August 2009 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] new socket API

by Schwab,Wilhelm K

Philippe, I've never had much use for UDP, but the common things should work the same, at least to a point. UDP is the protocol that does not guarantee packet order or even delivery, right? So anything remotely stream related is out, and there might be other things that should legitimately be different, if I'm thinking of the right thing. While I'm displaying ignorance, there is a negotiated rate serial protocol that ends up being implemented using sockets. I can't think of the name of it but know where to look for a reminder. It is something I would like to see supported "some day." Bill -----Original Message----- From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Philippe Marschall Sent: Wednesday, August 26, 2009 11:46 AM To: pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] new socket API Btw, while we're at it: TCP sockets and UDP sockets have a different interface. Some messages can only be sent to TCP sockets others only to UDP sockets. Yet they are both represented by the same Socket class. Cheers Philippe Schwab,Wilhelm K wrote: > Stef, > > I am not sure exactly what should be next, but I do not think we can simply kick it back to squeak-dev any more than we can the GUI/IDE, streams, etc., as sockets are very important. I am slowly chipping away at things that are standing in the way of my first Pharo success story; the impact of sockets has been minimal, but that will change. When I get going, I will be looking carefully at how things behave when peers do not respond (only the calling thread should suffer). Dolphin's socket streams are nice; Pharo seems to lean toward a single read-write stream, which should be fine as long as it manages not to hobble the image when the wrong wire gets kicked. SSL via OpenSSL should be easy to use; that will be an interesting adventure, but it needs to happen. > > With all due respect to those who have disagreed, I'm still highly suspicious that the networking problems have far more to do with IPv6 vs. 4 than the OS. > > Bill > > > > -----Original Message----- > From: pharo-project-bounces(a)lists.gforge.inria.fr > [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of > StÃ©phane Ducasse > Sent: Wednesday, August 26, 2009 7:21 AM > To: Pharo-project(a)lists.gforge.inria.fr > Subject: Re: [Pharo-project] new socket API > > so what is the next step? > > stef > > On Aug 26, 2009, at 8:19 AM, Michael Rueger wrote: > >> Philippe Marschall wrote: >> >>> There's gotta be a better way to do this. >>> >>> The whole API with arrays where you have to know the index, >>> ByteArray objects, integer flags and so forth seems like a blast from the past. >> >> We could put some default behavior wrapper methods around this. >> But this is really a discussion for squeak-dev, as the IPv6 >> implementation originates from etoys/olpc. >> >> Michael >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Aug. 26, 2009

Re: [Pharo-project] New Pharo based on core 10418

by Schwab,Wilhelm K

I just downloaded it. Re browsers and hierarchies, this indeed might be the solution for me. The win2k theme buttons are still sometimes reversed in state; I forget where that problem lives, but it's still with us. Right clicking in the browsers gives the #isClassNode and then #isMethodNode walkbacks. Bill -----Original Message----- From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of StÃ©phane Ducasse Sent: Wednesday, August 26, 2009 11:10 AM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] New Pharo based on core 10418 Youpi! Tx david/lukas/all :) Stef On Aug 26, 2009, at 6:00 PM, Damien Cassou wrote: > Here is a fresh new Pharo http://pharo-project.org/pharo-download > > This version contains the two browsers: one based on the categories > and one based on packages. To change the default one, open a browser > and click on the right-most button in the title bar. > > Happy Pharoing > > > -- > Damien Cassou > http://damiencassou.seasidehosting.st > > "Lambdas are relegated to relative obscurity until Java makes them > popular by not having them." James Iry > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Aug. 26, 2009

