Re: [orion-dev] Regarding Orion Web IDE PHP Editor (GSoC)

[John Arthorne <John_Arthorne@xxxxxxxxxx>]

> Karol Gusak <karol@xxxxxxxx> 
> 1) Are there any early documentation / design
plans available for new
> architecture in M7? For example about extension points, rules to keep
> in mind when extending the core functionality, etc?

It isn't really that M7 will have a "new architecture"
overall. The main difference from your perspective is that there should
be well defined extension points for inserting content assist, syntax highlighting,
etc, for different languages. The design discussions for these features
will likely happen via bugzilla:

Syntax highlighting: https://bugs.eclipse.org/bugs/show_bug.cgi?id=341938
Content assist: https://bugs.eclipse.org/bugs/show_bug.cgi?id=342157

> 3) In case when I'd be allowed to work with Orion this summer, what
> would be the best way to handle my commits? Another branch, or
> elsewhere?

My suggestion is creating a project on Eclipse Labs.
This would be lower overhead for you since it removes the fairly heavy
IP review rules for contributing to Eclipse projects. Also, since your
work appears to be entirely client side, it shouldn't need any build steps,
so pulling your work into the Orion build would just add unnecessary restrictions
for you. I think your integration should take the form of a single Orion
plugin that would be installed from the Orion plug-in registry page. That
would enable your work to be easily installed and used directly by anyone
using orionhub.org hosting, for example. Finally, having the code in a
separate plugin will avoid the temptation to tie the PHP support directly
to the framework code. We have done that today for some of the JS and CSS
language support, and we now have the costly problem of extracting that
into separately consumable plugins.

> 
> PS: Now you can also check out my proposal on wiki (always the
> freshest version)!
> http://wiki.eclipse.org/User:Karol.gusak.eu/
> GSoC_Proposal_Orion_Web_IDE_PHP_Editor

I read through your proposal and I like the general
direction. The idea of picking one or two key features and making them
great sounds better than a mediocre version of many features. Also since
Orion is still very early in its own development, having only one or two
well defined integration points to worry about will greatly increase your
chance of success. As you point out in the proposal, it will be great to
have someone pushing on different language support at the same time we
are developing the related extension points - it will keep us honest and
force us to provide infrastructure that is truly flexible and useful.

John