Hey Alex, Sean and all the others, I know we''re looking towards doing a possible 1.9.6 release, and as me and Alex discussed with the work towards using MinGW, that I personally would like to hold off on doing a 1.9.6 release, till we get the current issue with the Garbage Collection mark and sweep cleared out, as well as see about possibly getting OpenGL and MediaCtrl working, to keep it on Par with the MSVC version of wxRuby. I also know that Alex said that he was looking to hopefully move to a 2.0 release by the end of releasing 1.9.6, but I see that with the option of using MinGW opening up to us, that we may want to hold off on doing a 2.0 release for a while, till we can be sure that the MinGW version will work. We can still go beyond 1.9.6 release, as we have pleanty of numbers to go through past 1.9.9 (Look at Fox toolkit, with it''s 1.6.18 scheme to see what I mean). The only reason why I bring this up, is the fact that with the very strong possibility of the OCI going from MSVC to MinGW, people may end up getting the latest OCI, with it being MinGW, and find that wxRuby doesn''t work. I''d rather the Fox Toolkit run into that, then wxRuby. Just my two cents, plus I''m still looking at possibly adding back wxSocket classes, to wxRuby, since I''m becoming more comfortable with the Swig stuff, and I want to add Sockets into wxRuby very very strongly. Don''t get me wrong, I like the Ruby sockets, and they have gone a long way to keeping things going on all three platforms. At the same time, wxWidgets has the same operability, with the bonus addition of it being Event Driven, instead of Thread/Polling driven. To all honesty of the fact, trying to re-implement in Ruby, what''s already available in wxWidgets, seems a bit much like overkill. I''ve not seen many posts online, concerning the possible attacks against wxWidgets''s socket implementation being bugged and such, if you still have reservations about that Alex, could you point me to them, so that I can look and see what is going on? And weither it might warrent the same problems on our side or not. Again, these are just my feelings, and I''m not the only developer on the team, so I wanted to see what you guys think about this as well. -- Mario Steele http://www.trilake.net http://www.ruby-im.net http://rubyforge.org/projects/wxruby/ http://rubyforge.org/projects/wxride/ http://rubyforge.org/projects/vwmc/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://rubyforge.org/pipermail/wxruby-development/attachments/20080410/1e45fd17/attachment.html
Hi Mario Thanks for bringing this up. Overall, my view is that the greatest need is for a production version of wxRuby 2.0 with a guarantee of API and feature set, so that developers can develop against it with confidence. I think we''ve reached a level of stability that that''s possible now, and I''d like to push for it asap. Mario Steele wrote:> I know we''re looking towards doing a possible 1.9.6 release, and as me > and Alex discussed with the work towards using MinGW, that I > personally would like to hold off on doing a 1.9.6 release, till we > get the current issue with the Garbage Collection mark and sweep > cleared out,Yep, we must fix the problem with mark_wxEvent. I can reproduce the crash with the sample you sent (only in MingW, not OS X, haven''t tried VC) but currently cannot understand why it''s happening.> as well as see about possibly getting OpenGL and MediaCtrl working, to > keep it on Par with the MSVC version of wxRuby.I agree these would be good to have, but I don''t see these as blockers for the next release. Current stable wxPython doesn''t support MediaCtrl with MingW, so it may not be so easy. However, it''s likely that fixes here will be in the build system rather than the core library, so could be applied after 2.0.> I also know that Alex said that he was looking to hopefully move to a > 2.0 release by the end of releasing 1.9.6, but I see that with the > option of using MinGW opening up to us, that we may want to hold off > on doing a 2.0 release for a while, till we can be sure that the MinGW > version will work.I think it''s likely that the OCI will move to MingW, but I also think we can be pretty confident from our recent experiments that wxRuby will work. It''s just combines a platfrom (Windows) and a compiler (gcc) which we''ve well tested already.> Just my two cents, plus I''m still looking at possibly adding back > wxSocket classes, to wxRuby, since I''m becoming more comfortable with > the Swig stuff, and I want to add Sockets into wxRuby very very > strongly.OK, I think they would be nice to have, but not a vital feature for 2.0. It seems possible to build a range of socket-using apps already using ruby sockets and threads. I would prefer to spend time on polishing what''s there - tidying and commenting the samples so they serve more like tutorials, writing more overviews in the docs, ensuring samples cover new-ish classes - than adding features that may be marginal for most apps. But of course it''s not for me to tell anyone what to work on, and all contributions are welcome. I''d like to see Sockets added to a 2.1.x or 2.9.x next development series which was more modular in design, and shared the SWIG runtime code. Then it would have no impact on library size. I''m open to the views of other contributors and users on these topics. alex