Craig A. Berry wrote:
At 5:48 PM +0100 9/29/07, Nicholas Clark wrote:
At least one of the reasons ppm.t fails is that ppm is attempting to
do things with the tar format that may never work on VMS. For
example, you can't store object code in an archive that doesn't
preserve record boundaries. Perhaps current versions of the tar
specification have ways to store metadata in a way that could
preserve VMS attributes; I ought to look into it but have never found
the time. We could also reimplement the relevant pieces to use
Info-Zip or the native backup format. For now I think we need to
mark this as TODO.
d0:[craig.perl.lib.module.build.t]ppm.................ok 2/12
: checksum error
at d0:[craig.perl.lib.module.build.t]ppm.t line 122
: checksum error at d0:[craig.perl.lib.module.build.t]ppm.t line 122
: checksum error at d0:[craig.perl.lib.module.build.t]ppm.t line 122
: checksum error at d0:[craig.perl.lib.module.build.t]ppm.t line 122
Invalid header block at offset unknown at
d0:[craig.perl.lib.module.build.t]ppm.t line 122
Invalid header block at offset unknown at
d0:[craig.perl.lib.module.build.t]ppm.t line 122
Invalid header block at offset unknown at
d0:[craig.perl.lib.module.build.t]ppm.t line 122
The issue of not restoring VMS attributes does not appear to have
anything to do with the errors that are being displayed.
I have not yet determined if the tarball is being created corrupted, or
if the problem is in reading it.
-John
[EMAIL PROTECTED]
Personal Opinion Only
