On 07/25/2012 12:14 PM, Igor Fedorenko
wrote:
I
completely agree that Tycho extensibility story "is not good
enough",
although I'd use much strongly languahe myself :-). At the same
time,
I think it is a chicken-and-egg problem, at least partially. We
need
feedback/patches from real exploiters writing real extensions to
improve
our apis.
I think APIs are not that bad, process to write a plugin is easy
(it's standard Maven + dep on Tycho), process to write an extension
is easy to understand and mimic by looking at component.xml.
But people need to understand that Tycho is an extension of Maven
and that most of their use-cases can be fixed in "the Maven way", by
finding a plug-in to use on the right phase. And in case they can't
find what they want, they can write a plugin in the Maven way too.
There is no additional things to know to get started.
About Tycho extensions, it's probably better to ensure people share
their use-case on the mailing-list rather than letting them develop
10x the same thing on their side. Tycho is still young and some
concepts are still to discover. I'm afraid that if we encourage
everyone to write his own plugin or extension, we miss some
use-cases that are necessary to know in order to bring Tycho (and
CBI) to maturity.
It's more a concern about the status of Tycho: is it mature enough
to encourage external extensions, or should we focus on getting it
more complete and gather use-cases instead?
|