t3rmin4t0r (address@hidden) has joined channel #dotgnu
 Signoff: t3rmin4t0r (Client Quit)
<~urhysw~u> Last year, or last week?
<lstrcmpi> what nb t3r and FrePort were all talking about.
 t3rmin4t0r (address@hidden) has joined channel #dotgnu
<~urhysw~u> All aspects of pnet - runtime engine, compiler, class library, tools, documentation, etc, etc, etc
<~urhysw~u> That's last year - the last week I've been hacking on the compiler mostly.
 SidNL (address@hidden) has joined channel #dotgnu
<lstrcmpi> wow. that's cool. a runtime engine for c# ? you used m$ specifictions?
<bc_> what is the current status of the CLR ?
<~urhysw~u> I used the ECMA specifications plus and minus the odd MS embrace and extend.
<lstrcmpi> very cool. =)
<~urhysw~u> bc_: the clr can run simple command-line programs, but stuff that uses files, sockets, etc, needs more work in pnetlib and in the engine.
<t3rmin4t0r> rhysw: are you forgetting the curse of the frogger ?
 rhysw bangs head
<~urhysw~u> of course - I forget the most important part of pnet!  Silly me
 t3rmin4t0r laughs his head off....
<bc_> and the status of pnetlib (the class libraries) is approximately where ? (pretty empty it seems)
<~urhysw~u> bc_: that's about the size of it - lots of work to do on pnetlib
<bc_> if i were to put effort in developing the class libraries, is there a capable  build/runtime environment ? (as i can't bootstrap from windows ala the mono track)
<bc_> or are the pnetlibs lacking because of the lack of the runtime
<~urhysw~u> bc_: my highest priority task right now is to get cscc to the point of building pnetlib on non-Windows platforms.  That will make a lot of things easier.
<~urhysw~u> bc_: the runtime core is pretty much there, but work is needed on the native functions that pnetlib calls to perform various tasks
<t3rmin4t0r> rhysw: I think my lib_file.c is still with U ?
<lstrcmpi> rhysw: native functions?
<bc_> rhysw: this is C right ?
<t3rmin4t0r> yup
 Cannot find FreePort on irc
<lstrcmpi> rhysw: os's?
 rhysw ducks missile thrown by t3rmin4t0r
<bc_> where is the to do list ?
 t3rmin4t0r raises my umbrella
<~urhysw~u> sorry - yes, lib_file.c is still on my TODO list.  You do have cvs commit access and could always check it in yourself. :)
<t3rmin4t0r> I thought you'd like to have a look ;)
<t3rmin4t0r> y'know approval :)
<~urhysw~u> bc_: steve compall put one up on the dotgnu web site recently
 rhysw waves the magic wand and approves lib_file.c
<~urhysw~u> happy now? :)
<t3rmin4t0r> ah !
<t3rmin4t0r> yes
<t3rmin4t0r> Also one of my friends has promised to compile my C# for me
<t3rmin4t0r> for testing reasons ;)
<t3rmin4t0r> ahh
<~urhysw~u> lstricmpi: re native functions - there are some things that have to be written in C, and so pnetlib declares them "InternalCall" and the runtime does the work.  I call them "natives" because Java people are more familiar with the term.
<lstrcmpi> rhysw: oh you mean code that is written in c, but can be used in other languages. (i.e - java, c#).
<bc_> rhysw: sorry to nag, but is there a specific place needing work (as i can't find the todo readily)
<~urhysw~u> For those who are interested, here is the todo list: http://archive.dotgnu.org/new/tasklist.html
<bc_> thank you
<~urhysw~u> thank t3 - he just posted an e-mail on the dotgnu mailing list with the URL - I couldn't find it either. :)
<bc_> thanks t3rmin4t0r
 Lard (address@hidden) has joined channel #dotgnu
<FrePort> Rhys: A few questions I can recall from earlier are -
<FrePort> What is the level of compatibility between Mono and Pnet, and MS-CLR and Pnet?
<FrePort> and...
<lstrcmpi> hmm. i would like to develop the part that send the code to be executed on any local machine that wishes to d/l code from a web-service and run it using the SEE. that shouldn't be done, ofcourse, using SOAP etc. but by plain network protocol, though we need auth with that.
<~urhysw~u> Right now, basic stdin/stdout command-line programs should work in all 3.  Once we get pnetlib up to spec, anything ECMA-compliant should work in all 3.
