Hi there,
In summary, the reasons for sticking with Java 6 would be:
1. Users who use a Java 6 environment.
2. Users who use Eclipse 4.3 and below [1].
3. Enough time after the official end of life.
An additional condition for switching to Java 7 is:
4. SWTBot needs to work in that environment, i.e. with Eclipse 4.4+
Based on the current user data:
1. An estimated 1.5% percent of our users use Java 6 in about 1% of our
sessions (see [2]).
I think these number are low enough to safely ignore them.
2. About 16.5% of our users use Eclipse 4.3 or below (in 8.5% of the sessions).
A sixth of our users are a larger number, and I would not like to ignore
them (although they only account for a twelfth of the sessions).
I think we should be clearly under 5% here. So we have to wait.
3. Stefan said "a few years". What would you suggest?
By now, for Java 6 it has been 2 years and 5 months; for Eclipse 4.3 it's 1
year 7 months (since 4.3.2).
So how about this plan:
- We wait until the segment of up-to-Eclipse-4.3 users drops below 5%.
- Then we evaluate whether we still need the SWTBot (it might be superseded by
the HTML GUI and its testing framework by then)
and if so, whether it still works, and make adjustments when necessary.
- Then we switch the Saros project to Java 7.
Any comments?
Franz
[1] https://wiki.eclipse.org/Eclipse/Installation#Eclipse_4.3_.28Kepler.29
[2] http://article.gmane.org/gmane.comp.ide.eclipse.saros.devel/1558
-----Original Message-----
From: Stefan Rossbach [mailto:srossb...@arcor.de]
Sent: Tuesday, September 15, 2015 4:39 AM
To: Denis Washington <de...@denisw.de>; Zieris, Franz
<franz.zie...@fu-berlin.de>
Cc: dpp-devel@lists.sourceforge.net
Subject: Re: [DPP-Devel] IntelliJ 15 for OS X will come with JDK 8
Most of our students have a hard time to write code that will not fail
regarding the common basics of Java.
While some of the functional programming stuff in Java 8 is indeed nice
it will not solve the problem regarding the lack of imperative
programming for quite a bunch of members.
Most of the stuff you mentioned in Java 8 is just syntactical sugar. You
can archive the the same functionality with Java 7 and below with a
little overhead.
As we already agree that we do not want surprise our userbase with a
plugin that required a JRE that exceeds the minimal requirement of the
IDE I think
Java 7 will be fine but if and only if: We will bump the necessary
Eclipse Version to 4.4 AND ! we will get this SWTBot stuff working with
the new setup.
Those E2E SWT tests are more valueable than a simple JDK version upgrade.
On 14.09.2015 21:42, Denis Washington wrote:
> On 14.09.2015 20:26, Zieris, Franz wrote:
>> Hi Denis,
>>
>> What are the arguments against Java 8?
>
> There aren't really any. I mostly proposed Java 7 because it is the
> conservative choice: it would make the Eclipse plugin work everywhere
> where a recent Eclipse (>= 4.4) works. For the user, this might be
> less surprising than a plugin which requires a newer Java version than
> the IDE itself.
>
> Other than that, there is no real reason why we shouldn't upgrade to
> Java 8. Other than Stefan, I think that Java 8 has several great new
> features in the language and libraries that will be very beneficial
> for us. For instance, the readability of many inline Runnables we have
> in our code would benefit greatly from the new lambda expression
> syntax, and the functional tools in java.util.stream as well as the
> java.util.concurrent enhancements will probably also be useful [1].
>
> But as I said, personally I would be happy with a minimum of Java 7 as
> well.
>
> Regards,
> Denis
>
> [1]
> http://www.oracle.com/technetwork/java/javase/8-whats-new-2157071.html
------------------------------------------------------------------------------
_______________________________________________
DPP-Devel mailing list
DPP-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dpp-devel
