Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [orbit-dev] Orbit promotions and "release" plans

this would be _so_ much easier if you tossed out this whole I/M/R stuff and produced _one_ repository that was durable that I could depend on a particular version of a particular artifact being in for all time

*grumble*
jesse

--
jesse mcconnell
jesse.mcconnell@xxxxxxxxx


2011/11/16 David M Williams <david_williams@xxxxxxxxxx>

> Are there any plans yet for a new recommended release?

There is a plan for one for Indigo SR2 (February, 2011), but that will contain only minimal changes from previous one, to avoid risk. Only one or two bundles have changed, so far. These are currently our "M builds" and get renamed to an "R build" in February.

But, perhaps concept of an "R build" is still not clear ... we only produce those when we have a set of things we know we want to live forever (So to speak ... the webmasters hate it when I talk about files/builds/repositories lasting forever :). I think the way to think of it, is that R-builds are produced (primarily) for adopters that might depend on those exact versions for years and years. So, if anyone has a need for an "off cycle" R-build, they can certainly request it and we can discuss it here on this list ... otherwise, the next R build for "new things" is planned for June, 2012.

But, I plan to promote an I build later these week (or, earlier, if requested) and eventually our I-builds will be renamed to an S build for Juno M4, on December 2. The idea is that projects developing for Juno, that need or want to test some new Orbit bundle, can move to an I build, when its ready ... and then they (and everyone) should move to the latest S-build as they prepare their M4 deliverables. I hope this satisfies what you need for "new bundles since May".

Hope that helps ... if not, please ask again.


Inactive hide details for "Wetzold, Robert" ---11/16/2011 10:49:23 AM---Hi David, this did the trick. Great! We now have quite "Wetzold, Robert" ---11/16/2011 10:49:23 AM---Hi David, this did the trick. Great! We now have quite some new bundles in Orbit since May. There is

From: "Wetzold, Robert" <robert.wetzold@xxxxxxx>
To: David M Williams/Raleigh/IBM@IBMUS, "orbit-dev@xxxxxxxxxxx" <orbit-dev@xxxxxxxxxxx>,
Date: 11/16/2011 10:49 AM
Subject: RE: [orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed





Hi David,

this did the trick. Great! We now have quite some new bundles in Orbit since May. There is only xstream missing from my side. Are there any plans yet for a new recommended release?


Best regards, Robert

From: David M Williams [mailto:david_williams@xxxxxxxxxx]
Sent:
Mittwoch, 16. November 2011 14:42
To:
orbit-dev@xxxxxxxxxxx
Cc:
Wetzold, Robert
Subject:
RE: [orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed

> will be listed as “bundle.name” and not with their real name?

If you use the old-fashioned "plugin.properties" file, be sure to check you have

Bundle-Localization
: plugin
in the manifest.mf file.

If you use the "OSGi standard" (default) structure for NL properties, where you have a directory/file named
OSGI-INF/l10n/bundle.properties
then be sure you do
not have the header
Bundle-Localization
: plugin

In either case, check that the "variable name" in the manifest.mf file, such as

Bundle-Name
: %bundle.name
exactly matches the name used in the plugin.properties (or, bundle.properties file)

bundle.name
= Real Bundle Name
(without the '%' character)


If that doesn't help, let me know and I'll take a look at the actual code on cvs to see if I see anything else.
Or, perhaps others have tips on what to check?


Glad to "odd failure" turned out to be some fluke that "fixed itself" :)


Thanks,




Inactive hide details for "Wetzold, Robert" ---11/16/2011 08:15:40 AM---Hi David, thanks for your explanations. As you thought "Wetzold, Robert" ---11/16/2011 08:15:40 AM---Hi David, thanks for your explanations. As you thought the next build magically solved the issue. I

From:
"Wetzold, Robert" <robert.wetzold@xxxxxxx>
To:
David M Williams/Raleigh/IBM@IBMUS,
Date:
11/16/2011 08:15 AM
Subject:
RE: [orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed






Hi David,


thanks for your explanations. As you thought the next build magically solved the issue. I am now pretty much set but found one strange last thing: when including the Orbit into Eclipse some of my bundles will be listed as “bundle.name” and not with their real name.


The strange thing is, I cannot spot the difference between those of mine which work (e.g. lucene.core 3.0.3) and those which have no name (e.g. org.restlet). They all have a plugin.properties file containing a bundle.name variable which is also referenced in the manifest.


Do you know of any additional things which need to be done or did you see this before?



Best regards, Robert


From:
orbit-dev-bounces@xxxxxxxxxxx [mailto:orbit-dev-bounces@xxxxxxxxxxx] On Behalf Of David M Williams
Sent:
Dienstag, 15. November 2011 19:59
To:
Orbit Developer discussion
Subject:
Re: [orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed

I'll give a few quick tips, but upfront say it is hard to debug "from scratch" (i.e. not knowing what's changed, etc.).