<lstrcmpi> so as far as i understand i need to work on what you call the - seerun utility  - is that true?
 Silvernerd (address@hidden) has joined channel #dotgnu
<FrePort> "Would my friend be able to develop in VB.Net on Linux using Pnet?" (I told him most likely not, but that he should come back and ask you)
<~urhysw~u> After that, we will be using Mono's upper-level libraries on top of pnetlib, so we should be compatible with whatever they have.
<~urhysw~u> For VB.Net, it is a definite "dunno".  Supporting it is more than just compiling the code (I believe there is a MonoBasic).  It is also a library issue - MS'es libraries no doubt depend upon all kinds of Windows guff, and so making them work is highly unlikely.
<~urhysw~u> I'd prefer to make VB Mono's problem to solve. :)
 gopz (address@hidden) has joined channel #dotgnu
<~urhysw~u> And then we'll just use whatever they come up with.
 gopz had a power failure
 Cannot find t3 on irc
<FrePort> And finally - "Is there any plan to embrace MS extensions and then re-extend them for Pnet?" (practice our own embrace and extend was the point I gather)
<gopz> I think we already have started this -- _() stuff ?
<~urhysw~u> We will need to implement some MS extensions to the ECMA core libraries (ECMA defines 300 classes, but MS has 1000).  After that, it depends upon requirements.  There are differences between Windows and GNU/Linux that make extensions unavoidable.  And we will be supporting JVM side by side in the engine which will make things a little weird.
<~urhysw~u> gopz: that's an internal definition within pnetlib to replace the extremely ugly hack that MS used for internal resources.
<gopz> rhysw: what about the Javacc work ?
<gopz> gopz: rhys this is gopal -- I'm waiting for t3rmin4t0r to time out
 worm (address@hidden) has joined channel #dotgnu
<~urhysw~u> gopz: I'm waiting on the code from Marcin, and then I'll integrate it into the compiler. (I gathered that gopz=t3)
<~urhysw~u> javacc isn't really "embrace and extend" as the ECMA spec doesn't deal with it at all.
<gopz> rhysw: I think the _() stuff will restrict people compiling on MS CSC ?
<~urhysw~u> But the JVM stuff is because it is theoretically possible for a C# program to call a JVM class, which isn't ECMA-compliant of course.
<~urhysw~u> _() is an internal method in "System.Object".  As long as mcs can compile methods (like, duh! :) ), it shouldn't be a problem.
 worm is now known as worm^
<gopz> but Javacc outputs IL binary ?
<~urhysw~u> either IL or JVM, just like the C# front-end.
<gopz> even more wonderful
<bc_> javacc compiles to the CLR ?
<~urhysw~u> yup
<bc_> well
<bc_> thats slick
<bc_> )B
<bc_> B)
<bc_> is javacc complete ?
<~urhysw~u> one of the many benefits of the treecc-based compiler is the ease of adding front ends and back ends.
 worm^ is now known as masken
<~urhysw~u> I'm not sure where javacc is at, as Marcin needs to send me the code.  It was compiling simple examples the last time I saw it.
<~urhysw~u> It's not as far along as the C# front-end of course, but should get moving quickly once the code merge takes place.
<bc_> i don't see it in the "TODO" list
<bc_> under developer tools or runtime (i would think it should be under runtime)
<~urhysw~u> Could someone make a note on the IRC summary that the TODO list needs to be made more prominent, and updated some more?
<~urhysw~u> (the summary to be posted to the list, that is)
 Signoff: t3rmin4t0r (Read error: 110 (Connection timed out))
 t3rmin4t0r (address@hidden) has joined channel #dotgnu
<t3rmin4t0r> aah ! , it's a summer storm out at my place
<~urhysw~u> t3rmin4t0r: you are finally yourself again huh?
 gopal (address@hidden) has joined channel #dotgnu
<gopal> phone line hit this time, my protector tripped
 rhysw has an image in his head of a fried crispy-critter gopal :)
<gopal> rhysw: can imagine a fried crispy fuse right now
<gopal> oh wait I can see it
<~urhysw~u> any more technical pnet questions from the gallery?
<bc_> pnet includes pnet lib ?
<~urhysw~u> it's in a separate tar.gz file / cvs repository
<bc_> i am more interested in seeing pnetlib brought up to snuff
<~urhysw~u> me too - any help is good help
<bc_> in regard to technical questions on pnet, not the distribution details
 Signoff: Lamerix ("The Net Club")
<bc_> rhysw: is there a specific area you have in mind ?