Re: [Pharo-project] System vs. hierachy browsers

by Schwab,Wilhelm K

Damien, It might be that the browser I am using on performance grounds is the source of the trouble. Bill -----Original Message----- From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Damien Cassou Sent: Wednesday, August 26, 2009 12:01 PM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] System vs. hierachy browsers On Wed, Aug 26, 2009 at 4:31 PM, Schwab,Wilhelm K<bschwab(a)anest.ufl.edu> wrote: > I am not sure how much of this is self-inflicted and how much is already fixed; perhaps someone can remind me of a solution that already exists. Â If not, think of it as a wish list. > > I have typically never liked so-called "system browsers" because they tend to show classes alphabetically and hide inheritance relationshiops; that's really important in trying to understand a design. Â The toolset I am using now is reasonably fast (in fact, it's perfectly fine on my new quad core box, but I hate throwing hardware at efficiency problems) but it not the bleeding edge. Â The system browsers do not show hierachies, and the [hierarchy] buttons spawn off something that often shows only one of the paths from a super class. Â IIRC, I've seen a browser that does better, but I can't remember which one it was or whether it can run fast enough to work on most of the hardware I see every day. > > Sorry for the complaining tone of the above; I don't mean it that way, but I also do not have time to fix it right now. I don't really understand your problem. In all Pharo images, the default System Browser orders classes of the selected package/category hierarchically and not alphabetically. -- Damien Cassou http://damiencassou.seasidehosting.st "Lambdas are relegated to relative obscurity until Java makes them popular by not having them." James Iry _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Aug. 26, 2009

Re: [Pharo-project] new socket API

by StÃ©phane Ducasse

code code code :) Stef On Aug 26, 2009, at 7:20 PM, Philippe Marschall wrote: > Philippe Marschall wrote: >> Lukas Renggli wrote: >>>> TCP sockets and UDP sockets have a different interface. Some >>>> messages >>>> can only be sent to TCP sockets others only to UDP sockets. Yet >>>> they are >>>> both represented by the same Socket class. >>> That sounds like a really easy and backward compatible refactoring >>> (add two subclasses, fix the constructors, and push down methods). >>> Why >>> don't you submit a patch? >> >> Yeah if it's as simple as that and it doesn't break any plugin or >> stuff. > > And we find a solution for SocksSocket which is of course a subclass > of > Socket. > > Cheers > Philippe > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Aug. 26, 2009

Re: [Pharo-project] new socket API

by Philippe Marschall

Philippe Marschall wrote: > Lukas Renggli wrote: >>> TCP sockets and UDP sockets have a different interface. Some messages >>> can only be sent to TCP sockets others only to UDP sockets. Yet they are >>> both represented by the same Socket class. >> That sounds like a really easy and backward compatible refactoring >> (add two subclasses, fix the constructors, and push down methods). Why >> don't you submit a patch? > > Yeah if it's as simple as that and it doesn't break any plugin or stuff. And we find a solution for SocksSocket which is of course a subclass of Socket. Cheers Philippe

Aug. 26, 2009

Re: [Pharo-project] new socket API

by Philippe Marschall

Lukas Renggli wrote: >> TCP sockets and UDP sockets have a different interface. Some messages >> can only be sent to TCP sockets others only to UDP sockets. Yet they are >> both represented by the same Socket class. > > That sounds like a really easy and backward compatible refactoring > (add two subclasses, fix the constructors, and push down methods). Why > don't you submit a patch? Yeah if it's as simple as that and it doesn't break any plugin or stuff. Cheers Philippe

Aug. 26, 2009

Re: [Pharo-project] tallyIt nows trigger a complete profiling

by Alexandre Bergel

I fully agree, but this is not really a new feature, but an improvement of something existing. It basically makes tallyIt useful. Alexandre On 26 Aug 2009, at 02:22, Damien Cassou wrote: > On Tue, Aug 25, 2009 at 5:16 PM, Alexandre > Bergel<alexandre(a)bergel.eu> wrote: >> This is much more useful I think. Before, it simply shows the time >> the >> expression took to execute. Now you have the information about >> different calls. >> I put the tag Milestone-1.0 >> http://code.google.com/p/pharo/issues/detail?id=1087 > > After beta new features should normally go to next Milestone => 1.1. > > -- > Damien Cassou > http://damiencassou.seasidehosting.st > > "Lambdas are relegated to relative obscurity until Java makes them > popular by not having them." James Iry > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Aug. 26, 2009

