Hi
"I'm not sure why the length of time makes it easier to fix..."
After so long, the number of existing users making use of a really stale
URI such as http://download.eclipse.org/modeling/emf/updates is surely
very low and falling?
Conversely the number of new users getting confused by stale URIs is
comparatively stable.
Establishing a better policy for stale P2 repos:
+ less clutter for OOMPH / P2 Update
+ less confusion for new users
- inconvenience for existing users of prehistoric URIs
Seems like a clean up has useful benefits and only a trivial downside.
Those few users of the prehistoric URIs probably have the intelligence
to debug the clean up and may even be grateful for being moved on to
something not prehistoric.
Regards
Ed Willink
On 25/06/2016 14:11, Ed Merks wrote:
Ed,
Comments below.
On 25.06.2016 08:38, Ed Willink wrote:
Hi Ed
The confusion between EMF (container project) and EMF (core) has now
been in place for so long that it may be it now easy to fix.
I'm not sure why the length of time makes it easier to fix...
e.g. Is there really any point in providing
http://download.eclipse.org/modeling/emf/updates
No, it contains antiquated versions.
whose most recent contributions are from 2009 (EMF 2.5)?
Yes, that's quite old. :-P
when http://download.eclipse.org/modeling/emf/emf/updates (EMF
2.7...) was surely intended?
Yes, though that seems not to be maintained so well either because
2.12 is not composed.
Where did EMF 2.6 go to ?
http://download.eclipse.org/modeling/emf/emf/updates/2.6 though I
assume your point is it's not composed in any nice composite.
Perhaps P2 repos should be archived in the same way as downloads. i.e.
http://archive.eclipse.org/modeling/emf/emf/updates
It seems to me anyone using very old versions can use the
version-specific repository...
might be an aggregate of all EMF updates, whereas
http://download.eclipse.org/modeling/emf/emf/updates
might be only the last 3 SimRel's thereby reducing the pressure on
OOMPH etc.
All these composite are slow to use (involve many direct accesses to
the Eclipse server) so in general downstream clients are best off to
use a more specific update site.
For legacy consistency,
http://archive.eclipse.org/modeling/emf/updates could be all EMF
contained projects.
At this point moving it will just break people using it, though I
doubt anyone is using it...
Regards
Ed Willink
On 25/06/2016 13:20, Ed Merks wrote:
Konstantin,
Using
https://wiki.eclipse.org/Eclipse_Oomph_Authoring#How_to_find_a_P2_repository_at_Eclipse_using_the_Repository_Explorer
you can locate http://download.eclipse.org/modeling/emf/emf/updates/2.12
On 22.06.2016 19:01, Konstantin Komissarchik wrote:
I see that 2.12 final release now appears on the download page, but
itโs not included in the updates URL, nor is there an
emf/updates/2.12 repository.
http://download.eclipse.org/modeling/emf/emf/updates/compositeContent.xml
We are working on updating our build to use final Neon dependencies
and EMF is the last one we are waiting on. ๐
Thanks,
- Konstantin
_______________________________________________
emf-dev mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://dev.eclipse.org/mailman/listinfo/emf-dev
_______________________________________________
emf-dev mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://dev.eclipse.org/mailman/listinfo/emf-dev
_______________________________________________
emf-dev mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://dev.eclipse.org/mailman/listinfo/emf-dev
_______________________________________________
emf-dev mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://dev.eclipse.org/mailman/listinfo/emf-dev
_______________________________________________
emf-dev mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from
this list, visit
https://dev.eclipse.org/mailman/listinfo/emf-dev
