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

Bug 337743

Summary:

Need to react to changes coming in ICU 4.4.2

Product:

[WebTools] Dali JPA Tools

Reporter:

David Williams <david_williams>

Component:

General

Assignee:

Neil Hauge <neil.hauge>

Status:

RESOLVED FIXED

QA Contact:

Severity:

normal

Priority:

CC:

neil.hauge, tranle1

Version:

unspecified

Target Milestone:

3.0 M6

Hardware:

OS:

Windows 7

Whiteboard:

Attachments:

Description	Flags
one compile errror	none
two compile error	none
patch to restore I20110215-0800 pre-req	neil.hauge: iplog+
Patch	none

Description David Williams

2011-02-21 12:48:22 EST

I tried a to move up to a platform build from I20110215-0800

But there were three compile errors related to source incompatibility changes in ICU 4.4.2 (previously we were on 4.2.1). 

See 

http://build.eclipse.org/webtools/committers/wtp-R3.3.0-I/20110221042833/I-3.3.0-20110221042833/

I opened bug 337699 in Orbit to track from that end, but from what I can tell (from quick read) it will take a change in Dali's source to be compatible. (Feel free to comment in bug 337699 if you disagree or have other suggestions)  

I'll attach logs in case that temporary build gets removed. 

I'll also attach patch to once again pre-req I20110215-0800 ... in case there is no solution that works in both cases.

Comment 1 David Williams

2011-02-21 12:48:53 EST

Created attachment 189427 [details]
one compile errror

Comment 2 David Williams

2011-02-21 12:49:17 EST

Created attachment 189428 [details]
two compile error

Comment 3 David Williams

2011-02-21 12:50:05 EST

Created attachment 189429 [details]
patch to restore I20110215-0800 pre-req

Comment 4 David Williams

2011-02-21 12:56:23 EST

BTW, I would appreciate if this could be fixed soon ... today? ... as I20110215-0800 is the first (good) build with Ant 1.8.2 in it (instead of 1.7.x) so wanted to be sure to get that in to integration builds early this week. 

Thanks,

Comment 5 Neil Hauge

2011-02-21 13:07:24 EST

David,

I've been trying to download the platform build so I can test this out locally, but have had real trouble getting builds down from eclipse.org recently.  My last attempt failed (the download just died after running for 45 min at about 1.4k per second).  Will try again now.  Assuming I can get the build downloaded we should have a fix for this very soon.

Comment 6 Neil Hauge

2011-02-22 10:57:30 EST

I've got a fix for this.  Will check in and release ASAP.

Comment 7 Neil Hauge

2011-02-22 12:02:24 EST

Created attachment 189516 [details]
Patch

The following patch has been committed to head.  Please release this with the patch to restore the new platform pre-req.  After releasing this bug should be marked fixed.

Comment 8 Tran Le

2011-02-22 13:19:45 EST

Dali's code in head has been released, and Releng patch applied.

David, I also fixed the EMF Transaction & EMF Validation pre-reqs display problem, please restart the build at your convenience. Thanks.

Comment 9 David Williams

2011-02-22 13:39:01 EST

I have restarted the build ... it was still on "jst-sdk" even after 3 hours ... so, something was going slow (wrong?) anyway. 

Sounds like there should be a lesson here about pre-mature use of generics ... but ... I suspect it'd be fairly esoteric. Luckily in this case, I assume, it did not percolate up to part of our API, or anything.

Comment 10 Neil Hauge

2011-02-22 14:01:32 EST

(In reply to comment #9)
> I have restarted the build ... it was still on "jst-sdk" even after 3 hours ...
> so, something was going slow (wrong?) anyway. 
> 
> Sounds like there should be a lesson here about pre-mature use of generics ...
> but ... I suspect it'd be fairly esoteric. Luckily in this case, I assume, it
> did not percolate up to part of our API, or anything.

Thanks.  There may be a number of lessons here.  I do think this is a rather unlikely scenario, and no, it did not change any API.  The usage of generics by the Collator implementations seem very odd indeed in this case, and hopefully this type of usage would not often present itself.  And yes, sadly, pre-mature use of generics is still possible 6 years after being introduced. : )