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

[Pharo-project] [update 2.0] #20340

by Marcus Denker

20340 ----- Issue 6806: More cleanups related to ChangeSorters http://code.google.com/p/pharo/issues/detail?id=6806 -- Marcus Denker -- http://marcusdenker.de

Oct. 14, 2012

Re: [Pharo-project] Do you know any code that can be better understood with its test?

by Stephan Eggermont

Stef wrote: >bitShift: > >look at the tests > > ((1 bitShift: 100) bitShift: -100) = 1 For even better understanding, add ((1 bitShift: -100) bitShift: 100) = 0 Stephan

Oct. 14, 2012

Re: [Pharo-project] EncryptedMultiDisplayScanner, are you serious?

by StÃ©phane Ducasse

http://code.google.com/p/pharo/issues/detail?id=6808 :) On Oct 14, 2012, at 1:22 AM, Igor Stasenko wrote: > EncryptedMultiDisplayScanner

Oct. 14, 2012

Re: [Pharo-project] Need more consistent #lines

by StÃ©phane Ducasse

http://code.google.com/p/pharo/issues/detail?id=6807 Stef > A total number of lines in string should be: > one + number of line separators found. > > Explanation: a 'String cr/ crlf' stands for 'line separator'. > By definition, a separator, is thing which separates two others.. i > mean, there is no way how you can use 'separator' term > without having TWO things which it separates. > > The problem is that #lines does not follows common sense: > > (String empty, String cr, String empty) lines size 1 > > so, our line separator "separates" one thing :) > It means that we overload the definition: > - a line separator separates two lines unless it is at the end of text. > > But think , how dearly this 'unless' will costs in terms of implementation! > Everywhere where you may need to handle text, you should put this > extra rule, and if you forget it, > you will be punished.. > > > Here what various #lines produces: > > ( 'A' ) lines size 1 > ( 'A', String cr ) lines size 1 !!oops!!. > > ( String cr, 'A' ) lines size 2 > ( 'A', String cr, 'A' ) lines size 2 > > ( 'A', String cr, 'A', String cr ) lines size 2 > ( 'A', String cr, 'A', String cr ) lines size 2 !!oops!! > > You may think this is fine, but lets look at the problem from another angle: > > self assert: (string1, String cr, string2) lineCount >= 2 > > The 'string1, String cr, string2' above is an insertion operation. > We inserting a line separator between two arbitrary strings. > Now, in what universe the total number of lines may not get increased > after such insertion? > > So, it is simply inconsistent. Or maybe you happy to put extra code > everywhere where you handling text, to do a special handling if last > character is line separator? > But we can avoid all that mess it in a first place, if we obey to common sense. > > -- > Best regards, > Igor Stasenko. >

Oct. 14, 2012

Re: [Pharo-project] EncryptedMultiDisplayScanner, are you serious?

by StÃ©phane Ducasse

On Oct 14, 2012, at 1:22 AM, Igor Stasenko wrote: > An EncryptedMultiDisplayScanner is a MultiDisplayScanner displaying > stars instead of characters (for password by example) > > Wow... this is soooo overkill! Indeed I was not aware of that system. > > Look at the amount of extra code we have just to simply displaying > stupid stars ... ************ > > See the: > EncryptedTextMorphForFieldView>>drawOn: > ... > aCanvas encryptedParagraph: self paragraph bounds: fauxBounds color: color. > > ehh?.. Since when canvas (a tool for drawing graphic primitives and > text) needs to know anything like that? > > I'm not sure if there could be better/simpler solution (considering > that we don't have a simple line editor and need to go over > overcomplicated text->composer->scanner->displayscanner chain to be > able to edit even single string.. > but that's really, really stupid to penetrate multiple layers of > responsibility just to display stars instead of characters... > > What if i would like to display '%' instead of chars, should i also go > and repeat same way?? > > We should get rid of this thing.. Yes. > > -- > Best regards, > Igor Stasenko. >

Oct. 14, 2012

Re: [Pharo-project] Article On Modern Smalltalk

by James Foster

GemStone should be included in a list of commercial offerings. James On Oct 13, 2012, at 6:24 PM, blake wrote: > Hello, folks-- > > I've been commissioned to write an article on modern Smalltalk and > wanted to solicit feedback on what people thought the most interesting > and useful aspects to cover would be. Obviously I wanted to talk about > Squeak, Pharo, Amber, and at least touch on Lively, and talk about the > commercial offerings (Dolphin, VAST, VW), and then there's Seaside and > GLORP and Spoon and Cuis and...etc. etc. etc. > > Since it's not a book, I can't possibly cover everything, so I want to > hit the highlights, and I'm trying to avoid being overly insular > (e.g., only talking about what are the highlights for me). > > Any and all comments welcome. > > ===Blake=== > >

