- Pharo-users - lists.pharo.org

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

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

by Sean P. DeNigris

Igor Stasenko wrote > looks like 64-bit value, which fits the size of time_t (which some ppl > refer as unix timestamp) Googling now... it's really strange... if I send #asInteger, it looks *really* like a unix timestamp, but shifted 2 to the left... -- View this message in context: http://forum.world.st/It-s-time-to-play-Decode-This-tp4652979p4652982.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 25, 2012

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

by Igor Stasenko

On 26 October 2012 00:55, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > A REST API passed me #[0 0 1 58 153 199 12 211], which it swears is a unix > timestamp within the last day... I don't see how that's possible... Any > ideas? > looks like 64-bit value, which fits the size of time_t (which some ppl refer as unix timestamp) > Thanks, > Sean > > > > -- > View this message in context: http://forum.world.st/It-s-time-to-play-Decode-This-tp4652979.html > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. > -- Best regards, Igor Stasenko.

Oct. 25, 2012

It's time to play... "Decode This"

by Sean P. DeNigris

A REST API passed me #[0 0 1 58 153 199 12 211], which it swears is a unix timestamp within the last day... I don't see how that's possible... Any ideas? Thanks, Sean -- View this message in context: http://forum.world.st/It-s-time-to-play-Decode-This-tp4652979.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 25, 2012

Re: [Pharo-users] Regex search and replace with groups

by Davide Varvello

:-) yep, Peter, it works. I was looking for something semantically related to grouping, but your solution do the stuff Thanks Davide Peter H. Meadows wrote > How about: > > 'abc123/456def/ghi'copyWithRegex: '(\d)/(\d)' matchesTranslatedUsing: > [ :each | each first asString , each third asString ]. > > Not sure if that's the best way, but it seems to work. -- View this message in context: http://forum.world.st/Regex-search-and-replace-with-groups-tp4652801p465289… Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Oct. 25, 2012