December 2011 - Pharo-dev - lists.pharo.org

[Pharo-project] [update 1.4] #14244

by Marcus Denker

14244 ----- Issue 5048: Move Transcript to Tools Package http://code.google.com/p/pharo/issues/detail?id=5048 Issue 5047: Stream should not print its contents in printOn: http://code.google.com/p/pharo/issues/detail?id=5047 Issue 5053: ZnChunkedReadStream doesNotUnderstand: #next:into: http://code.google.com/p/pharo/issues/detail?id=5053 -- Marcus Denker -- http://marcusdenker.de

Dec. 4, 2011

Re: [Pharo-project] VM freezes sending #basicNew to Stream subclass

by StÃ©phane Ducasse

I agree with you. Nevertheless it seems that there is a bug in the debugger. > In a workspace, type this and debug it: > Stream basicNew printString. > > Step over basicNBew. > Step into printString. > Step into printStringLimitedTo: 50000 > Step into streamContents: [:s | self printOn: s] limitedTo: limit > Step over 100 min: sizeLimit > Step over self new: (100 min: sizeLimit) > --> error: the debugger tries to execute (String new: String) > > I also note that all class methods in stack window left column are > miss-printed - like 'Block in class)' new: Stef On Dec 4, 2011, at 12:07 AM, Igor Stasenko wrote: > On 3 December 2011 01:30, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >> We will have to track when this error was introduced may be this is the runtimeError fix that we pushed. >> We should check on images to spot this regression. >> >> Now printing the contents of a stream by default looks to me a bad idea. >> May be when the stream is positionable it makes sense but why would I force to read the complete contents of a stream >> when I open an inspector on it. >> >> > > I think that 'contents' concept in streams are controversial. > Just consider: > - a stream which generates random numbers. every time you send #next, > you get another random number. what is a contents of such stream? > all random numbers it can generate till universe die? > > - a real-time stream broadcasting video/audio. what is a contents of > such stream? > you cannot capture whole contents of it, because these kinds of > streams has no 'end'. > > It is because streams are for processing data in portions, and when > you requesting contents of it, you turning it into collection. > > Here the Collection class comment: > " > I am the abstract superclass of all classes that represent a group of elements. > " > > And Stream: > " > I am an abstract class that represents an accessor for a sequence of > objects. This sequence is referred to as my "contents". > " > Yes. Of course, abstractly speaking, all streams containing something > , but you can access it only by portions. > Otherwise they will be no different to collections and therefore we > won't need them. > > -- > Best regards, > Igor Stasenko. >

Dec. 4, 2011

Re: [Pharo-project] Pharo VMs should be more Pharo-like

by Stefan Marr

Hi: On 04 Dec 2011, at 03:47, Esteban Lorenzano wrote: >> 2. Is it possible to have a 'smart' version of the fullscreen feature, at least on Snow Leopard? >> Currently, it is blanking out my second screen. I read that VMware Fusion introduce such a 'smart' full screen functionality, that does not blank out the other screens. > > I don't understand this... what you mean? > there is a full-screen mode already in latest builds... and it is working fine for me. There seem to be people who agree that making all other displays just black is not a useful feature. See what they advertise for 4.1: http://blogs.vmware.com/teamfusion/2011/11/weekly-tips-and-tricks-whats-new… Being able to use the second screen for something is something I care about. And it was possible with the old fullscreen mode. Thus, I consider the new behavior a bug that is worthwhile to be fixed. Thanks Stefan -- Stefan Marr Software Languages Lab Vrije Universiteit Brussel Pleinlaan 2 / B-1050 Brussels / Belgium http://soft.vub.ac.be/~smarr Phone: +32 2 629 2974 Fax: +32 2 629 3525

