Hi David
Sorry my bad; too engrossed in one problem to heed advice on the
next step.
I think I have overinterpreted the versioning rules. We only a need
a major increment for UML-related plugins and almost all features.
This should be friendly downstream where consumers just reference by
feature name and mostly use non-UML plugins; consumers will just
need to get their platform, EMF, UML repos right.
This seems to align with advice in
http://wiki.eclipse.org/Version_Numbering#Require_features which
points out that feature versions are brittle. It seems we will have
over 90% of the project at 3.2 but a small part at 4.0 and so almost
all features at 4.0.
Wouldn't it be better if features were release and SR-named not
version numbered? i.e. org.eclipse.ocl.master_Juno.0.v20120606.
Regards
Ed Willink
On 23/09/2011 02:45, David M Williams wrote:
[I saw Kenn's reply, just before
pressing "send"
... he says it all ... but, if you like long wordy, detailed
explanations,
here's my version.]
Here's my understanding (little as that is) ...
of
the EMF and platform situation for M2.
If you use EMF (common, or ecore features) you
must
use
http://download.eclipse.org/modeling/emf/emf/updates/2.8milestones
as it is the one that has the two features:
org.eclipse.emf.common_2.7.0.v20110916-1359
org.eclipse.emf.ecore_2.8.0.v20110916-1359
And, those are the two features that the
Eclipse M2
repository _also_ provides, in
http://download.eclipse.org/eclipse/updates/4.2milestones/S-4.2M2-201109161615
And ... at least some of the bundles in there
are
singletons, so only one can be installed at a time. And,
currently,
the platform does an "include" instead of
a "require" which tightens the constraints even further (which
is the subject of bug 356644).
Sort of a perfect storm.
The b3 aggregator (and, likely the buckminster
build,
the output of which included in note below) checks to make
sure "all
is installable together" (That is, after all, one of the main
reason
of using the aggregator, it does not do a blind mirror of
bundles and features
(if it did, sure, the mirroring might succeed, but the results
would not
be installable ... a "build time error" detection rather than
a waiting to find a "runtime error").
Oh, and, I should have said, EMF's I-build
location
(in the output below) does not have that "most recent" EMF
features
that the platform and EMF's milestone site includes ... I am
not sure what
the EMF team's typical practice is in this regard, but know
that for this
milestone they did "jump through hoops" to work around this
include/require
issue at the same time building _for_ the platform, as well as
_against_
the platform. (And, it is appreciated).
Is your head spinning yet? I know mine is :)
... but,
I hope this helps explain the current situation. Things should
be better
for M3, but until then, there is this tight requirement that
everyone use
exactly these versions of emf.common and emf.ecore -- well, if
you do any
sort of 'include' and/or if your build process requires a
correct and consistent
target. [And, teams involved, feel free to correct me where
I've erred].
From:
Ed Willink
<ed@xxxxxxxxxxxxx>
To:
Cross project issues
<cross-project-issues-dev@xxxxxxxxxxx>
Date:
09/22/2011 03:04 PM
Subject:
Re:
[cross-project-issues-dev]
Juno M2 aggregation -- will it be the worst milestone ever?
Sent by:
cross-project-issues-dev-bounces@xxxxxxxxxxx
Hi
For MDT/OCL
> We plan to commit a dependency change to [4,5) shortly
that should
> allow other downstream projects to build, provided that
they too
> change to UML [4,5). This should be available perhaps
within 12 hours
> if Hudson and other factors are co-operative.
We are trying to build a candidate, but get
(https://hudson.eclipse.org/hudson/job/buckminster-mdt-ocl-core-3.2-master/440/console)
WARNING [0004] : Component request
org.eclipse.emf.common:eclipse.feature/[2.7.0.v20110912-1000,2.7.0.v20110912-1000]
is
inconflict with request
org.eclipse.emf.common:eclipse.feature/[2.7.0.v20110916-1359,2.7.0.v20110916-1359]
WARNING [0004] : Component request
org.eclipse.emf.ecore:eclipse.feature/[2.8.0.v20110912-1000,2.8.0.v20110912-1000]
is
inconflict with request
org.eclipse.emf.ecore:eclipse.feature/[2.8.0.v20110916-1359,2.8.0.v20110916-1359]
ERROR [0113] : No suitable provider for component
org.eclipse.emf.common:eclipse.feature/[2.7.0.v20110916-1359,2.7.0.v20110916-1359]
was
found in resourceMap
file:/opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.2-master/org.eclipse.ocl.git/releng/org.eclipse.ocl.releng.buckminster/releng/ocl-platform.rmap
ERROR [0113] : No suitable provider for component
org.eclipse.emf.common:eclipse.feature/[2.7.0.v20110916-1359,2.7.0.v20110916-1359]
was
found in searchPath emf
ERROR [0113] : Rejecting provider
p2({0}/modeling/emf/emf/updates/2.8-I-builds[file:/home/data/httpd/download.eclipse.org/modeling/emf/emf/updates/2.8-I-builds]):
No
component match was found
ERROR [0113] : No suitable provider for component
org.eclipse.emf.ecore:eclipse.feature/[2.8.0.v20110916-1359,2.8.0.v20110916-1359]
was
found in resourceMap
file:/opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.2-master/org.eclipse.ocl.git/releng/org.eclipse.ocl.releng.buckminster/releng/ocl-platform.rmap
ERROR [0113] : No suitable provider for component
org.eclipse.emf.ecore:eclipse.feature/[2.8.0.v20110916-1359,2.8.0.v20110916-1359]
was
found in searchPath emf
ERROR [0113] : Rejecting provider
p2({0}/modeling/emf/emf/updates/2.8-I-builds[file:/home/data/httpd/download.eclipse.org/modeling/emf/emf/updates/2.8-I-builds]):
No
component match was found
INFO: TAG-ID 0004 = Query for
org.eclipse.ocl.releng.buckminster:buckminster,
path:
org.eclipse.ocl.releng.buckminster:buckminster$0.1.0.qualifier
->
org.eclipse.emf:eclipse.feature$2.8.0.v20110913-1544
TAG-ID 0113 = Query for
org.eclipse.ocl.releng.buckminster:buckminster,
path:
org.eclipse.ocl.releng.buckminster:buckminster$0.1.0.qualifier
->
org.eclipse.platform:eclipse.feature$4.1.0.v20110612-1800-9JF70HDuFo6EMhMISblH-cp6WYcpnVwlmX9L-0_CgNLen
->
org.eclipse.rcp:eclipse.feature$4.1.0.v20110822-9-CVG2DFlt5xpcTFqQa2DE2YQl6C
->
org.eclipse.e4.rcp:eclipse.feature$1.1.0.v20110712-1859-7HAN6FTy21UwAsXPk5BIP
Are there conflicting EMF repositories?
Regards
Ed
Willink
_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cross-project-issues-dev
_______________________________________________
cross-project-issues-dev mailing list
cross-project-issues-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cross-project-issues-dev
No virus found in
this message.
Checked by AVG - www.avg.com
Version: 10.0.1410 / Virus Database: 1520/3912 - Release Date:
09/22/11
|