ECL gaining Unicode support

Until now, ECL accepted strings made of extended characters, with codes up to 21 bits. This feature was accessible when built with --enable-unicode, but it wasn't really useful because these strings could not be written or read from files.

As part of the effort to make ECL safer with respect to threads and interrupts, we had to change the way C streams are implemented. It didn't seem that much implementing different input-output formats.

And indeed it was not. The change involves right now fixes in the code for handling readtables, the code for printing symbols and other places, and it is probably incomplete. Nevertheless, as the screenshot shows https://sourceforge.net/project/screenshots.php?group_id=30035 it seems to work :-)

Posted by Juan Jose Garcia Ripoll 2008-10-24