Re: m4 ebcdic patch

[Michael Elizabeth Chastain]

> That's not sufficient.  You also have to show that, for example, the
> pattern [/-:] is equivalent to the pattern [/:0-9].

My theory of maintenance is to create incremental patches that can be
processed independently.  If you want a jumbo patch that fixes ALL
ebcdic problems, you'll have to wait for someone else to write such a
patch, test it, and execute an FSF assignment.

Also, my patch already contains all the code to process the pattern
[/-:].  It would just be a matter of changing the initialization strings
in the range_list table.  Thus it brings m4 much closer to your desired
behavior.

But I don't want to wade into the rat's nest of locale sensitivity.
One patch at a time.

> Sorry, I don't have a copy of your earlier email, but perhaps you can
> deduce some test cases from the above comments.

Okay, I've spent enough time on this.  The patch is in the bug-gnu-utils
archive here:

  http://mail.gnu.org/archive/html/bug-gnu-utils/2003-12/msg00055.html

Do whatever you want, including nothing.

Michael C