Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 143242

Summary: Set default encoding of CBE XML log parser to UTF-8
Product: z_Archived Reporter: Alex Nan <apnan>
Component: TPTP.monitoringAssignee: Rohit Shetty <rohit.shetty>
Status: CLOSED FIXED QA Contact:
Severity: normal    
Priority: P1 CC: apnan, jkubasta, labadie, Lina.Kemmel, rohit.shetty
Version: unspecifiedKeywords: plan
Target Milestone: ---   
Hardware: PC   
OS: Windows 2000   
Whiteboard:
Attachments:
Description Flags
Patch for this problem. rohit.shetty: review?

Description Alex Nan CLA 2006-05-23 11:51:12 EDT
Currently the CBE XML log import is defaulting to the native character encoding of the system where the parser is running and tries to use this value when parsing log files. Since most XML files are UTF-8 encoded the import will work fine for log files containing only ASCII characters. Logs containing other characters then ASCII will be incorrectly decoded and characters will show corrupted in the log view (as an example a CBE log file containing hebrew characters that was used for testing and reporting this defect).
Comment 1 Dave Smith CLA 2006-05-29 10:17:23 EDT
Will look at this again in 4.2 i4.
Comment 2 Dave Smith CLA 2006-06-06 15:34:37 EDT
Deferring this to 4.3 as I am concerned about changing the encoding.  On some platforms the encoding of a CommonBaseEvent XML file may not be UTF-8.  For example on z/OS the encoding will be EBCDIC.  More investigation of the implications is required before making the change.

A 4.2 readme item will be added for this problem.
Comment 3 Dave Smith CLA 2006-06-06 15:41:50 EDT
Created bugzilla 145612 to add readme item for this problem.
Comment 4 Dave Smith CLA 2006-10-24 17:19:46 EDT
Deferring to 4.4 as it is not a stop ship issue and cannot be contained in 4.3.
Comment 5 Dave Smith CLA 2007-01-17 23:49:19 EST
Added sizing.
Comment 6 Dave Smith CLA 2007-04-09 17:04:47 EDT
Did not make the 4.4 plan.  Moving to future.
Comment 7 Eric Labadie CLA 2007-08-09 10:51:19 EDT
This is required from AC perspective
Comment 8 Dave Smith CLA 2007-08-31 17:37:51 EDT
This can be fixed by changing the characterEncoding commonFields value in the logParser extension for the CommonBaseEvent parser in org.eclipse.hyades.logging.parsers plugin.xml from "DEFAULT" to "UTF-8".  For example:

<commonFields characterEncoding="UTF-8"/>

I am reducing the sizing for this work to 16 hours because it is mostly a testing effort to ensure the fix does not cause a regression in both the local and remote import cases and with older Agent Controllers. 
Comment 9 Alex Nan CLA 2007-10-12 13:59:17 EDT
Reassigning to Rohit.
Comment 10 Alex Nan CLA 2007-10-30 19:43:01 EDT
Targetting 4.5 i2.
Comment 11 Rohit Shetty CLA 2007-11-03 07:32:07 EDT
Created attachment 82030 [details]
Patch for this problem.

Verified, that this works in the above described scenario's.
Comment 12 Alex Nan CLA 2007-11-05 18:50:35 EST
Targetting to 4.5 i4.
Comment 13 Alex Nan CLA 2007-11-14 18:25:15 EST
Patch reviewed. Looks OK. Please check in once i4 is open for development.
Comment 14 Rohit Shetty CLA 2007-11-19 07:39:28 EST
Fixed, Code is in CVS.
Comment 15 Alex Nan CLA 2008-06-27 20:18:36 EDT
Verified with TPTP-4.5.0 GA.
Comment 16 Alex Nan CLA 2008-06-27 20:18:47 EDT
Closing.