Ok. Give me some days to find the time to compile from the source and do the
tests.
On Wednesday, October 05, 2011 06:03:44 PM Tommaso Cucinotta wrote:
> Il 05/10/2011 22:54, Tommaso Cucinotta ha scritto:
> > 2) try to recompile LyX from the sources, to check whether the problem
> > still shows
Il 05/10/2011 22:54, Tommaso Cucinotta ha scritto:
2) try to recompile LyX from the sources, to check whether the problem
still shows up on your system or not in the current trunk. Shortly
(but it's going to take half an hour or so):
svn co svn://svn.lyx.org/lyx/lyx-devel/trunk lyx-trunk
cd
Il 05/10/2011 20:37, Rudi Gaelzer ha scritto:
All right, I installed valgrind and ran it as per instructions.
When it finally opened the text, the process massif-amd64-li... was
using 13.5%MEM and the demand was increasing very slowly without any
operation on the text.
When I opened the A
Le 04/10/2011 16:33, Tommaso Cucinotta a écrit :
I have some doubts about how UndoElement instances are handled.
a) they're pushed into the stack *by value*
std::deque c_;
so, more instances of UndoElement are created and copied wastefully:
void push(UndoElement const & v) {
Le 04/10/2011 16:33, Tommaso Cucinotta a écrit :
Thanks for looking at these. However, one of the biggest mess seems to
me the Undo related logic.
Let me report below my original question/doubts:
It look like I missed the original posting on this subject.
The questions look very interesting, I'
Il 04/10/2011 15:28, Jean-Marc Lasgouttes ha scritto:
Le 01/10/2011 02:04, Tommaso Cucinotta a écrit :
These are just cleanup ops that are not explicitly done on program exit,
but not a big deal, actually. They could help in having a better cleanup
of LyX at exit, but I'm not sure any of these d
Le 01/10/2011 02:04, Tommaso Cucinotta a écrit :
These are just cleanup ops that are not explicitly done on program exit,
but not a big deal, actually. They could help in having a better cleanup
of LyX at exit, but I'm not sure any of these deserves to be committed
(pls, comment).
I think such
Il 01/10/2011 19:01, Rudi Gaelzer ha scritto:
On Friday 30 September 2011 21:04:35 Tommaso Cucinotta wrote:
> For the increased memory usage notified by Rudi, if there were any
> problem with handling the Undo instances, then we should have seen
> something about Undo into the valgrind log. I
Il 30/09/2011 12:15, Jean-Marc Lasgouttes ha scritto:
What do we see?
1/ there is a continuous drift of memory use
2/ the most important part is the agregation of small allocations (not
very useful...)
3/ the second one is related to font loading by qt (remains stable)
4/ the third one is the
Le 30/09/2011 08:56, Tommaso Cucinotta a écrit :
Ok, please, find attached the obtained output parsed with ms_print.
Now, what do I do with this monster :-) ?
Where's the info about memory leaks ?
It is not necessarily a problem of memory leak (could be an ever-growing
cache, for example). Her
Le 29/09/11 01:13, Tommaso Cucinotta a écrit :
==26690== LEAK SUMMARY:
==26690== definitely lost: 12,126 bytes in 73 blocks
==26690== indirectly lost: 18,336 bytes in 571 blocks
==26690== possibly lost: 696,631 bytes in 2,590 blocks
==26690== still reachable: 898,279 bytes in 7,833 blocks
==26690
On Wednesday 28 September 2011 19:05:28 Tommaso Cucinotta wrote:
> Il 28/09/2011 20:40, Rudi Gaelzer ha scritto:
> > The document I'm working right now has thousands of formulas and I've
> > been using the advanced search pretty heavily. I can detect the
> > increasing memory load even with the ver
Il 29/09/2011 00:30, Jean-Marc Lasgouttes ha scritto:
Le 29/09/2011 00:05, Tommaso Cucinotta a écrit :
Also, I noticed a monotonic memory usage increase while editing the
document and especially creating math insets and deleting them later. I
guess this is due to the Undo feature, that keeps sto
Le 29/09/2011 00:05, Tommaso Cucinotta a écrit :
Also, I noticed a monotonic memory usage increase while editing the
document and especially creating math insets and deleting them later. I
guess this is due to the Undo feature, that keeps storing (a copy of)
all the insets even though you delete
Il 28/09/2011 20:40, Rudi Gaelzer ha scritto:
The document I'm working right now has thousands of formulas and I've
been using the advanced search pretty heavily. I can detect the
increasing memory load even with the very simple text that I attached
(memory.lyx). I was continually searching for
On Wednesday 28 September 2011 15:11:02 Tommaso Cucinotta wrote:
> Il 28/09/2011 16:36, Rudi Gaelzer ha scritto:
> > I don't know if this has already been reported. I made a quick scan
> > through both the user and devel lists and couldn't find anything related.
> >
> >
> > While editing a text
Il 28/09/2011 16:36, Rudi Gaelzer ha scritto:
I don't know if this has already been reported. I made a quick scan
through both the user and devel lists and couldn't find anything related.
While editing a text riddled with mathematical formulae, I had to use
advanced search several times to
17 matches
Mail list logo
