FTBFS against openjdk-6 6b23 due to internal Sun API changes.

Bug #800082 reported by James Page
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenJDK
Fix Released
Medium
batik (Ubuntu)
Fix Released
Medium
James Page
openjdk-6 (Ubuntu)
Fix Released
Medium
Unassigned
xmlgraphics-commons (Ubuntu)
Fix Released
Medium
James Page

Bug Description

batik makes use of internal sun(oracle) API's for creation of TIFF and JPEG images.

com.sun.image.codec.jpeg.ImageFormatException prior to 6b23 extended RuntimeException which means that code does not need to throw or catch this exception; however in 6b23 this class now extends Exception which means that it now need to be caught or declared as thrown by calling code.

James Page (james-page)
Changed in batik (Ubuntu):
status: New → Confirmed
status: Confirmed → In Progress
assignee: nobody → James Page (james-page)
importance: Undecided → Medium
Revision history for this message
In , James Page (james-page) wrote :

Operating System:
 OS - Ubuntu Development Release - Oneiric Ocelot

java -version:
 java version "1.6.0_22"
 OpenJDK Runtime Environment (IcedTea6 1.11pre) (6b23~pre2-0ubuntu2)
 OpenJDK 64-Bit Server VM (build 20.0-b11, mixed mode)

Prior to 6b23 com.sun.image.codec.jpeg.ImageFormatException extended RuntimeException rather than Exception so it did not need to be caught or declared as throw in code that was using this internal API.

>= 6b23 is class extends Exception - this has caused a couple of build failures in packages that use this package - see linked bug report in Ubuntu.

This change was introduced with changeset http://icedtea.classpath.org/hg/icedtea6/rev/8720a7df46e3?revcount=120

It would appear that this makes the IcedTea6 implementation API different from the original Java 1.5 API for this class.

Changed in xmlgraphics-commons (Ubuntu):
importance: Undecided → Medium
Changed in openjdk-6 (Ubuntu):
importance: Undecided → Medium
Changed in xmlgraphics-commons (Ubuntu):
status: New → In Progress
assignee: nobody → James Page (james-page)
Revision history for this message
James Page (james-page) wrote :

This API change also impacts xmlgraphics-commons

Changed in openjdk:
importance: Unknown → Medium
status: Unknown → Confirmed
Revision history for this message
In , Andrew John Hughes (ahughes) wrote :

I'll look into fixing this. Note that the version concerned is a pre-release of IcedTea6 1.11 and not '6b23' which confused me at first (I thought you were referring to a proprietary Oracle JDK).

What disturbs me is that this is not yet in a release of IcedTea6 yet, but you seem to have acquired it via an Ubuntu package.

Revision history for this message
Andrew John Hughes (ahughes) wrote :

I'll look into this in IcedTea, but can someone please explain why Ubuntu is shipping an unreleased version of IcedTea6?

Revision history for this message
In , James Page (james-page) wrote :

(In reply to comment #1)
> I'll look into fixing this. Note that the version concerned is a pre-release
> of IcedTea6 1.11 and not '6b23' which confused me at first (I thought you were
> referring to a proprietary Oracle JDK).

Thanks for looking at this and apologies for the confusion - this is probably my ignorance of how IcedTea and OpenJDK/OracleJDK related to each other - I was purely going on the version number of the Ubuntu package.

> What disturbs me is that this is not yet in a release of IcedTea6 yet, but you
> seem to have acquired it via an Ubuntu package.

This is in the latest development release for Ubuntu; I guess that having the pre release of 1.11 helps flush out this sort of bug.

Revision history for this message
In , Andrew John Hughes (ahughes) wrote :

IcedTea builds a specific OpenJDK build drop with additional patches and extensions to make it usable. This ID is even more confusing because there is no b23 of OpenJDK6 yet. Both IcedTea6 1.10 and the forthcoming 1.11 are still using b22. So I don't know where this 'b23' is coming from...:-S

I ran a full comparison:

http://fuseyism.com/japi/jpeg-icedtea6-oracle.html

and fixed up some other differences too. As far as I can tell, the exception reports in the above are either wrong or undocumented. The 1.5 docs for the package state that these exceptions are thrown. All the other issues I've fixed and I'm building the result now. Will post here once the commit is in.

If this is an Ubuntu pre-release, then it's understandable but it would still be good of them to keep us in the loop. It shouldn't go out in a released Ubuntu as we only provide security updates for releases, not random snapshots.

Revision history for this message
In , Andrew John Hughes (ahughes) wrote :

http://fuseyism.com/japi/jpeg-oracle-icedtea6.html is the comparison the other way round (Oracle as the standard which IcedTea6 is implementing).

Revision history for this message
In , Andrew John Hughes (ahughes) wrote :
Changed in openjdk:
status: Confirmed → Fix Released
James Page (james-page)
Changed in batik (Ubuntu):
status: In Progress → Fix Released
Changed in xmlgraphics-commons (Ubuntu):
status: In Progress → Fix Released
Changed in openjdk-6 (Ubuntu):
status: New → Fix Released
Revision history for this message
In , James Page (james-page) wrote :

Hi Andrew

Thanks for fixing this issue; this change has now landed in the Ubuntu development release and has resolved the issues that I was having with this API.

Cheers

James

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.