Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-06 Thread Morten Welinder
The hang has been fixed upstream.  It was (and the as-values
thing mostly was) a matter of taking an extremely long time,
not a hang per se.

Morten



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread Morten Welinder
The as values problems are fixed upstream.

Morten



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread J.H.M. Dassen (Ray)
On Thu, Nov 04, 2010 at 11:21:07 -0400, Morten Welinder wrote:
 The as values problems are fixed upstream.

Specifically:
http://git.gnome.org/browse/gnumeric/commit/?id=bca4b6ccadbad690f17d28ad710ff67bed3e6f14
plus
http://git.gnome.org/browse/gnumeric/commit/?id=9ee41a9620aa902e3d7d1658983afa3a075e16b2

although Morten notes
15:50 @gmorten fixing that is easy (and done) but it doesn't work for array
 formulas.
-- 
Brought to you by Microsoft .NOT technology: just say No.



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread Morten Welinder
I am unable to reproduce the ods save issue.  It does use a lot of cpu
for a short
while, but then finishes as expected.  Your stack trace suggests that it did
finish.

I am unclear on how to reproduce the the BadAlloc crash.  Could you please spell
out the steps in detail?  I.e., what is selected and what is dragged.

If I read things right, you are using gtk+ 2.20.x  Around that time,
gtk+ underwent
serious changes in its window handling switching to client side
windows.  It was quite
buggy for a while.  Is it possible for you to upgrade gtk+?

Thanks,
M.



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread Andres Cimmarusti
 I am unable to reproduce the ods save issue.  It does use a lot of cpu
 for a short
 while, but then finishes as expected.  Your stack trace suggests that it did
 finish.

I had to press Ctrl-C and then used to the command bt to obtain the
trace. I then promptly exited gdb and killed gnumeric with it. I had
to do this because the saving process was taking a very very long
time. I'll give it another try.

 I am unclear on how to reproduce the the BadAlloc crash.  Could you please 
 spell
 out the steps in detail?  I.e., what is selected and what is dragged.

I tried to simplify the steps (skip the fourier analysis part) and got
the same issue:
1. Open the ods file attached
2. In an adjacent column type: =imabs( complex ( B2 , C2 ) )
3. Drag downward about 2000 cells
4. Now select the whole new column, i.e. go to the first cell then
hold shift, press end and then the downward arrow.
5. Drag this column selection to the right
6. BadAlloc crash...

 If I read things right, you are using gtk+ 2.20.x  Around that time,
 gtk+ underwent
 serious changes in its window handling switching to client side
 windows.  It was quite
 buggy for a while.  Is it possible for you to upgrade gtk+?

You are correct, I'm using gtk+ 2.20.1, which is in the standard
debian repositories.
I'm hesitant to upgrade a core library in my Squeeze system for one in
the experimental repository (gtk+ 2.22.0), but if you are not able to
reproduce the issue after my detailed explanation then I'll give it a
try.

Thanks

Andres



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread Morten Welinder
Ok, got it.  Fix will land upstream shortly.  The problem is that
the dragging of the little square (as opposed to dragging of the
data) is creating a crazy-tall tooltip that gtk+ cannot handle.
You can use copy+paste instead.

Note: you are better off using HYPOT over IMABS+COMPLEX.
The latter creates a string and some kind of rounding is therefore
involved.



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread J.H.M. Dassen (Ray)
On Thu, Nov 04, 2010 at 16:26:27 -0400, Morten Welinder wrote:
 Ok, got it.  Fix will land upstream shortly.

Landed as
http://git.gnome.org/browse/gnumeric/commit/?id=f40daf97c2c369dce9572b93feaf896ceb47a30c
-- 
First they ignore you.
Then they laugh at you.
Then they fight you.
Then you win.- Gandhi



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#602257: Debian 602257 -- Gnumeric crashes constantly

2010-11-04 Thread Andres Cimmarusti
 I am unable to reproduce the ods save issue.  It does use a lot of cpu
 for a short
 while, but then finishes as expected.  Your stack trace suggests that it did
 finish.

Ok, so these are the steps that cause my cpu usage to soar to 100% and
stay there until I kill the process when saving to an existing ods
file:

1. Open ods file originally attached
2. Open Fourier analysis ( check labels )
3. Select all the data EXCEPT the first column
4. Do fourier analysis
5. In one of the empty sheets (say sheet3)
6. Do: =hypot( fourier output real column 1 , fourier output imaginary
column 2 ) (the IMABS + COMPLEX also works to get the problem)
7. Drag this down till you encompass all the data in the fourier output colums
8. Try to save by overwriting the ods file you are working on (option
- with foreign extras)
9. Takes forever to save and I have to kill the process...

Note that I'm using gnumeric on a 5-year old laptop that might be not
good enough to do the job quicklyfor some reason..



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org