From: Peter Graves <peter@ar...> - 2002-06-17 17:54:30
There have been three reports recently of problems saving files.
One of them was bug 568736, which involved j hanging while running "ls
-ld", which is done in ExtFile.getPermissions(), and which apparently
doesn't work over a SOCKS proxy.
I don't really know anything about SOCKS proxies, except that
apparently you can't run j over one, as things stand.
There are quite a few places where j does things like exec("ls -ld"),
and if that sort of thing doesn't work over a SOCKS proxy, a lot of
things are likely to fail, although j may appear to work correctly in
So bug 568736 isn't really a bug per se, so much as a Feature Request
for "SOCKS proxy support", which is not a bad idea, but it would
require an audit of all the places where j tries to do things that
won't work in a SOCKS proxy environment. In some cases a workaround
might be possible, and in others it might be necessary to lose some
functionality. It could be done, but it seems like a pretty big job,
and I'm not going to undertake it at this point.
The remaining two reports appear to be real bugs.
The first was "occasional lockups with file corruption (turned all
chars into zeros) on save (CTRL+S)", on Win2K with Java 1.4, with j
0.15.0, reported on May 23. No word of any recurrence since then.
The second is bug 569172:
After using J for an hour or two, I am no longer able
to save documents, and J refuses to close. The message
I get when trying to save is "Unable to write backup
file for <file>" which is followed by "Unable to save
<file>." I am able to close J if I kill a new process
that it brings up for saving.
I'm worried about these two, but there's not quite enough information
to pin down exactly what's happening.
The Win2K case was thought by the user reporting it to be related to
weirdness in his environment.
The Linux case seems like it might be some sort of resource leakage,
since it only occurs after an hour or two, but I can't reproduce it
here, and I've never seen anything like that myself, even though I
routinely run j on Linux for many hours straight. (The problem might
also be something quite mundane, like running out of disk space, or a
file or directory permission issue.)
If you experience a problem in this area, it would be very helpful if
you would enable logging by adding these lines to ~/.j/prefs (or
debug = true
log = true
With these settings, you'll get trace output to the console window or
xterm from which you started j, and in addition there will be a log
file called "log" created in ~/.j (or C:\.j). (It's actually a two-
generation log; when you restart j, "log" gets copied into "log.0".)
If something goes astray in the process of saving a file, there is
likely to be detailed information about it in the trace output and/or
For 0.15.3 I'll try to improve the error reporting in this area in the
default case where logging is not enabled.