Re: Qshell jar zipped file has strange characters

[John Yeung <gallium.arsenide@xxxxxxxxx>]

On Mon, Mar 2, 2015 at 4:48 PM, Tim Bronski <tim.bronski@xxxxxxxxx> wrote:
> While it's true that the contents of the archive might unzip correctly, the
> actual files names that are restored on the destination system might very
> well be material to the receiver.

Right, which is why I then showed that extracting using the same jar
program worked fine.

> Since a zip archive has no record of the ccsid
> used on the creating system it is sometimes more than just helpful if the
> archive creator formats the name so it appears on the receiving system the
> same way as it did to the sender.

And if a ZIP archive indeed has no record of the CCSID of the creating
system, then how does does the receiving system know which CCSID to
use when extracting the names?  Hint:  It doesn't.  And what's more,
it doesn't really matter!

I haven't read the official JAR specification, but I imagine that it
has to specify an encoding.  And it would be as universal as possible
(UTF-8? That's what almost any sane person would use today, but I am
pretty sure Java was originally developed with UCS-2, so maybe it's
that).  So it's the archiver's responsibility to encode properly, and
the extractor's responsibility to decode properly.  With "properly"
being defined by the JAR specification.

Apparently, some software doesn't adhere to the spec, possibly
including some portions of the jar program itself; but as far as I can
tell, the archiving and extraction seem fine.  But I don't have other
machines to test on.

John Y.