[ 
http://jira.codehaus.org/browse/GEOT-2538?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Andrea Aime reopened GEOT-2538:
-------------------------------


So, it turns out I messed up and committed and incomplete patch, and thus the 
problem is still there :(

> GEOT-2073  fix intoduced severere performance degradation due to open SUN Bug 
> ID:6568969
> ----------------------------------------------------------------------------------------
>
>                 Key: GEOT-2538
>                 URL: http://jira.codehaus.org/browse/GEOT-2538
>             Project: GeoTools
>          Issue Type: Bug
>          Components: core render, core styling
>         Environment: WinXP, java 1.6
>            Reporter: Per S Hustad
>            Assignee: Andrea Aime
>             Fix For: 2.5.6
>
>         Attachments: patch-stroke-bug.txt, StrokeBugDemo.java
>
>
> The Stroke fix for GEOT-2073 introduced in 2.5.2, may cause severe rendering 
> degradation. This should be seen in conjunction with [SUN Bug 
> ID:6568969|http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6568969] 
> reporting that drawing long dashed lines takes extremely long time.
> Drawing very long lines will typically happen when zooming far into a 
> relatively large feature. 
> Before the 2073 patch, plain solid SLD strokes  were just converted to 
> {{java.awt.BasicStroke}} using a {{null  dash}} array. However, after the 
> 2073 patch, a dash array of {{[1.0, 0.0]}} is passed to the   {{BasicStroke}} 
>  constructor if the original stroke is solid. Running the attached test 
> program shows the output from drawing a very long text line on my 8 core 
> Intel Xeon 3.37 Ghz machine with an ATI FireGL V7350 graphics card:
> {noformat}
> Drawing with dash= [1.0, 0.0] took 14570.986102 ms 
> Drawing with dash= null took 0.326722 ms
> {noformat}
> I.e. the difference in performance is immense. Note also there is no 
> workaround for this. Even if you specify a solid stroke, geotools will 
> automatically
> create a dummy [1,0] stroke for you. 
> We have created a patch for the 2.5.5 version. The patch substitutes any 
> {{[1.0, 0.0]}} dash array with a null argument before constructing the 
> BasicStroke. All existing unit tests in the gt-render module pass with the 
> patch installed and our application performs extremely much better when 
> zooming in.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: 
http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to