On Wed, Oct 14, 2009 at 03:39:30PM +0000, The Fungi wrote:
[...]
> I'll see if I can get my hands on a virtual machine to replicate
> the issue and hopefully figure out a fix.
[...]
> I'll follow up again when I have something.
Okay, I was able to replicate the problem under IE 8 in XP within
VirtualBox, and can see the same behavior you described.
Unfortunately, neither the original AjaxTerm source nor the UTF-8
patch have useful inline comments, so I'm having a little trouble
following them. After a lot of unsuccessful shotgun debugging (which
is never really a good idea anyway, silly me) I finally resorted to
hacking out the HTTP compression in qweb.py and taking some before
and after packet captures with tcpdump. Right away I saw an obvious
difference...
The original AjaxTerm code performed inline replacement of all
spaces (really, anything from nul/0x00 up to and including
space/0x20, except lf/0x0a) with a non-breaking space (0xa0),
presumably to prevent some browsers from attempting to break lines
and word-wrap them. I can see evidence of this in my particular
application, where some ASCII art is getting mangled due to spurious
linebreaks. I was able to revert this regression by changing
span.encode('utf8') to span.replace(u' ',u'\xa0').encode('utf8') at
ajaxterm.py:326. After that, not only are the unwanted linebreaks
gone, but the cursor also once again looks like it should.
Also, somewhat less important, the ANSI bold attribute is no longer
shifting color numbers by one bit (0-7 -> 8-15). The ANSI escape and
color handling code, both in original AjaxTerm and the patched
version, is significantly opaque that it requires a bit more
meditation before I can unravel it and find a fix. More to come.
--
{ IRL(Jeremy_Stanley); PGP(9E8DFF2E4F5995F8FEADDC5829ABF7441FB84657);
SMTP(fu...@yuggoth.org); IRC(fu...@irc.yuggoth.org#ccl); ICQ(114362511);
AIM(dreadazathoth); YAHOO(crawlingchaoslabs); FINGER(fu...@yuggoth.org);
MUD(fu...@katarsis.mudpy.org:6669); WWW(http://fungi.yuggoth.org/); }
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
