[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ecf-dev] ECF project and Tycho
- From: Scott Lewis <slewis@xxxxxxxxxxxxx>
- Date: Sun, 09 Nov 2014 11:49:43 -0800
- Delivered-to: firstname.lastname@example.org
- User-agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0
On 11/9/2014 11:31 AM, Wim Jongman wrote:
Yes. Given that we already have our own website, builder, and test infrastructure, I don't see any advantages of creating/using a HIPP instance. And there are a lot of associated hassles if we were to do so (restrictions or complexity of distribution of non-EPL components, OSGi CT usage/testing, our support for other target platforms...e.g. Karaf, etc).
Our HIPP already exists and is building some of our example stuff. Building the new tycho on our HIPP is better because it is sandboxed away from our main build server.
Once the Tycho build is running it is easy to restore it on the main build.
The missing link here Wim is that before we switch to a new Tycho-based
build, it has to build/run/test all of the pieces of our build that
cannot be run/tested on the HIPP instance:
1) Non-EPL projects (e.g. JMS provider, MQTT provider)
2) OSGI CT (as per Markus' reply just now)
3) Discovery and other config needed for multi-process automated testing
(e.g. dns, xmpp server, etc.)
I can see this working if the conversion process is:
a) Setup Tycho on HIPP instance and convert EF projects (ECF core,
filetransfer, remote services, etc)
b) Move setup to new VM on OSU OSL instance
c) Do work to convert remaining proejcts (i.e. 1, 2, 3, and any others)
to Tycho build
d) Switch to depending upon the new Tycho build
If this is what you have intended then that's fine. But it seems to me
extra work to do 'a', as opposed to just jumping to/starting with 'b'.