Comments for Villane

Comment on Baseline Eclipse for Java Game Development? by dmdevito

dmdevito — Tue, 29 Sep 2009 10:27:34 +0000

Nice idea.
Why not starting a game distribution on top of Eclipse ? See my idea: http://www.jroller.com/dmdevito/entry/why_not_a_game_distribution
It may drive the creation of a Eclipse distribution for game development…

Comment on Baseline Eclipse for Java Game Development? by Baseline Eclipse for Java Game Development? « Villane | Webmaster Tools

Baseline Eclipse for Java Game Development? « Villane | Webmaster Tools — Sat, 29 Aug 2009 14:22:57 +0000

[...] See the original post here: Baseline Eclipse for Java Game Development? « Villane [...]

Comment on Baseline Eclipse for Java Game Development? by Twitted by t4ffer

Twitted by t4ffer — Sat, 29 Aug 2009 10:49:47 +0000

[...] This post was Twitted by t4ffer [...]

Comment on P2 Still Not Awesome by leberre

leberre — Tue, 30 Jun 2009 09:06:28 +0000

Hi Erkki,

p2 gets a semi-structured explanation from SAT4J, not plain text.

Please fill a bug report if you feel that a simplified explanation should be shown first (just the first lines of the explanation for instance).

Comment on Python is great, but __ by kumar303

kumar303 — Mon, 29 Jun 2009 19:31:58 +0000

In addition to what eengbrec said, double underscores also magically create functions that you cannot override or call. In other words, you’d have to type:

void private final init function() {
…
}

if it were Java. So, in Python, __init__() is private and final but init() or _init() would not be final or private. In general I find private and final methods to be annoying when refactoring or testing but when they are magic hooks into the low level object system (like __init__, __len__, __iter__) then it kinda makes sense.

Comment on P2 Still Not Awesome by Erkki Lindpere

Erkki Lindpere — Sat, 27 Jun 2009 13:38:08 +0000

So P2 gets this message simply as text from the solver? Or is it more structured? Even separating the “software being installed” and “software already installed” from the actual unsolved constraints would be progress, IMHO.

Comment on P2 Still Not Awesome by irbull

irbull — Sat, 27 Jun 2009 03:15:04 +0000

Erkki,

Yes, I agree that a more readable form would be good. However, this is very difficult using a pseudo boolean solver. In p2 dependencies are calculated by running a variety of optimization functions through SAT4J, and if a solution is available, it will find it. If not, this is a different can of worms. Researchers at Université d’Artois have done a lot of work to improve the error reporting in 3.5, but more is needed. I know they are looking for help (research funding, research, etc…).

Comment on P2 Still Not Awesome by Twitted by planet_lang

Twitted by planet_lang — Fri, 26 Jun 2009 23:21:50 +0000

[...] This post was Twitted by planet_lang [...]

Comment on P2 Still Not Awesome by Erkki Lindpere

Erkki Lindpere — Fri, 26 Jun 2009 21:21:21 +0000

I just think this information should be presented in a more readable form. I’ll file a report when I come up with a good suggestion.

Comment on P2 Still Not Awesome by jarthorn

jarthorn — Fri, 26 Jun 2009 20:58:18 +0000

You don’t need to select dependencies at all using p2. Dependencies are computed and installed automatically. It’s only because you’re accustomed to the old UI that (unreasonably) required the user to manually figure out and select dependencies. Just don’t select EMF at all and it should work.

The root of the error message is on the fifth line, “Only one of the following can be installed at once”. You are trying to install something that is incompatible with what you have installed (they require different versions of the bundle org.eclipse.jem.util, but there can only be one version of that bundle installed at any given time). If you think this is still unclear please enter a bug report. p2 is just enforcing the constraints defined by the authors of these features.