Re: [Pharo-project] [Vm-dev] Using VMMaker on Pharo

by Damien Cassou

On Wed, Aug 26, 2009 at 5:52 PM, jean baptiste arnaud<jbaptiste.arnaud(a)gmail.com> wrote: > just a few feedback, about VMMaker, on Pharo some plugin like (GeniePlugin) can't be used because, > when generate on pharo they raise a error "InvalidDirectoryError". > see with : > Last Image Core PharoCore-1.0-10414-BETA.zip. > clean install :Â FFI with ScriptLoader loadFFI. > install fromÂ squeaksource of Speech and VMMaker. > Hi Jean-Baptiste, could you please describe the whole process of building a VM from Pharo on the Pharo wiki (http://code.google.com/p/pharo/w/list)? Thank you -- Damien Cassou http://damiencassou.seasidehosting.st "Lambdas are relegated to relative obscurity until Java makes them popular by not having them." James Iry

Aug. 26, 2009

Re: [Pharo-project] new socket API

by Lukas Renggli

> TCP sockets and UDP sockets have a different interface. Some messages > can only be sent to TCP sockets others only to UDP sockets. Yet they are > both represented by the same Socket class. That sounds like a really easy and backward compatible refactoring (add two subclasses, fix the constructors, and push down methods). Why don't you submit a patch? Lukas > > Cheers > Philippe > > Schwab,Wilhelm K wrote: >> Stef, >> >> I am not sure exactly what should be next, but I do not think we can simply kick it back to squeak-dev any more than we can the GUI/IDE, streams, etc., as sockets are very important. Â I am slowly chipping away at things that are standing in the way of my first Pharo success story; the impact of sockets has been minimal, but that will change. Â When I get going, I will be looking carefully at how things behave when peers do not respond (only the calling thread should suffer). Â Dolphin's socket streams are nice; Pharo seems to lean toward a single read-write stream, which should be fine as long as it manages not to hobble the image when the wrong wire gets kicked. Â SSL via OpenSSL should be easy to use; that will be an interesting adventure, but it needs to happen. >> >> With all due respect to those who have disagreed, I'm still highly suspicious that the networking problems have far more to do with IPv6 vs. 4 than the OS. >> >> Bill >> >> >> >> -----Original Message----- >> From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of StÃ©phane Ducasse >> Sent: Wednesday, August 26, 2009 7:21 AM >> To: Pharo-project(a)lists.gforge.inria.fr >> Subject: Re: [Pharo-project] new socket API >> >> so what is the next step? >> >> stef >> >> On Aug 26, 2009, at 8:19 AM, Michael Rueger wrote: >> >>> Philippe Marschall wrote: >>> >>>> There's gotta be a better way to do this. >>>> >>>> The whole API with arrays where you have to know the index, ByteArray >>>> objects, integer flags and so forth seems like a blast from the past. >>> >>> We could put some default behavior wrapper methods around this. >>> But this is really a discussion for squeak-dev, as the IPv6 >>> implementation originates from etoys/olpc. >>> >>> Michael >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project -- Lukas Renggli http://www.lukas-renggli.ch

Aug. 26, 2009

Re: [Pharo-project] System vs. hierachy browsers

by Damien Cassou

On Wed, Aug 26, 2009 at 4:31 PM, Schwab,Wilhelm K<bschwab(a)anest.ufl.edu> wrote: > I am not sure how much of this is self-inflicted and how much is already fixed; perhaps someone can remind me of a solution that already exists. Â If not, think of it as a wish list. > > I have typically never liked so-called "system browsers" because they tend to show classes alphabetically and hide inheritance relationshiops; that's really important in trying to understand a design. Â The toolset I am using now is reasonably fast (in fact, it's perfectly fine on my new quad core box, but I hate throwing hardware at efficiency problems) but it not the bleeding edge. Â The system browsers do not show hierachies, and the [hierarchy] buttons spawn off something that often shows only one of the paths from a super class. Â IIRC, I've seen a browser that does better, but I can't remember which one it was or whether it can run fast enough to work on most of the hardware I see every day. > > Sorry for the complaining tone of the above; I don't mean it that way, but I also do not have time to fix it right now. I don't really understand your problem. In all Pharo images, the default System Browser orders classes of the selected package/category hierarchically and not alphabetically. -- Damien Cassou http://damiencassou.seasidehosting.st "Lambdas are relegated to relative obscurity until Java makes them popular by not having them." James Iry

Aug. 26, 2009