MIME Corruption

[Robert Atkinson]

Our support company is looking at this problem, but I wanted to know if anyone else has any clues.

I'm using MIME.EXE to Mime-encode a PDF file.
If I then extract the file back out, I find it's changed.

For instance, the first 4 lines change from this :-

%PDF-1.3
%JetForm PDF Support Version 2.3.000
%EncodingObject=0
%ûüÿ␊

...to this....

%PDF-1.3%JetForm PDF Support Version 2.3.000%EncodingObject=0%ûüÿ

This has been tried on another system using MIME 1.7, 1.8 and 1.9, but works fine.

I've taken their 1.7 and our 1.8, but both fail. I'm now wondering if a patch has upset this utility.

Rob.

[Ian Miller.]

do the file attributes change?

[Robert Gezelter]

Rob,

Check the line termination characters and record formats of both files. I suspect that you losing one or the other (i.e., if the file was variable length with implied carriage control, and then was output as a LF-UNIX style file), that would explain the problem.

- Bob Gezelter, http://www.rlgsc.com

[Robert Atkinson]

The PDF starts life as Stream. After being extracted, it ends up as Stream_LF.

If I use EXTRACT/FDL= with a Stream FDL, the attributes are then OK, but the file is still corrupt.

My feeling is that it's corrupted when stored.

Rob.

[Volker Halle]

Rob,

to compare the actual file contents, consider to use DUMP/BL=COUNT=1 on the original file and the corrupted one.

How is the attached file encoded in the MIME message (MIME> LIST) ?

Volker.

[Robert Atkinson]

I use /BINARY, so it's Base64, Octet/Stream.

Rob.

[Volker Halle]

Rob,

MIME uses about 13 shareable images (SDA> SHOW PROC/IMA/ID=).

You could compare their versions using anal/ima/select=(ident,build,link_time) sys$library:filename between the working and failing environments. If you believe you've found the 'bad' one, you could try to use a 'good' one in your process by defining a process logical pointing to a 'working' copy of the godd shareable library.

If you copy the MIME-encoded file from the failing to the working environment and extract the .PDF there, you can prove, if your theory (corruption during encoding) is right.

Volker.

[Robert Atkinson]

Cheers Volker.

We've identified the shareable images already and I'm checking these against patch kits we've applied to see if I can see anything.

I'm pretty confident that it's the packing, as when I email the MIME file, the attachment is corrupt, which is how I found the problem originally.

Unfortunately, I don't have a system that is working locally, but I might get our support company to send their files over, to test against.

Cheers, Rob.