Dec. 4, 2011

Re: [Pharo-project] VM freezes sending #basicNew to Stream subclass

by Sven Van Caekenberghe

+100 On 04 Dec 2011, at 00:07, Igor Stasenko wrote: > On 3 December 2011 01:30, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >> We will have to track when this error was introduced may be this is the runtimeError fix that we pushed. >> We should check on images to spot this regression. >> >> Now printing the contents of a stream by default looks to me a bad idea. >> May be when the stream is positionable it makes sense but why would I force to read the complete contents of a stream >> when I open an inspector on it. >> >> > > I think that 'contents' concept in streams are controversial. > Just consider: > - a stream which generates random numbers. every time you send #next, > you get another random number. what is a contents of such stream? > all random numbers it can generate till universe die? > > - a real-time stream broadcasting video/audio. what is a contents of > such stream? > you cannot capture whole contents of it, because these kinds of > streams has no 'end'. > > It is because streams are for processing data in portions, and when > you requesting contents of it, you turning it into collection. > > Here the Collection class comment: > " > I am the abstract superclass of all classes that represent a group of elements. > " > > And Stream: > " > I am an abstract class that represents an accessor for a sequence of > objects. This sequence is referred to as my "contents". > " > Yes. Of course, abstractly speaking, all streams containing something > , but you can access it only by portions. > Otherwise they will be no different to collections and therefore we > won't need them. > > -- > Best regards, > Igor Stasenko. >

Dec. 4, 2011

Re: [Pharo-project] Pharo VMs should be more Pharo-like

by Sean P. DeNigris

EstebanLM wrote > >> 2. Is it possible to have a 'smart' version of the fullscreen feature, at >> least on Snow Leopard? >> Currently, it is blanking out my second screen. I read that VMware >> Fusion introduce such a 'smart' full screen functionality, that does not >> blank out the other screens. > I don't remember about Snow Leopard, but this is how Lion's fullscreen mode (stupidly) works. -- View this message in context: http://forum.world.st/Pharo-VMs-should-be-more-Pharo-like-tp4119042p4156011… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Dec. 4, 2011

Re: [Pharo-project] Pharo VMs should be more Pharo-like

by Esteban Lorenzano

Hi, El 03/12/2011, a las 9:38p.m., Stefan Marr escribiÃ³: > 1. there is still the SqueakOSXApp string all over the menu (CogVM [..] 32bits). yeah... I will solve that soon :) > 2. Is it possible to have a 'smart' version of the fullscreen feature, at least on Snow Leopard? > Currently, it is blanking out my second screen. I read that VMware Fusion introduce such a 'smart' full screen functionality, that does not blank out the other screens. I don't understand this... what you mean? there is a full-screen mode already in latest builds... and it is working fine for me. best, Esteban

Dec. 4, 2011

Re: [Pharo-project] Pharo VMs should be more Pharo-like

by Stefan Marr

Hi Esteban: On 04 Dec 2011, at 00:19, Esteban Lorenzano wrote: > Take a look: https://ci.lille.inria.fr/pharo/view/VM/job/Pharo-Mac-Cocoa/ Looks good. But I noticed two things: 1. there is still the SqueakOSXApp string all over the menu (CogVM [..] 32bits). 2. Is it possible to have a 'smart' version of the fullscreen feature, at least on Snow Leopard? Currently, it is blanking out my second screen. I read that VMware Fusion introduce such a 'smart' full screen functionality, that does not blank out the other screens. Thanks Stefan -- Stefan Marr Software Languages Lab Vrije Universiteit Brussel Pleinlaan 2 / B-1050 Brussels / Belgium http://soft.vub.ac.be/~smarr Phone: +32 2 629 2974 Fax: +32 2 629 3525

