[rkward-devel] workspace browser does not list variables in amd64 arch

[Thomas Friedrichsmeier]

Hi,

On Monday 01 October 2007, Prasenjit Kapat wrote:
> I really hate to bring this up, when we are so close to releasing
> 0.4.8, but unfortunately, I was too involved with school...  I think
> we can go ahead with releasing 0.4.8 since most users are on 32 bit
> machiens (i guess).

ok, well, if we can't fix it in time, we'll release with the bug. But let's 
give it a try.

Unfortunately, I do not have access to a 64bit system, so I'll need your help 
with this. Anybody else have a 64bit system at hand? If so, please join in!

> > 2) Do you get commandline completion (console/script editor) for objects
> > in globalenv()?
>
> No.

Ok, so this happens at a rather basic level. Probably the workspace browser 
itself is not to blame, rather it looks like fetching information about the 
objects in the workspace fails.

> > 3) You mention, it's been there for a long time. Any idea, when precisely
> > it first turned up?
>
> I guess it has always been there in the 64 bit arch installations. I
> moved to a native 64 bit machine around July, and this issue has been
> there since then. I did not (and still do not) see it in 32 bit
> machines.

Ok, but I take it you used the SVN version? Could you perhaps try with 0.4.7 
and/or 0.4.6, and see whether those give the same result? Otherwise, this 
would narrow it down quite a bit. Also, could you try turning off all 
optimization (./configure with --enable-debug=full should do the trick), just 
to make sure?

Some more questions:
Does the data editor work all right? Does it work at all?

> Some more details:
> The occurence is not predictable. I will try to explain what I mean:
> (with attachments)

Hm, those unpredictable ones are the worst. Probably somewhere we're somehow 
relying on a variable, internally, which is 64bit long on a 64bit system with 
only the lower 32 bits zeroes out, and the rest random. Unfortunately, I have 
no idea, *where* this happens.

Let's try the usual procedure: Please run with
# rkward --debug-level 5 --debug-flags 132
and post the output somewhere (despite limiting the output 
using --debug-flags, this will create a *lot* of output).

If you have the time, please install valgrind, and run rkward through it 
(i.e. "valgrind /path/to/rkward"), also posting the tons of output somewhere 
(if you have the choice to put it online somewhere, that would be great, else 
send it to the list in zipped form).

Regards
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20071001/0d62245f/attachment.sig>