[Info-vax] Unexpected error using ZIP for OpenVMS
Jeremy Begg
jeremy.removethis at vsm.com.au
Mon Dec 12 00:05:51 EST 2011
Hi Steven,
I too ran into this issue a couple of months back but it wasn't critical so
I ignored it.
I've now tested your changes and have found just one minor shortcoming.
The file being ZIPped has these attributes:
VAX-11 RMS attributes
Record type: RMS-11 stream
File organization: Sequential
Record attributes: Implied carriage control
Record size: 0
Highest block: 1552
End of file block: 1553
End of file byte: 0
Bucket size: 0
Fixed control area size: 0
Maximum record size: 0
Default extension size: 0
Global buffer count: 0
Directory version limit: 0
I ZIPped it using your latest mods to ZIP (with LARGE FILE support) then
UNZIPped it using both UNZIP 5.52 and UNZIP 6.0. The resulting unzipped
file has these attributes:
VAX-11 RMS attributes
Record type: LF-terminated stream
File organization: Sequential
Record attributes: Implied carriage control
Record size: 0
Highest block: 1552
End of file block: 1553
End of file byte: 0
Bucket size: 0
Fixed control area size: 0
Maximum record size: 0
Default extension size: 16384
Global buffer count: 0
Directory version limit: 0
The Record type has changed from "Stream" to "StreamLF".
The actual bytes on disk appear to be the same as the original.
(I tested this on Itanium on VMS 8.4 and 8.3-1H1. Both produced the same
result. BACKUP/COMPARE on V8.4 ignored the different record formats but on
V8.3-1H1 it complained about every record!)
Thanks,
Jeremy Begg
Steven Schweda wrote:
>>I bumped across this error a while ago [...]
>
>
> It's been around since at least "Zip 2.0 (Sept 7th 1993)".
> (I prefer Zip bugs which are old enough to vote, because they
> can't be my fault.)
>
>
>> I haven't done enough testing yet, [...]
>
>
> I've now tried the proposed fix with long and short
> records in Stream, Stream_CR, and Stream_LF (and also
> Undefined), and the results all look plausible to me.
> DIFFERENCES may have some trouble with the long records, but
> BACKUP /COMPARE seems happy enough. Other record types
> should be unaffected by these changes. What could go wrong?
>
> For the record, accessing these long-record files over
> DECnet could lead to similar but different errors. For
> example:
>
> zip warning: non-translatable vms error code: 0x183A2
> %rms-e-netbts, network buffer too small for !ul byte record
> zip warning: could not open for reading:
> [.UTILITY.SOURCE.ZIP.DEV]LONG_REC_CR.DAT
>
> The proposed fix seems to clear those, too.
>
> Pending further complaints, I believe that I'm satisfied.
More information about the Info-vax
mailing list