[rkward-users] [Help] RE: Rkward windows do not refresh

SourceForge.net noreply at sourceforge.net
Wed May 25 11:28:50 UTC 2011


The following forum message was posted by tfry at http://sourceforge.net/projects/rkward/forums/forum/165574/topic/4523913:

Hi!

I'm afraid, I'm going to have to give up on this, soon. A few more things, before
I do:
1) Could you perhaps store a backup of the affected workspace/workplace and
all related files, somewhere? In case the problem goes away one day, it would
be good to be able to verify whether it was actually fixed, or whether you simply
no longer trigger it, then.
2) Is flux1_log.txt very large? Could there be anything else unusual about this
file?
3) Suppose you copy flux1_log.txt to an internal disk, instead of a USB drive
(or whatever it is), open it from there, instead, save the workplace, and restart
RKWard with that? Same issue?

Regarding the errors you cite in comment #11: Yes, these are almost certainly
unrelated. In case you copied the lines for the .rkworkplace-file, manually,
with a text-editor, please check for any trailing or leading spaces, and make
sure that there is a hard line break after each line. ATM, I can't think of
any other possible source of error. This [i]should[/i] work (I tried, here,
replacing flux1_log.txt for a different file, but did not manage to reproduce
the bug).

Some more loud thinking:
- You said that the workspace is very big, and suggested a memory issue. I think
this is very unlikely. You would get much less confined symptoms in this case.
- Of course, a race condition comes to mind. But I can't make any sense of this,
either. Only after the R workspace has been fully loaded, and all associated
events have been processed, RKWard goes on with restoring the .rkworkplace.
So that's guaranteed to be sequential.
- ATM, I'm suspecting a problem in KDE / Qt, rather than RKWard code, but I
can't find any pointers on that, either.

Regards
Thomas




More information about the Rkward-users mailing list