<~urhysw~u> The first phase of pnetlib is ECMA compatibility.  Any ECMA class or method that we don't have yet is fair game.  The dotgnu Web site has a status on what we don't have yet (url in pnetlib's HACKING file).
 Signoff: Lard (Read error: 110 (Connection timed out))
 Signoff: gopz (Read error: 110 (Connection timed out))
 SachaS_ has left channel #dotgnu
<bc_> how big of a learning curve will i have to deal with from the C# perspective if I were targeting the native calls (and i know java well)
 nb (address@hidden) has joined channel #dotgnu
<~urhysw~u> C# is almost Java with a handful of extra syntax (properties, delegates, etc).  It's really easy to pick up if you know Java well.  The syntactic differences are annoying, not difficult.
<bc_> alright
<bc_> i would prefer not to get bogged down in those idiosynchrocies
<bc_> any status on getting the IL to the jvm ?
<~urhysw~u> ? what do you mean ?
<bc_> compiling C# to the JVM
<~urhysw~u> oh
<~urhysw~u> the jvm code generator is in roughly the same state as the IL code generator, but quite a bit of work is needed on the assembler to convert JL assembly syntax into .class files.
 Signoff: t3rmin4t0r (Read error: 110 (Connection timed out))
<bc_> hrmm
<bc_> JL ?
<lstrcmpi> il?
<~urhysw~u> IL == CIL == MSIL == Common Intermediate Language
<lstrcmpi> yes.
<lstrcmpi> what is JL?
<bc_> IL == C# bytecode, C# source  is compiled to IL, which runs in the CLR (the C# JVM)
<lstrcmpi> [i meant shouldn't JL be IL, in what rhysw wrote]
<~urhysw~u> JL is a weird combination of IL and JVM instructions - it uses an IL-like assembler syntax for classes and methods, and JVM instructions inside the methods instead of IL instructions.
<bc_> lstrcmpi: i'm asking as well
<~urhysw~u> Basically, I needed an assembly code for JVM, and I already had a huge amount of IL parsing code around which I reused.
 t3rmin4t0r (address@hidden) has joined channel #dotgnu
<FrePort> rhys: in discussing the SEE with another room member, the following was asked (approximately, even though it was originally about Java, the question also applies to pnet)... anyhow: "If pnet is being run under the auspices of the SEE, and the SEE receives a request to execute code on Pnet, does the SEE simply pass the code down, or should it be doing some checking first? If so what sort of checks would you envision?"
<bc_> i have the code to write out the java classfiles, you just need to map the IL-like classes/methods to java classfiles and insert the bytecodes ?
<~urhysw~u> bc_: more or less
<bc_> got a url for the description of the IL-"classfile" meta data ?
 decex (address@hidden) has joined channel #dotgnu
 FrePort apologizes for the interruption. (I didn't know that messagess autosubmit when i overflow the typing buffer)
<bc_> analog to the java classfile spec
<~urhysw~u> FrePort: I'm not really up on what was finally decided on SEE.  Pnet already has a lot of code in place to check the structure of IL programs for correctness.  Having SEE do that wouldn't be wise.  But other security things like "can this program access this file?" can be done by plugging a SEE-supplied library into pnet or pnetlib at some point.
<~urhysw~u> bc_: the ECMA specifications for CLR contain extensive documentation on the binary metadata format, and the textual assembly syntax.
<t3rmin4t0r> I think a callback to the SEE security manager would be the ideal solution
<t3rmin4t0r> ie pnet --strict-verify --secure-file stuff
<t3rmin4t0r> which calls a SEE callback
 davi (address@hidden) has joined channel #dotgnu
<t3rmin4t0r> to check if a resource is allowed
<~urhysw~u> t3rmin4t0r: yep, that's basically the approach I think we should take.
 Signoff: gopal (Read error: 110 (Connection timed out))
 ss (address@hidden) has joined channel #dotgnu
<lstrcmpi> rhysw: so far the only thing that is written in SEE is the - andromeda ?
<~urhysw~u> that's correct - nb can say more on that, as SEE is "not my department" right now.
 Silvernerd is away: I'm busy
<lstrcmpi> ok. thanks.
 gopz (address@hidden) has joined channel #dotgnu
<gopz> that was another blink in the voltage-time continum
<~urhysw~u> I can see the summary of the IRC chat now: Gopal V. kept the group entertained all night with his popular game of "guess what handle I am now"
<gopz> I wish someone could kick t3
<gopz> so that I could switch back
<gopz> speaking about the summary the IRCLOGS are missing !