Oct. 14, 2012

[Pharo-project] Need more consistent #lines

by Igor Stasenko

A total number of lines in string should be: one + number of line separators found. Explanation: a 'String cr/ crlf' stands for 'line separator'. By definition, a separator, is thing which separates two others.. i mean, there is no way how you can use 'separator' term without having TWO things which it separates. The problem is that #lines does not follows common sense: (String empty, String cr, String empty) lines size 1 so, our line separator "separates" one thing :) It means that we overload the definition: - a line separator separates two lines unless it is at the end of text. But think , how dearly this 'unless' will costs in terms of implementation! Everywhere where you may need to handle text, you should put this extra rule, and if you forget it, you will be punished.. Here what various #lines produces: ( 'A' ) lines size 1 ( 'A', String cr ) lines size 1 !!oops!!. ( String cr, 'A' ) lines size 2 ( 'A', String cr, 'A' ) lines size 2 ( 'A', String cr, 'A', String cr ) lines size 2 ( 'A', String cr, 'A', String cr ) lines size 2 !!oops!! You may think this is fine, but lets look at the problem from another angle: self assert: (string1, String cr, string2) lineCount >= 2 The 'string1, String cr, string2' above is an insertion operation. We inserting a line separator between two arbitrary strings. Now, in what universe the total number of lines may not get increased after such insertion? So, it is simply inconsistent. Or maybe you happy to put extra code everywhere where you handling text, to do a special handling if last character is line separator? But we can avoid all that mess it in a first place, if we obey to common sense. -- Best regards, Igor Stasenko.

Oct. 14, 2012

[Pharo-project] Article On Modern Smalltalk

by blake

Hello, folks-- I've been commissioned to write an article on modern Smalltalk and wanted to solicit feedback on what people thought the most interesting and useful aspects to cover would be. Obviously I wanted to talk about Squeak, Pharo, Amber, and at least touch on Lively, and talk about the commercial offerings (Dolphin, VAST, VW), and then there's Seaside and GLORP and Spoon and Cuis and...etc. etc. etc. Since it's not a book, I can't possibly cover everything, so I want to hit the highlights, and I'm trying to avoid being overly insular (e.g., only talking about what are the highlights for me). Any and all comments welcome. ===Blake===

Oct. 14, 2012

[Pharo-project] EncryptedMultiDisplayScanner, are you serious?

by Igor Stasenko

An EncryptedMultiDisplayScanner is a MultiDisplayScanner displaying stars instead of characters (for password by example) Wow... this is soooo overkill! Look at the amount of extra code we have just to simply displaying stupid stars ... ************ See the: EncryptedTextMorphForFieldView>>drawOn: ... aCanvas encryptedParagraph: self paragraph bounds: fauxBounds color: color. ehh?.. Since when canvas (a tool for drawing graphic primitives and text) needs to know anything like that? I'm not sure if there could be better/simpler solution (considering that we don't have a simple line editor and need to go over overcomplicated text->composer->scanner->displayscanner chain to be able to edit even single string.. but that's really, really stupid to penetrate multiple layers of responsibility just to display stars instead of characters... What if i would like to display '%' instead of chars, should i also go and repeat same way?? We should get rid of this thing.. -- Best regards, Igor Stasenko.

Oct. 13, 2012

Re: [Pharo-project] image format with shebang

by Camillo Bruni

On 2012-10-13, at 22:10, Stefan Marr <smalltalk(a)stefan-marr.de> wrote: > On 13 Oct 2012, at 20:44, Camillo Bruni wrote: > >> how hard would it be to change the image file to include a shebang on the first line? >> >> for instance: >> >> #!/usr/bin/env pharo -headless >> >> like that images would serve as "almost" script files? >> I image something like everything until the first 0 or >> newline byte is ignored when loading the image... > > It is supported, see for instance: http://marc.info/?l=squeak-vm-dev&m=124049248310307&w=2 awesome! strange that nobody promoted that so far... well I guess yet another of these hidden treasures ;)

Oct. 13, 2012