

         [-.lib.Module.Build.t]xs.t

Can't locate object method "blib
" via package "Module::Build" at 
/PROJECT_ROOT/PERL-BLEAD/lib/Module/Build/Base.pm line
2208.
# Looks like you planned 22 tests but only ran 15.
# Looks like your test died just after 15.

         [-.lib.Test.Harness.t]failure.t

This one is confusing, It claims all 6 tests pass when run 
interactively, but there is also much output about tests
that fail, and 
that output indicates that the tests are expected to fail.

The output is not prefixed with "#".

         op/utftaint.t

This test is dependent on the value of the logical name PATH
before it 
is run.  More investigation is needed.

-John
wb8tywqsl.net
Personal Opinion Only

Re: Status of VMS with Symlinks on blead@31806
user name
2007-09-08 19:26:29
At 12:37 PM -0500 9/8/07, John E. Malmberg wrote:
>With my submitted patches in place, and with one
tentative bugfix to stat() in VMS.C, here is a status of
where things are for me on VMS.
>
>As rsync://ftp.linux.activestate.com is stuck at 31806,
I have not been able to easily test patches after that
point.

It's working again.  I haven't been building with symlink
support so
I'm not seeing all the same failures you are, but here are a
few
comments in those few cases where I have gotten a little
further in
determining what's going on.

>NPLUS-Internals-Utils.t
>
>New failure:
>not ok 19 -    Looks like a proper version
>#   Failed test '   Looks like a proper version'
>#   at [-.lib.cpanplus.t]00_cpanplus-internals-utils.t
line 78.
>#                   '5_10_0'
>#     doesn't match '(?-xism:d.d.d)'
>
>My guess is that somewhere a directory named '5.10.0'
was converted to '5_10_0' and that this regex can not deal
with that.

It was because 10 doesn't match d, and #31800 should've
taken care of it.


>        [-.lib.Module.Build.t]basic.t
>
>This passes when run interactively.  Lots of output not
prefixed by #

The Module::Build tests have three completely different
failure
signatures when run with t/TEST, t/harness, and
individually.  Not
sure what all is involved, but one problem is that when a
test dies,
it doesn't clean up after itself, so the next test cleans up
and
emits a warning about having to clean up first.  When t/TEST
sees the
warning, it gives up the ghost with "Unexpected output
at test 0."

>
>        [-.lib.Module.Build.t]ext.t
>
>This passes when run interactively.
>
>        [-.lib.Module.Build.t]extend.t
>
>not ok 7
>#   Failed test at [-.lib.module.build.t]extend.t line
55.
>
>Looking for a VMS directory as a subdirectory of a VMS
directory, and the directory delimiters are causing the test
to fail.
>
>        [-.lib.Module.Build.t]install.t
>
>Manifying [.blib.lib]simple.pm ->
[.blib.libdoc]__simple.rno
>Use of uninitialized value within dirs in pattern match (m//)
at /PROJECT_ROOT/PERL-BLEAD/lib/Pod/Man.pm line 795.

There is a bug in Pod::Man::devise_title where it is
surprised when
an array element beyond the end of the array is undefined. 
I have a
band-aid in progress.



-- 
________________________________________
Craig A. Berry
mailto:craigberrymac.com

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser

[1-2]

about | contact  Other archives ( Real Estate discussion Medical topics )