[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [cbi-dev] Minutes from October 2, 2012 cbi-dev call
|
Alex,
I really appreciate your enthusiasm, and of course Red Hat's support
via. Krzysztof's work as you know.
CBI is being used and will be used. Hand in hand with accepting new
technology is negotiating and meeting acceptance criteria for deployment
as well as the team ramping up on it. That's what we're talking about
here. No one is sitting back and waiting. No one wants to punish early
CBI adopters, what ever that means. And no one is talking about
abandoning CBI. Well, except your mention of it, but I think hurry is
what you really mean rather than abandon.
Cheers,
Andrew
On 10/02/2012 11:05 AM, Aleksandar Kurtakov wrote:
Time is ticking and we have to either start using CBI or abandon it and move back to our custom build system. As Krzysztof is the first guy to ever made use of CBI to provide a working full featured platform I propose accepting his patches more eagerly even if they are not in "the Maven way" (tm) and get the whole thing rolling. Improvements can be made after that but sitting and waiting gains Eclipse community nothing. It's important to have CBI working ASAP.
Krzysztof patches are using customCallbacks, call ant, call eclipse - but still it makes everyone able to build on his own machine. It might be even looked as intermediate step until the "pure maven way" happens but please don't punish/discourage early adopters and contributors with another delay.
Alexander Kurtakov
Red Hat Eclipse team
----- Original Message -----
From: "Andrew Ross" <andrew.ross@xxxxxxxxxxx>
To: "Common-build Developers discussion" <cbi-dev@xxxxxxxxxxx>
Sent: Tuesday, October 2, 2012 5:43:05 PM
Subject: [cbi-dev] Minutes from October 2, 2012 cbi-dev call
Hi Everyone,
These are the minutes from today's cbi-dev call. I'll also post them
in the wiki.
Attendees:
Paul Webster, Krzysztof Daniel, Thanh Ha, Andrew Ross
Cut-over Timing:
The goal of this work is to have the Eclipse platform cut over to use
the CBI built platform instead of the PDE built version.
* ASAP to be honest. Maps to the following milestones:
* Juno SR2
* Kepler M3 if possible (4 weeks to go) if not M4
It is worth noting that Krzysztof shared Red Hat has already cut over
to CBI for the platform, is happy with it, and not waiting on
anything.
Actions:
* Andrew to follow-up with Igor re: Tycho related bugs
* Andrew to follow-up with Jason re: having a Maven guru from
Sonatype work on the maven-related list below.
* Jan, Tobias... can you guys help with any of the Tycho related
bugs?
The following are our short list of bugs we need for the CBI platform
team to cut over to use CBI.
Tycho related bugs:
* Include the ISV doc in our SDK:
* Bug 389762 - Add support to tycho-source-feature-plugin to
allow including non-source bundles (this is the short term
blocker)
* Create root files with binaries for the Mac (this is an
adoption blocker)
* Bug 386263 - Add support for root.folder.<subfolder> and
root.<config>.folder.<subfolder>
* Generate the .api_description used by our tools to manage
versions and public API (this has to be sorted out before SR2)
* Bug 381057 - CBI should be able to generate the API tools
description
* Make sure our executables are coming from the correct place,
and that 3rd party RCP apps that use o.e.eclipse.executables or
o.e.rcp.configuration still work
* Bug 390111 - Make sure that tycho product materialization
doesn't interfere with SDK features ( (this has to be sorted out
before SR2)
* Support generating the EclipseSource-References into the
MANIFEST.MF (this has to be sorted out before Kepler)
* Bug 380872 - Add support for Eclipse-SourceReference headers
Maven related bugs:
* Bug 377115 - Build SWT and consume native fragments
* - this would be converting ant scripts to maven magic: the
source is copied, pre-processed (switching "int" for "long" for
example), and then compiled for each fragment. This is a
high-priority for maven help as my build on my computer fails to
produce SWT jars.
* Bug 370704 - equinox launchers build
* - Waiting on a maven friendly way of publishing the artifacts
earlier on in the build and consuming them for the many features
that include these root files, this effects about 4 features.
Sorting out these root files is a high-priority for maven help .
* Bug 390020 - Create the swt-<buildId>-gtk-linux-x86.zip files
* - swt-*.zip files need to be created using the appropriate
maven magic. We need help on this before adoption.
* Bug 385154 - JDT core is not ready for tycho builds
* - I think we need some maven magic here to post-assemble the
ecj jar. We need help on this before adoption.
* Bug 385959 - org.eclipse.jdt.launching uses
customBuildCallbacks
* - a jar needs to be compiled earlier in the build and then
consumed by this plugin. We need help on this before adoption.
* Bug 384873 - org/eclipse/core/runtime/IPluginDescriptor is
missing in runtime_registry_compatibility.jar
* - we need a way to include this pre-compiled class. We need
help on this before SR2.
General bugs:
* Bug 385410 - Build documentation using tycho
* - we were going to try Christopher's patch, in a profile, and
require an eclipse install for schema conversion and javadoc
inclusion.
* (may need some maven help with this too)
* Bug 375554 - Be able to specify the jdt.core used for compiling
* - We can test out if this works.
* Bug 369798 - Add signing capability to the CBI build for the
eclipse platform
* - we'll move forward with a fix for this once the discussion
dies down (I think we have consensus).
New bugs:
* Bug 390119 - branding plugins buildId aren't updated
* - need to replace token with build tag. Possibly a maven
related bug. 2nd part is the qualifier being set. Can we
override?
* Bug 390721 - The product IU qualifier is not updated
* - qualifier should be it's own or the highest of what it
includes. Likely a Tycho related fix.
