Issue 96090

Summary: Increasing memory usage when opening several presentations
Product: Impress Reporter: odd_bloke <daniel.watkins>
Component: programmingAssignee: AOO issues mailing list <issues>
Status: CONFIRMED --- QA Contact:
Severity: Trivial    
Priority: P3 CC: caolanm, hdu, issues
Version: OOO300m9   
Target Milestone: ---   
Hardware: PC   
OS: Unix, all   
Issue Type: DEFECT Latest Confirmation in: ---
Developer Difficulty: ---
Issue Depends on: 97906, 102142    
Issue Blocks:    
Attachments:
Description Flags
Python script to reproduce issue
none
current OOo
none
patched OOo
none
slightly tweaked script to bump xrestop results for OOo after each sleep none

Description odd_bloke 2008-11-11 12:15:04 UTC 
When opening and displaying several presentations, the memory for previous
presentations is not freed, resulting in massive memory usage over a relatively
short period of time (i.e. 24 hours).

The attached Python script will show an increase in memory used by X, because
OpenOffice isn't cleaning up after itself.  The easiest way to see the increase
is using 'xrestop'.
Comment 1 odd_bloke 2008-11-11 12:16:59 UTC 
Created attachment 57870 [details]
Python script to reproduce issue
Comment 2 wolframgarten 2008-11-11 12:53:19 UTC 
Reassigned. Please have a look. Thanks.
Comment 3 clippka 2009-06-03 12:18:20 UTC 
cl->hdu: is that an x server issue?
Comment 4 hdu@apache.org 2009-06-08 08:40:30 UTC 
@cmc: you have this nice tool xresgrind that could answer cl's question...
Other than that this issue reminds me of issue 97906 (X resource usage in presentation)
Comment 5 caolanm 2009-06-08 12:00:11 UTC 
The majority of this is definitely 102142 anyway, i.e. I'll attach results
before/after that patch wrt usage. So I'd call it a duplicate of that, as the
runaway effect can be seen in the attached results.

(With patch for 102142 in place its still a little hard with the tooling to
determine if there's another continual leak, or a little bit of bouncing around
a stable cache position. Tooling can't distinguish between pixmaps/windows that
are definitely leaked or those that are stuck into a endless-growing-cache or
something like that where there are "effectively" leaked, but get released on
exit so aren't "definitely" leaked.)
Comment 6 caolanm 2009-06-08 12:01:06 UTC 
Created attachment 62847 [details]
current OOo
Comment 7 caolanm 2009-06-08 12:01:23 UTC 
Created attachment 62848 [details]
patched OOo
Comment 8 caolanm 2009-06-08 12:05:38 UTC 
sorry, those are transposed, but summary here..., i.e.

Before:
start...
        windows       : 40
        pixmaps       : 96
        pictures      : 29
        unknowns      : 26
end...
        windows       : 58
        pixmaps       : 120
        pictures      : 36
        unknowns      : 44

After:
start...
5 - OpenOffice.org ( PID:18290 ):
        windows       : 40
        pixmaps       : 96
        pictures      : 29
        unknowns      : 26
end...
        windows       : 40
        pixmaps       : 96
        pictures      : 29
        unknowns      : 26
Comment 9 hdu@apache.org 2009-06-08 12:16:22 UTC 
@cmc: wonderful, thank you!

@af: as Caolan pointed out issue 102142 is the most probable root cause so this issue is a dupe. The test 
script here is too good to just close this issue too soon.
Comment 10 caolanm 2009-06-08 12:34:06 UTC 
Created attachment 62852 [details]
slightly tweaked script to bump xrestop results for OOo after each sleep
Comment 11 groucho266 2009-09-25 10:31:45 UTC 
Setting target to OOo 3.3.
Comment 12 groucho266 2010-07-13 16:00:50 UTC 
Changing target.
Comment 13 Martin Hollmichel 2011-03-15 11:06:38 UTC 
set target to 3.x since they are not release relevant for 3.4
Comment 14 Marcus 2017-05-20 11:11:17 UTC 
Reset assigne to the default "issues@openoffice.apache.org".