[
https://issues.apache.org/jira/browse/FLEX-34769?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15045169#comment-15045169
]
Harbs commented on FLEX-34769:
------------------------------
For comparison, here's the old behavior (using tlf 3.0 from the Apache repo):
1. Composes lines 0-25. (exactly like current version although I just got 67ms
instead of 200ms for this stage. It might not be an exact comparison)
2. scroll to range composes line 25 through 6299 and takes 1509ms to complete.
3. verticalScrollPosition is set, but recompose here.
4. triggers a recompose of lines 6292 through 6299.
5. set verticalScrollPosition causes a recompose of lines 6299 and 6300.
So:
1 seems OK.
Step 2 should compose from the current point where step 1 left off and compose
to the end.
I have a feeling this will all be resolved if we remove the resetting of the
start composition which was committed by Adobe in the donation. I'll work on
that.
> TLF Performance issue (applyLeafFormat method)
> ----------------------------------------------
>
> Key: FLEX-34769
> URL: https://issues.apache.org/jira/browse/FLEX-34769
> Project: Apache Flex
> Issue Type: Bug
> Components: TLF
> Affects Versions: Apache Flex 4.10.0, Apache Flex 4.11.0, Apache Flex
> 4.12.0, Apache Flex 4.13.0, Apache Flex 4.14.0
> Environment: OS: Windows 7
> AIR version: 15 and 16
> CPU: Intel i7 2630h
> RAM: 4GB
> Reporter: goratz
> Assignee: Harbs
> Labels: flex, performance, tlf
> Fix For: Apache Flex 4.9.0
>
> Attachments: TLFPerfTest.as, TLF_Bench.fla, textLayout_14.swc,
> textLayout_9.1.swc
>
>
> I have a very strange performance difference between 2 versions of
> TextLayout.swc. The first one is the swc included in Apache Flex 4.14 SDK and
> the second one is included in Apache Flex 4.9.1.
> I have a very large text, more or less 200 DIN-A4 pages in a Textflow. I
> scroll to the end of the document and I underline a word. This is my result
> underling a word in both versions.
> Flex SDK Version Time
> 4.9.1 21 ms (miliseconds). 0,02 Seconds
> 4.14 3.405 ms (miliseconds). 3,4 Seconds
> The time difference happens in this line of code:
> EditManager.applyLeafFormat(oFormat);
> Anyone has any idea what's the problem?
> More info:
> https://issues.apache.org/jira/browse/FLEX/?selectedTab=com.atlassian.jira.jira-projects-plugin:issues-panel
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
