<br><br><div class="gmail_quote">On Sun, Mar 6, 2011 at 3:40 AM, Levente Uzonyi <span dir="ltr"><<a href="mailto:leves@elte.hu">leves@elte.hu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On Fri, 4 Mar 2011, Mariano Martinez Peck wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
We are really happy to announce that ESUG will sponsor us once again through<br>
the ESUG Summer Talk project. This means that we have reached the ESUG<br>
expectations and that they still think that relational database access is an<br>
important matter in Smalltalk.<br>
</blockquote>
<br></div>
snip<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
5) Cog compatibility.<br>
</blockquote>
<br>
What has to be done?<div class="im"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
6) Use Alien instead of FFI.<br>
Eliot is working on a threaded CogVM. One of the projects of the GSoC of<br>
this year was to make something similar to a threaded FFI. What the student<br>
did is a modification in Alien (I think) that can be run in a multithreaded<br>
envirorment. He worked with Eliot. The idea is when Eliot releases the<br>
threaded CogVM, this FFI would work our of the box, and would avoid locking<br>
the WHOLE vm while a C function is being invoked (as it happens today with<br>
FFI).....So....when that VM is released, we MUST migrate to that).<br>
</blockquote>
<br></div>
How is Alien better than FFI? AFAIK Alien itself is not developed anymore,<br>
parts of it will be merged with FFI.<br></blockquote><div><br></div><div>Rather than get stuck on the names I think the direction is clear. Use the threaded FFI which is integrated with the new Alien callback facilities. I'll try to use part of the course at Lille to realise integrating the FFI and callback facilities into the standard VM and hence be able to publish the callback support and have it ready for general use. So the "new" FFI is initially Alien-style callbacks integrated into the current FFI, and evolves to a fully threaded FFI. I'm going to try to publish a new version of OSCog today that includes the threading work. Since the threaded VM is still a prototype it'll need plenty of pounding on to get ready for production.</div>
<div><br></div><div>does this make sense?</div><div><br></div><div>best,</div><div>Eliot</div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><font color="#888888">
<br>
<br>
Levente<br>
</font><br>
P.S.: <a href="http://www.squeaksource.com/SqueakDBX.html" target="_blank">http://www.squeaksource.com/SqueakDBX.html</a> states that code in the repository has MIT license, but it includes Glorp which has LGPL(S) license.<br>
<br>
</blockquote></div><br>