Dec. 4, 2011

Re: [Pharo-project] Pharo VMs should be more Pharo-like

by Esteban Lorenzano

Hi, So, I took "the bull by the horns" and hacked a jenkins job to do a Pharo app. I also created a new squeaksource project (PharoVM) to keep the specific configurations for Pharo. Right now, I made it just for Macs, but if everything is ok, I'll add a job for Linux and when windows slave is working, also for windows. Take a look: https://ci.lille.inria.fr/pharo/view/VM/job/Pharo-Mac-Cocoa/ Cheers, Esteban El 29/11/2011, a las 9:56p.m., Serge Stinckwich escribiÃ³: > On Tue, Nov 29, 2011 at 9:46 PM, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >> Hi, >> >> Currently, we are linking official VMs for pharo taking directly from Jenkins a version and packaging for download (or for Pharo One-Click). >> I think this is a temporal solution, we need something better: >> >> 1) Pharo VMs should have Pharo Icons. >> 2) Pharo VMs should be named "pharo" or "pharovm" (better the fist), not "CogVM", which is an internal name. >> 3) We should include a PLUGINS.txt file with information about present plugins (internal and external) >> 4) We should include a LICENCE.txt file with all licence information. >> >> ... I think that's all :) >> >> >> >> I know... this is very dumb and clearly not mandatory, but it says a lot about the maturity of our Pharo :) >> >> All agree? Should I create a ticket? > > Yes, nobody could disagree with that ;-) > >> Cheers, >> Esteban >> >> pd: I'm the only one who does not likes One-Click versions? I would prefer a different package for each platform (Is ugly to have a directory named Pharo.app in linux or windows. Also, no need to have all those duplicated-but-for-different-platform files) > > Most of the time, i'm using One-click images. This is really > convenient to move images between machines and give them to friends. > If the process of creating releases is automated, both can be provided. > Regards, > -- > Serge Stinckwich > UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam > Matsuno Laboratory, Kyoto University, Japan (until 12/2011) > http://www.mechatronics.me.kyoto-u.ac.jp/ > Every DSL ends up being Smalltalk > http://doesnotunderstand.org/ >

Dec. 3, 2011

Re: [Pharo-project] VM freezes sending #basicNew to Stream subclass

by Igor Stasenko

On 3 December 2011 01:30, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > We will have to track when this error was introduced may be this is the runtimeError fix that we pushed. > We should check on images to spot this regression. > > Now printing the contents of a stream by default looks to me a bad idea. > May be when the stream is positionable it makes sense but why would I force to read the complete contents of a stream > when I open an inspector on it. > > I think that 'contents' concept in streams are controversial. Just consider: - a stream which generates random numbers. every time you send #next, you get another random number. what is a contents of such stream? all random numbers it can generate till universe die? - a real-time stream broadcasting video/audio. what is a contents of such stream? you cannot capture whole contents of it, because these kinds of streams has no 'end'. It is because streams are for processing data in portions, and when you requesting contents of it, you turning it into collection. Here the Collection class comment: " I am the abstract superclass of all classes that represent a group of elements. " And Stream: " I am an abstract class that represents an accessor for a sequence of objects. This sequence is referred to as my "contents". " Yes. Of course, abstractly speaking, all streams containing something , but you can access it only by portions. Otherwise they will be no different to collections and therefore we won't need them. -- Best regards, Igor Stasenko.

Dec. 3, 2011

Re: [Pharo-project] VM freezes sending #basicNew to Stream subclass

by StÃ©phane Ducasse

Thanks I entered this description http://code.google.com/p/pharo/issues/detail?id=5054 And I will start to see if I can identify what changed using dichotomyâ¦. On Dec 2, 2011, at 7:37 PM, Nicolas Cellier wrote: > Yes. With user interaction... > > In a workspace, type this and debug it: > Stream basicNew printString. > > Step over basicNBew. > Step into printString. > Step into printStringLimitedTo: 50000 > Step into streamContents: [:s | self printOn: s] limitedTo: limit > Step over 100 min: sizeLimit > Step over self new: (100 min: sizeLimit) > --> error: the debugger tries to execute (String new: String) > > I also note that all class methods in stack window left column are > miss-printed - like 'Block in class)' new:

Dec. 3, 2011