incubator-callback-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wim Lingier (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CB-107) Possible socket leaking in CallbackServer
Date Mon, 19 Dec 2011 09:56:30 GMT

    [ https://issues.apache.org/jira/browse/CB-107?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13172155#comment-13172155
] 

Wim Lingier commented on CB-107:
--------------------------------

I tested https://gist.github.com/1446383 on two devices:

A HTC Desire A8181 where the original error occured: 968 lines were appended before the next
error;



D/PhoneGapLog(21565): file:///android_asset/www/index.html: Line 35 : write success

I/Web Console(21565): write success at file:///android_asset/www/index.html:35

E/dalvikvm-gc(21565): Could not create 1196032-byte ashmem mark stack: Too many open files

E/dalvikvm-heap(21565): dvmHeapBeginMarkStep failed; aborting

E/dalvikvm(21565): VM aborting

I/ActivityManager(   96): Process com.phonegap.example (pid 21565) has died.

I/WindowManager(   96): WIN DEATH: Window{463587c8 com.phonegap.example/com.phonegap.example.PhoneGapExample
paused=false}

On A Sony-Ericsson WT19i there were no problems, so probably it is a HTC/Sense problem?
                
> Possible socket leaking in CallbackServer
> -----------------------------------------
>
>                 Key: CB-107
>                 URL: https://issues.apache.org/jira/browse/CB-107
>             Project: Apache Callback
>          Issue Type: Bug
>          Components: Android
>    Affects Versions: 1.2.0
>            Reporter: Simon MacDonald
>            Priority: Critical
>
> I think we may be leaking a socket in CallbackServer. On line 216 of the CallbackServer
we create a new Socket but I can't see anywhere in CallbackServer where the Socket is closed.
Over time this could cause the problem described below. The user is probably running into
it much sooner than usual as he's making a file write every second.
> Here is the original bug report from the Google Groups:
> I'm trying to write approx every second to a file. This for writing a
> geolocation to this file.
> So I make a filewriter when I start tracking, then start a
> watchposition en in the successfunction I write to this file.
> In my first tests everything went well, but with some more intensive
> testing we found out that after more than 900 points (or
> filewritervar.write() callings) the app fails. In the logcat the next
> error is vissible:
> [code]
> W/System.err(27919): java.net.SocketException: Too many open files
> W/System.err(27919):    at
> org.apache.harmony.luni.platform.OSNetworkSystem.acceptSocketImpl(Native
> Method)
> W/System.err(27919):    at
> org.apache.harmony.luni.platform.OSNetworkSystem.accept(OSNetworkSystem.java:
> 69)
> W/System.err(27919):    at
> org.apache.harmony.luni.net.PlainSocketImpl.accept(PlainSocketImpl.java:
> 126)
> W/System.err(27919):    at
> java.net.ServerSocket.implAccept(ServerSocket.java:281)
> W/System.err(27919):    at
> java.net.ServerSocket.accept(ServerSocket.java:165)
> W/System.err(27919):    at
> com.phonegap.CallbackServer.run(CallbackServer.java:216)
> W/System.err(27919):    at java.lang.Thread.run(Thread.java:1102)
> [/code]
> I looked in the FileUtils.java file and the filestream is everytime
> correctly closed, so I dont see the problem.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message