On Jun 29, 2010, at 6:15 PM, Schwab,Wilhelm K wrote: > Stef, > > It all sounds great. I probably care less about Alien than I do about what > it is supposed to do well: callbacks. We need callbacks, and we need to be > able to do callbacks that return double. That can come from Alien, Native > Boost, or a plugin that makes FFI able to handle callbacks. Callbacks should > not come at a cost of making every call harder to do, which **might** be the > case with Alien.
the only way to make progress on this front is to propose some code. > It would also be nice to have an analog to Dolphin's overlapped calls (calls > made on a separate thread that block only the calling Smalltalk process, and > not the entire image). I think that eliot is working on threaded something for Cog > > SSL would also be nice. Overlapped calls would greatly simplify creating it, > as then one could safely use OpenSSL's blocking calls on separate threads. > The alternative is to do non-blocking calls, which is considerably more work > than letting the OS handle the multi-tasking. > > Bill > > > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Stéphane > Ducasse > Sent: Sunday, June 27, 2010 3:32 AM > To: Pharo Development > Subject: [Pharo-project] 1.2 vision > > Hi all > > Of course we will integrate changes and changes and improvements.... now I > believe that it would be nice to build a common vision. Here what I would > like to get in 1.2. > > Now it does not mean that this is a definitive list and that we will have the > energy to do all but like that you know that you can put energy do build this > common vision :) > > - better infrastructure for integrating rb composite queries into the > system > The idea is to introduce a superclass on top of > systemDictionary and to define there an API that > is compatible with the one of RB (and potentially modified the > one of RB) > > - another step towards using announcement for system notification > > - better tools for dev > > - have a look at the RB change model so that we could get a real undo > and changes > > - clean Monticello so that package logic stays in package so that we > can plug RPackage into the system. > Right now this looks too brittle to do anything > > - make sure that we can have RPackage on the side of PakageInfo (we got > really nice benchmark) > > - New Compiler in beta > > - Helvetia hooks > > - ROME API > > - Sophie Event Hierarchy > > - Define some architectural action to get the system more modular > > - Hudson infrastructure > > - metacello to manage core > > - metacelloRepository for 1.0/1.1/1.2 > Gofer > > - squeak collection optimisation? > > - Alien > > Open points: > > - What do we do with ToolBuilder? Not clear to me. > > - What should be done at the level of polymorph? > - What should be done a the level of Pluggable**** > can we use the fact that we have now block closure to > get it better. > > As you guess it will not come simply but this is important that we all head > towards a common understanding > > > Stef > > > > > _______________________________________________ > Pharo-project mailing list > [email protected] > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > _______________________________________________ > Pharo-project mailing list > [email protected] > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project _______________________________________________ Pharo-project mailing list [email protected] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
