On 09/14/2016 04:34 PM, Bruno Medeiros
wrote:
1. Eclipse 4
What end?
It was for sure the beginning of a difficult era for the IDE (as
perceived by users), but I don't see any end on the horizon.
Disclaimer:
http://cdn-a.community.bonitasoft.com/sites/community/files/blog/wp-content/uploads/2010/03/IMG_0113.jpg
Do you remember the times before p2? Basically, having a plugin
working was really a miracle, taking hours, requiring to read a lot
of similar story, requiring to read some Eclipse log to figure out
what's wrong and imagine what could be a solution.
p2 is definitely something really helpful, for sure it wasn't a
trivial change. But see where we are now: p2 delivers the promise
that everything that installs runs, that you can install old content
on new Eclipse and vice-versa - and still if it installs then it
runs-, it has very useful (although not trivial) error messages
allowing to figure out the cause of any issue, it has allowed to
move build to maven, it allows the Simultaneous Release and
Tycho-based builds to detect at build-time of possible conflicts, it
has a remediation page that is quite helpful in most cases...
With the size of the Eclipse ecosystem, a dependency management at
install-time was and is a must-have; p2 is a good solution. I don't
know how IntelliJ or other platforms deal with the fact that content
can come from multiple sources, maybe they require people to rebuild
content or to re-test it to get to a newer release. With p2, those
steps can be avoided.
Bashing p2 as it's done in the article is IMO a clear sign that the
author doesn't have a sufficient knowledge of the needs nor of the
solution related to the installation story to write about it.
I like how Eclipse IDE looks on Fedora Linux. There is an important
bug about UX and theming:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=325937 . However,
no-one really considered it as important enough for other OSs right
now.
Are we a product focused organization? Do every involved
contributors want this?
|