[ 
https://issues.apache.org/jira/browse/GUACAMOLE-884?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16940602#comment-16940602
 ] 

Mike Jumper commented on GUACAMOLE-884:
---------------------------------------

Yes, I did as well. In my case, Firefox stayed level at 7.7 MB. Either way, 
that's not really the point. The facts are as follows:

* There exists a bug in some browsers where instances of {{Image}} will leak.
* In cases where this leak occurs, it is the memory of the browser/tab process 
that increases. The JavaScript heap usage remains unaffected.
* There are ways to draw received image data without {{Image}}. In the past, 
tests have shown those methods to be significantly slower than {{Image}} and 
data URIs.

With the availability of {{createImageBitmap()}}, I am checking whether this is 
now doable without a performance hit, in which case we can do that and simply 
fall back to data URIs if {{createImageBitmap()}} is unavailable. If there _is_ 
a performance hit, then we cannot make this change.

Either way, as it can be shown that there is no leak within the JS, I highly 
recommend reporting the leak itself to the groups maintaining the affected 
browsers.

> Canvas Render Browser Memory Leakage
> ------------------------------------
>
>                 Key: GUACAMOLE-884
>                 URL: https://issues.apache.org/jira/browse/GUACAMOLE-884
>             Project: Guacamole
>          Issue Type: Bug
>          Components: guacamole-common-js
>    Affects Versions: 1.0.0
>         Environment: guacamole 1.0
>            Reporter: jianfeng chen
>            Priority: Minor
>         Attachments: 001.png, CorrectCode.png, firefox.png, guac-js-heap.png, 
> pysh.py
>
>
> In SSH protocol, executing Python scripts will constantly render and refresh 
> the interface, resulting in a gradual increase in browser memory, and finally 
> browser crash.
> The script content is in the attachment.
> Execute script instructions:python pysh.py 1024 1



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to