October 2012 - Pharo-users - lists.pharo.org

Re: [Pharo-users] [OT] RTalk

by dimitris chloupis

yes I thought I made this clear in my first post, the reason why I find RTalk so tempting is that it does not only offer the language but also the environment which AFAIK redline does not. Please correct me if I am mistaken, but I have not seen redline docs/wiki mention the enviroment anywhere and I have to say that smalltalk environment is pretty much the reason why I chose smalltalk and not so much the language. ________________________________ From: Geert Claes <geert.wl.claes(a)gmail.com> To: pharo-users(a)lists.gforge.inria.fr Sent: Tuesday, 30 October 2012, 12:54 Subject: Re: [Pharo-users] [OT] RTalk for Smalltalk on the JVM there is Redline Smalltalk I thought ... Ref: http://astares.blogspot.de/2012/10/talk-about-rtalk.html http://www.redline.st -- View this message in context: http://forum.world.st/OT-RTalk-tp4653286p4653297.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 30, 2012

Re: [Pharo-users] [OT] RTalk

by Geert Claes

for Smalltalk on the JVM there is Redline Smalltalk I thought ... Ref: http://astares.blogspot.de/2012/10/talk-about-rtalk.html http://www.redline.st -- View this message in context: http://forum.world.st/OT-RTalk-tp4653286p4653297.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 30, 2012

[OT] RTalk

by dimitris chloupis

I found this http://www.jarober.com/blog/blogView?showComments=true&printTitle=IM_98:_Rt… ,about RTalk , that is not only smalltalk run on top of JVM but almost the entire environment.Â The problem is that I cant seem to be able to find via google any site that contains the source code or something I could use ? Does anyone know if there is any official website for RTalk ? It looks quite promising and I would like to give it a try.

Oct. 30, 2012

Re: [Pharo-users] Zodiac long-term streaming

by Sean P. DeNigris

Sven Van Caekenberghe-2 wrote > Yes, I guess you could set an extreme timeout, like > ... timeout: 24*60*60; > ... > A real indefinite 'forever' timeout is considered bad form by many Sounds like we're on the same page... I'll keep the daily timeout, although I'll write "1 day asSeconds" because... well isn't that why we love Smalltalk ;) Thanks again, Sean -- View this message in context: http://forum.world.st/Zodiac-long-term-streaming-tp4653138p4653159.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 28, 2012

Re: [Pharo-users] Zodiac long-term streaming

by Sven Van Caekenberghe

On 27 Oct 2012, at 18:39, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > My image is connected to an ssl streaming server with Zodiac... Okay, right > off the bat, how cool is that?! Thanks, Sven! You're welcome. It is however the result of a collective effort by many people. > Anyway, using "ZnClient streaming: true; post: 'https://', url contents: > data", I have a ZdcSecureSocketStream waiting for more data. Is there a way > to specify "no timeout"? Since the connection is only good for 24 hours, I > was thinking to just set the timeout to be a little longer than that, but I > was curious. Yes, I guess you could set an extreme timeout, like ZnClient new timeout: 24*60*60; â¦ A real indefinite 'forever' timeout is considered bad form by many. Having a defined timeout and then looping expresses your intention clearer, makes you aware of the infinite loop and gives you a chance to do something useful (logging, resetting stuff, ..). > Thanks, > Sean -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Oct. 27, 2012

Zodiac long-term streaming

by Sean P. DeNigris

My image is connected to an ssl streaming server with Zodiac... Okay, right off the bat, how cool is that?! Thanks, Sven! Anyway, using "ZnClient streaming: true; post: 'https://', url contents: data", I have a ZdcSecureSocketStream waiting for more data. Is there a way to specify "no timeout"? Since the connection is only good for 24 hours, I was thinking to just set the timeout to be a little longer than that, but I was curious. Thanks, Sean -- View this message in context: http://forum.world.st/Zodiac-long-term-streaming-tp4653138.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 27, 2012

