#jython IRC Log


IRC Log for 2015-05-19

Timestamps are in GMT/BST.

[0:43] * FunkyELF (~FunkyELF@2606:a000:63a2:b200:a801:e347:2372:2ce7) Quit (Ping timeout: 265 seconds)
[1:00] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 264 seconds)
[3:01] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) Quit (Quit: Leaving)
[3:11] * koo6 (~koo5@236.152.broadband3.iol.cz) has joined #jython
[3:29] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) has joined #jython
[5:12] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 245 seconds)
[5:40] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) Quit (Quit: Leaving)
[9:09] * koo6 (~koo5@236.152.broadband3.iol.cz) has joined #jython
[9:16] <agronholm> the C API will have to change eventually to get rid of the GIL in favor of transactional memory
[13:11] <chrisseaton> agronholm: is that actually what they're going to do, or just what you think would be a good idea?
[13:11] <agronholm> chrisseaton: it would be a good idea
[13:11] <agronholm> at least I don't see any other practical way out of GIL
[13:11] <chrisseaton> agronholm: I'd be in favour of interpreting the C code so it's not running natively any more, then you can abstract from things like the GIL but still run the C code
[13:12] <agronholm> interpreting the C code...?
[13:12] <agronholm> if you just want to call C code, there's CFFI for that
[13:12] <agronholm> but C extensions (that interact with Python) will need a new API
[13:31] <chrisseaton> In JRuby we're looking at using an interpreter (with JIT) for the C language, to run Ruby C extensions
[13:31] <chrisseaton> Then you can provide the same C API to the extensions, but really do everything differently under the hood
[13:32] <agronholm> do you emulate the GIL (or whatever it's called in Ruby)?
[13:32] <chrisseaton> We have our own GIL for the moment, but we have research on STM to remove it
[13:32] <agronholm> when you do, you will have to change the C API to take advantage of it
[13:33] <agronholm> since the design itself does not let more than one thread run python (or ruby) stuff
[13:33] <chrisseaton> agronholm: I think we could make certain API calls a transactional commit point
[13:33] <chrisseaton> IO and things like that
[13:35] <chrisseaton> agronholm: you would provide the illusion of a GIL, and run the same code, but really a transaction is enforcing isolation. If you get conflicts, run sequentially
[13:35] <agronholm> mmm
[13:35] <agronholm> okay
[13:36] <chrisseaton> as an example of what we can do - we can give a C extension a VALUE* that points to the backing store of a Ruby array, but we can still specialise the actual storage to an int[] or a double[] and just provide the illusion of having a pointer to boxed values
[13:36] <chrisseaton> so we can abstract quite far as long as provide the same API the C code
[14:12] * xemdetia (xemdetia@nat/ibm/x-idlruqdkmnlshlqa) has joined #jython
[14:16] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[14:19] * agronholm (~agronholm@2001:1bc8:102:6f29:c0bc:8c53:ba65:66bc) Quit (Ping timeout: 265 seconds)
[14:23] * agronholm (~agronholm@nblzone-211-20.nblnetworks.fi) has joined #jython
[14:23] * ChanServ sets mode +o agronholm
[15:29] * xemdetia (xemdetia@nat/ibm/x-idlruqdkmnlshlqa) Quit (Quit: Leaving)
[16:20] * xemdetia (xemdetia@nat/ibm/x-hkblgwbjsiisavif) has joined #jython
[17:07] * srcerer (~chatzilla@dns2.klsairexpress.com) Quit (Quit: ChatZilla [Firefox 37.0.2/20150415140819])
[17:09] * srcerer (~chatzilla@dns2.klsairexpress.com) has joined #jython
[19:59] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) has joined #jython
[20:33] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[21:03] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Quit: Leaving)
[21:04] * srcerer (~chatzilla@dns2.klsairexpress.com) Quit (Read error: Connection timed out)
[21:06] * srcerer (~chatzilla@dns2.klsairexpress.com) has joined #jython
[21:46] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) Quit (Remote host closed the connection)
[22:54] * xemdetia (xemdetia@nat/ibm/x-hkblgwbjsiisavif) Quit (Quit: Leaving)
[22:54] * xemdetia (xemdetia@nat/ibm/x-pccupsyqfjnicfac) has joined #jython
[23:27] * gopar (~gopar@c-73-162-171-146.hsd1.ca.comcast.net) has joined #jython


These logs were automatically created by JythonLogBot on irc.freenode.net using a slightly modified version of the Java IRC LogBot.