Hi Calatus, again many thanks to you for sharing this with us :-) But to reuse some part of code we have some constraints from Apache, and contributed code must be attached to a JIRA ticket (with the grant enabled for inclusion in asf work) ... otherwise we could NOT use it. For now, thanks for having included some details here.
In any case, we have to see what Greg and other Developers say, but I think that a feature like this could really be useful, at least in a way where a Pivot newbie need to better understand some of the internals of BXML, and in my opinion it's always a good thing. > Although BXML based declarative style makes the GUI specification shorter, > when I started using Pivot, I faced a lots of runtime error, and it was > quite difficult to debug, I couldn't see any clue to debug. This > heterogeneous approach mixing, XML, Javascript, etc reminds me old JSP > style approach. I agree with you, because for some things I'm in the same situation :-) ... So my proposal for this (but only as a starting point for discussions): https://issues.apache.org/jira/browse/PIVOT-578 Then, like in JSP I proposed some flag to forbid the inclusion in BXML of scripting logic ... but this probably is not so important here (there is a discussion on this). And us you, I think that more checks at compile time (by Java/Groovy/Scala Builders, annotations, etc) are always a good thing, for everyone, at least as alternative options. > By the way, this version was based on the idea to use AOP like > interception for Pivot objects. This approach is quite general, but use a > lots of variables. > And after I recognized the pattern, I think this code should not be used > for this BXML to Java converter. Personally I don't like too much AOP interceptors, they are a powerful concept, but to me seems to lose control over things, and to have a too complex part running under ... but it's a question of taste. > Now I have almost completed a new version which generates codes with > suggested anonymous class based coding style. I will commit this version > later in github after I fixed some issue. Very good. > But if you consider this should be committed to a storage related to > apache project, I would commit this to appropriate place for that. Ok, but for committing to our main repository you should be a Pivot Developer, so for the moment probably the best thing could be to enable you on our "secondary" pivot-related projects (if you are interested, and there aren't objections ... but I don't think there will be). And on this, we are moving some of these projects (but I can't tell you more details at the moment), so if you want I can grant you to write in my "temporary" projects, like this: http://code.google.com/p/pivot-stuff/ the only requirement is to create inside a dedicated project, and for example (if useful) use Pivot libraries from projects already inside. But I understand that this could be not the best workflow for you at the moment, so don't worry, proceed as you prefer: tell me if you want some grant, or tell us if/when you will have news (even on your projects on GITHub). > Anyway I think this requires more testing and review. No problem, we are here even for this :-) , the only problem is to find time ... > Actually I found one problem for using this approach for current Pivot > source code. > In order to use this coding style, the Pivot API classes should not be > final. > Although most of them are not final, but I found TableView.Column is > defined as final class. > (I locally changed this to none final class to avoid this restriction. Are > there any good reason to make this final?) This is my fear, see if something in Pivot prevents someone to extend features. On this Greg can tell you better if it's only because we are here in an un-covered use case, or if there is a real need to have this final. Can you give some more info on limitations you found (please maybe starting a new thread on Dev mailing list, or forwarding this) ? Comments ? Note: at the moment I'm using Nabble, and from here it's not possible to forward this thread to the Dev Mailing List, is someone is able to do it in my place .. thanks. Bye, Sandro -- View this message in context: http://apache-pivot-users.399431.n3.nabble.com/pivot-Collection-overrides-equals-hashCode-bug-tp2134634p2162408.html Sent from the Apache Pivot - Users mailing list archive at Nabble.com.