Re: [Pharo-users] garbage collection questions

by Igor Stasenko

On 27 October 2012 01:23, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > On 26 Oct 2012, at 21:41, Paul DeBruicker <pdebruic(a)gmail.com> wrote: > >> 2. If I have two objects with accessors: >> >> Object subclass: #ObjectA >> instanceVariableNames: 'objectB' >> classVariableNames: '' >> poolDictionaries: '' >> category: 'Temp' >> >> Object subclass: #ObjectB >> instanceVariableNames: 'objectA' >> classVariableNames: '' >> poolDictionaries: '' >> category: 'Temp' >> >> and create an instance of each in a workspace and then set them up: >> >> |a b| >> a:=ObjectA new. >> b:=ObjectB new. >> a objectB: b. >> b objectA: a. >> >> and close the workspace then those instances of ObjectA & ObjectB will never be garbage collected, correct? Its my understanding that I'd need to nil out the instance variables of both instances to have them GCd as they both refer strongly to each other. Is that right? > > That's a circular reference and should be no problem for the garbage collector. > Unless something else holds either a or b, like the workspace, an inspector, a debugger. > or "restore workspace contents" utility. > -- > Sven Van Caekenberghe > http://stfx.eu > Smalltalk is the Red Pill -- Best regards, Igor Stasenko.

Oct. 26, 2012

Re: [Pharo-users] garbage collection questions

by Sven Van Caekenberghe

On 26 Oct 2012, at 21:41, Paul DeBruicker <pdebruic(a)gmail.com> wrote: > 2. If I have two objects with accessors: > > Object subclass: #ObjectA > instanceVariableNames: 'objectB' > classVariableNames: '' > poolDictionaries: '' > category: 'Temp' > > Object subclass: #ObjectB > instanceVariableNames: 'objectA' > classVariableNames: '' > poolDictionaries: '' > category: 'Temp' > > and create an instance of each in a workspace and then set them up: > > |a b| > a:=ObjectA new. > b:=ObjectB new. > a objectB: b. > b objectA: a. > > and close the workspace then those instances of ObjectA & ObjectB will never be garbage collected, correct? Its my understanding that I'd need to nil out the instance variables of both instances to have them GCd as they both refer strongly to each other. Is that right? That's a circular reference and should be no problem for the garbage collector. Unless something else holds either a or b, like the workspace, an inspector, a debugger. -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Oct. 26, 2012

garbage collection questions

by Paul DeBruicker

1. Is there a wiki entry about the GC somewhere that provides advice about how to tune the GC using #setGCParameters? 2. If I have two objects with accessors: Object subclass: #ObjectA instanceVariableNames: 'objectB' classVariableNames: '' poolDictionaries: '' category: 'Temp' Object subclass: #ObjectB instanceVariableNames: 'objectA' classVariableNames: '' poolDictionaries: '' category: 'Temp' and create an instance of each in a workspace and then set them up: |a b| a:=ObjectA new. b:=ObjectB new. a objectB: b. b objectA: a. and close the workspace then those instances of ObjectA & ObjectB will never be garbage collected, correct? Its my understanding that I'd need to nil out the instance variables of both instances to have them GCd as they both refer strongly to each other. Is that right? 3. How can I learn when an incremental or full GC run will happen? Is it dependent upon on the size of new or old memory or the overall size of the image or ... ? It seems like putting a #halt in a random spot could easily lead to a stuck image. Thanks for your help Paul

Oct. 26, 2012

Re: [Pharo-users] It's time to play... "Decode This"

by Sean P. DeNigris

Sean P. DeNigris wrote > A REST API passed me #[0 0 1 58 153 199 12 211], which it swears is a unix > timestamp My bad. I RTFM, but not so well "time in *milliseconds* from 00:00:00 UTC on January 1, 1970", so I have to do something like: DateAndTime fromUnixTime: #[0 0 1 58 153 199 12 211] asInteger / 1000. Obviously need to stop hacking for today ;) Cheers, Sean -- View this message in context: http://forum.world.st/It-s-time-to-play-Decode-This-tp4652979p4652989.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 26, 2012