First, when its something pretty vague (like "Java returned code 13") it does not hurt to simply "try again" ... occasionally, there are burbs in the infrastructure and something just isn't checked out, or something.

Second a good debugging tip is to revert your changes, and make sure it still builds without them. :)

Third, most out-right build failures are due to cvs tagging problems ... such as map says one tag, but that version doesn't exist ... due to typos, etc.

Fourth, there is an easier to read (plain text) "antBuilderOutput.log" produced during the build, but when there is an early, bad error, this log is not saved (so, must be captured during the build ... usually by me ... while the build is in progress). So, if none of the above help, don't hesitate to ask for help here on this list. There's more details that could be said ... but, most committers should not need to read through the build log, so if the above types of actions don't resolve an issue, best to ask the "build master" (i.e. me :).

The current build has gone longer than the failed build, and made it to the signing stage, so not sure if you reverted or fixed something ... or, if it was just one of those burbs ... but I'll keep an eye on it.

And, there could always be areas where we could improve build error handling or reporting ... so, if you see areas for improvement, feel free to open a bug in Orbit's releng component.

HTH, at least a little.


Inactive hide details for "Wetzold, Robert" ---11/15/2011 12:36:00 PM---Hi, I need a piece of advice please. How do you usually"Wetzold, Robert" ---11/15/2011 12:36:00 PM---Hi, I need a piece of advice please. How do you usually go ahead and find out what the reason for th

From:
"Wetzold, Robert" <robert.wetzold@xxxxxxx>
To:
Orbit Developer discussion <orbit-dev@xxxxxxxxxxx>,
Date:
11/15/2011 12:36 PM
Subject:
Re: [orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed
Sent by:
orbit-dev-bounces@xxxxxxxxxxx







Hi,

I need a piece of advice please. How do you usually go ahead and find out what the reason for the failing build is? The log is not really this helpful or I don’t know how to read it yet. Thanks a lot in advance!



Best regards, Robert


From:
orbit-dev-bounces@xxxxxxxxxxx [mailto:orbit-dev-bounces@xxxxxxxxxxx] On Behalf Of orbitBuilder
Sent:
Dienstag, 15. November 2011 18:13
To:
david_williams@xxxxxxxxxx; orbit-dev@xxxxxxxxxxx; rwetzold@xxxxxxxxxxx
Subject:
[orbit-dev] [ORBIT-BUILD: orbit-I] orbit-I Build Failed

View results here ->
http://build.eclipse.org:9777/dashboard/tab/build/detail/orbit-I?log=log20111115165334

BUILD FAILED
Ant Error Message:/shared/orbit/org.eclipse.orbit.releng.control/cc_project_build.xml:69: The following error occurred while executing this line: /shared/orbit/projectBuilders/orbit-I/org.eclipse.orbit/org.eclipse.orbit.releng.builder/build.xml:89: The following error occurred while executing this line: /shared/orbit/projectBuilders/orbit-I/org.eclipse.orbit/org.eclipse.orbit.releng.builder/distribution/orbit.build/build.xml:36: The following error occurred while executing this line: /shared/orbit/projectBuilders/orbit-I/org.eclipse.orbit/org.eclipse.orbit.releng.builder/scripts/build/runbuild.xml:29: The following error occurred while executing this line: /shared/orbit/projectBuilders/orbit-I/org.eclipse.orbit/org.eclipse.orbit.releng.builder/scripts/build/runbuild.xml:177: Java returned: 13
Date of build:2011-11-15T16:53:34
Time to build:19 minutes 4 seconds
Last changed:2011-11-15T16:33:57
Last log entry:Resolved strange merge conflict in bundle map

Errors/Warnings: (10)
checkout.builder.clean: true
retieved new version of build.cfg
Version tag for org.eclipse.orbit/org.eclipse.orbit.releng.builder is: v201110241852
retieved new version of org.eclipse.orbit/org.eclipse.orbit.releng.builder
cctimestamp: 20111115165334
datetime: 201111151656
timestamp: 20111115165334
buildfile: /shared/orbit/projectBuilders/orbit-I/org.eclipse.orbit/org.eclipse.orbit.releng.builder/distribution/orbit.build/build.xml
resolution.devMode: true
cvsaccess: pserver:anonymous

Unit Tests: (0)
No Tests Run
This project doesn't have any tests

Modifications since last successful build: (1)
modifiedrwetzold /org.eclipse.orbit/org.eclipse.orbit.releng/maps/bundles.map1.727 2011-11-15T16:33:57Resolved strange merge conflict in bundle map


_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx

https://dev.eclipse.org/mailman/listinfo/orbit-dev


_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/orbit-dev


GIF image


Back to the top