<bc_> any representative of the xml-rpc group arrive yet ?
<gopz> nope, they are in the -GMT group
<gopz> ie US timezones
<bc_> i am GMT-5 ...B)
<lstrcmpi> where is that bc?
<bc_> North Carolina
<bc_> USA
<lstrcmpi> cool.
<gopz> I'm in +530 GMT
 lstrcmpi GMT +02:00
<gopz> so rhysw what is the response of the meet ?
<~urhysw~u> it's been worthwhile - looking forward to tomorrow's sessions (my time).
<gopz> so rhys what all have been covered ?

<gopz> I mean , I've missed almost everything
 Silvernerd is back (gone 00:12:35)
<gopz> Silvernerd: how's hacking ?
<rhysw> the usual questions: what pnet/see/pnetlib/etc are, how to contribute, and pointers to more "TODO" information.
<gopz> do you think we have any more developers ?
<bc_> maybe...
<rhysw> I probably won't be able to make the next two sessions (12:00am and 4:00am my time), but I should be online again at 8:00am and 12:00pm tomorrow.
<gopz> ok
<rhysw> Perhaps I can run a live demo of "how to do X with pnet", if someone can come with with a good "X" to use as a candidate (native methods, adding classes to pnetlib, hacking on cscc, etc)
<bc_> anyone got a direct link to the ecma spec ?
<rhysw>     http://www.ecma.ch/ecma1/STAND/ECMA-334.htm   (C#)
<rhysw>     http://www.ecma.ch/ecma1/STAND/ECMA-335.htm   (CLI)
<lstrcmpi> rhysw: "hacking on cscc?"
<bc_> native methods, adding classes to pnetlib
<rhysw> lstrcmpi: mostly for compiler people who want to help out - cscc uses a pretty powerful tool called treecc - it takes some getting used to but makes the compiler writing process much easier.
<bc_> enough of the C# to get me very interested
 chiraz (address@hidden) has joined channel #dotgnu
<lstrcmpi> ok. i'll look google for data about treecc. thanks.
<rhysw> we'll need to try to get the irc bot going next time if I'm giving a live tutorial - it will be easier to turn it into a "HOWTO" guide afterwards.
<rhysw> lstrcmpi: http://www.southern-storm.com.au/treecc.html
<rhysw> check out the recently-posted essay: http://www.southern-storm.com.au/treecc_essay.html
<lstrcmpi> rhysw <--- faster than google. you are what they say you are =))
<rhysw> lstrcmpi: since it is my tool, on my site, it was pretty easy to find :)
 njanabhikshu (address@hidden) has joined channel #dotgnu
<lstrcmpi> rhysw: you're Rhys Weatherley?
<gopz> also read the article on FreeSoftware Mag http://www.rons.net.cn/english/FSM/issue02/treecc.pdf
<gopz> njanabhikshu: hi
<lstrcmpi> gopz: thank you.
<njanabhikshu> hi gopz
<rhysw> gopz: the pdf and html essays are the same article - FSM hasn't got their HTML version up yet, so I'm providing a copy.
<gopz> ok,
<gopz> njanabhikshu: so, what do you wanna know about DotGNU ;)
<bc_> lstrcmpi: absurd overslept ...heh
<njanabhikshu> yes
<gopz> all the experts are right here
<njanabhikshu> ok
<njanabhikshu> any site which give me basic info about the dotgnu project
<gopz> dotgnu.org is an ideal choice
<njanabhikshu> ok
<gopz> njanabhikshu: also the Guru of Portable.NET -- Rhys is here now
<rhysw> njanabhikshu: hi!
<njanabhikshu> oh
 Signoff: t3rmin4t0r (Read error: 110 (Connection timed out))
<chiraz> Can anyone tell me a little bit more about the distributed execution environment? Is this comparable to BeoWulf?
 gopz thinks he'll stay as gopz
<nb> ok, see...
 Signoff: ss ("Client Exiting")
<nb> no it's not like Beowulf, which is a cluster of computers...
<bc_> more like CORBA / java RMI
<chiraz> does it build on those CORBA standards?
<nb> the idea of SEE is to have something that looks like a single server to the outside...
<nb> but which is really implemented by multiple machines, with redundancy and load balancing.