User-agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-GB; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
Hi Adolfo
Orbit now has an R20110523182458 (as well as an S20110523182458).
Changing our builds to R made no difference.
I think the missing parameters means than none of the platform
build.type guards in ocl-platform.rmap pass, so no platform.
Looking at the Xtext parameters, they don't look too hard to clone;
only the reference repository looks harder.
(They have a "sign" parameter which could save time on N-builds.)
(There is also a Publish Javadoc config: is this the 'same' as our
'MANAGE_JAVADOC")
I'll leave it to you.
Ed
On 31/05/2011 06:08, Ed Willink wrote:
Hi Adolfo
My start ability was back so I tried a build; it went straight
off; indeed no job parameterization. Failed on an Orbit access.
I'll see what's missing.
80% of recent jobs have failed, glancing at a couple suggests file
permission issues.
I suspect Denis is busy; strangely no further update message.
Ed
On 31/05/2011 00:40, Adolfo Sanchez Barbudo wrote:
Ed,
I've realized that there is missing info in the job's
configuration... among other things, users' access, which was
the cause why we weren't able to run the job... The job
parameters are also missed...
I hope to fix the issue tomorrow's morning... Perhaps, we should
also warn Denis and/or cross project mail list about the
issue...
The point is that I CAN'T START a job.
So I can't do anything right now... I'd
definitely wait until servers are
stable. If the servers are stable,
perhaps the current state of project can
successfully be built...
My opinion is:
1. Announce the delay Eclipse OCL RC3
until tomorrow's morning.
2. Try to run an S-build in the morning:
a) If we can't run builds yet
=> raise a bugzilla to Eclipse
Hudson.
b) If we can, we could take a
decision depending of the result of the
built.
The Download for docbook zips
failed while I watched. So
it's failed now.
It seems that with limited
network connectivity, the
download may be a problem.
Option 1:
revert the documentation to
the non-auto-generated form.
will probably need a retry, so
realistically 3 builds through
a massive
slow Hudson queue.
Option 2:
manually do the zip downloads
so that they don't need a get
each build
Option 3:
notify cross-project-issues
that RC3 will be late.
- the only changes are
examples and doc (? and source
bundles)
-- very unlikely to affect
other projects: they can use
RC2
- hopefully connectivity will
be back within 24 hours
I favour starting on 3, and
see how David Williams reacts.
I'm sure other projects will
stretch too.
Regards
Ed
On 30/05/2011 21:08,
Adolfo Sanchez Barbudo
wrote:
I did the change to
the master... it's not
really worthy since it
is an N build... the
point is that now I
can't either stop the
running job nor start a
new one .... :\