- Pharo-dev - lists.pharo.org

[Pharo-project] [ANN] Pharo Sprint, 26.10

by Marcus Denker

Hi, We are organizing a Pharo Sprint (that is: a hacking session) about Pharo this Sunday. Date: 26.10 Time: Starting 11h Where? Lille, France Depending on how many people want to come. Likely it will be at Stef's place. Please send me an email ASAP if you want to join. Then we decide about the place and organize some food. Marcus -- Marcus Denker -- denker(a)iam.unibe.ch http://www.iam.unibe.ch/~denker

Oct. 22, 2008

Re: [Pharo-project] Size of objects

by StÃ©phane Ducasse

may be compare with the vw version I remember that lomg time ago in book of the guy doing june there was some code to do that http://www.sra.co.jp/people/aoki/ in freedocument then Smalltalk Idioms in English then chapter 3 Here let me introduce a program that finds the size in bytes of an arbitrary object. Finding the size of an object is the quickest route toward understanding the internal structure of objects. "Calculate the size in bytes of an arbitrary object." | byteSizeOfObject | byteSizeOfObject := [:anObject | | bytesInOTE bytesInOOP aClass indexableFieldSize instVarFieldSize byteSize | bytesInOTE := ObjectMemory current bytesPerOTE. bytesInOOP := ObjectMemory current bytesPerOOP. aClass := anObject class. aClass isPointers ifTrue: [instVarFieldSize := aClass instSize * bytesInOOP. aClass isVariable ifTrue: [indexableFieldSize := anObject basicSize * bytesInOOP] ifFalse: [indexableFieldSize := 0]] ifFalse: [instVarFieldSize := 0. aClass isVariable ifTrue: [indexableFieldSize := anObject basicSize + (bytesInOOP - 1) bitAnd: bytesInOOP negated] ifFalse: [indexableFieldSize := 0. aClass hasImmediateInstances ifTrue: [bytesInOTE := 0]]]. byteSize := bytesInOTE + instVarFieldSize + indexableFieldSize. byteSize yourself]. ^byteSizeOfObject value: (OrderedCollection new: 10) have a look at the book. now in VW numOopsNumBytes "Answer an Array containing the number of objects in the system and the number of bytes they occupy. This method is implementation dependent." "ObjectMemory current numOopsNumBytes" | nobjects nbytes class bytesInHeader bytesInOOP | nobjects := nbytes := 0. bytesInHeader := self bytesPerOTE. bytesInOOP := self bytesPerOOP. self class allObjectsDo: [:object | nobjects := nobjects + 1. nbytes := nbytes + bytesInHeader. (class := object class) isBits ifTrue: [nbytes := nbytes + (object basicSize + (bytesInOOP - 1) bitAnd: bytesInOOP negated)] ifFalse: [nbytes := nbytes + (class instSize * bytesInOOP). class isVariable ifTrue: [nbytes := nbytes + (object basicSize * bytesInOOP)]]]. ^Array with: nobjects with: nbytes On Oct 21, 2008, at 6:15 PM, Adrian Lienhard wrote: > Hi FrÃ©dÃ©ric > > Nice to see you on the Pharo mailing list! > > Just the other day I needed this too and I implemented the following > method: > > Object>>sizeInMemory > "Returns the number of bytes used by this object in memory > (including its header)" > > | headerSize instanceSize | > headerSize := (self class indexIfCompact > 0 ifTrue: [ 4 ] ifFalse: > [ 8 ]). > instanceSize := (self class isVariable > ifFalse: [ self class instSize * Smalltalk wordSize ] > ifTrue: [ (self basicSize * (self class isBytes > ifTrue: [ 1 ] ifFalse: [ Smalltalk wordSize ])) ]). > ^ headerSize + instanceSize > > > BTW, I think that would be a meaningful addition. Could somebody > quickly check the method to verify that I didn't do any > miscalculation? > > Cheers, > Adrian > > On Oct 21, 2008, at 17:58 , Frederic Pluquet wrote: > >> Hello, >> >> For an experimentation, I need to know the size in bytes of any >> object. Is there a way to obtain this information in Pharo ? >> >> FrÃ©d >> >> -- >> FrÃ©dÃ©ric Pluquet >> UniversitÃ© Libre de Bruxelles (ULB) >> Assistant >> http://www.ulb.ac.be/di/fpluquet >> _______________________________________________ >> 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 >

Oct. 22, 2008

Re: [Pharo-project] Looking for ImageSegment tests

by Adrian Lienhard

On Oct 22, 2008, at 22:04 , StÃ©phane Ducasse wrote: > > On Oct 21, 2008, at 5:25 PM, Adrian Lienhard wrote: > >> >> On Oct 21, 2008, at 16:05 , StÃ©phane Ducasse wrote: >> >>> Hi all >>> >>> if by any chance you have some code that could be turned into >>> tests, or >>> some tests or some time to write tests for image segments this >>> would be great >>> because etoy is really deep inside image segment and I would >>> really like to keep this >>> beautiful idea working :) >> >> yes please. We use this in production (like DabbleDB), and it works >> very well. I just hope to never really need to understand the code... > > really read the code because it is full of etoy related ugly hack > (really ugly) I know... > I would not rely on that. > So we should really write some tests. This is really important. yes, especially if we make changes in this area. I fear this a bit... > I never used imageSegment but someone with experience should do that. > at least start. I'll add it to my todo list Adrian

Oct. 22, 2008

Re: [Pharo-project] Looking for ImageSegment tests

by StÃ©phane Ducasse

On Oct 21, 2008, at 5:25 PM, Adrian Lienhard wrote: > > On Oct 21, 2008, at 16:05 , StÃ©phane Ducasse wrote: > >> Hi all >> >> if by any chance you have some code that could be turned into >> tests, or >> some tests or some time to write tests for image segments this >> would be great >> because etoy is really deep inside image segment and I would really >> like to keep this >> beautiful idea working :) > > yes please. We use this in production (like DabbleDB), and it works > very well. I just hope to never really need to understand the code... really read the code because it is full of etoy related ugly hack (really ugly) I would not rely on that. So we should really write some tests. This is really important. I never used imageSegment but someone with experience should do that. at least start. stef > > > Adrian > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 22, 2008

[Pharo-project] changes file

by John M McIntosh

I note I downloaded the current pharo image/changes for pushing out on the iPhone and noticed the changes was 10.5MB, but if I condense it becomes 4.2MB which is much easier to deal with if I include it in an iPhone.app I'm wondering here if you need to push out the changes with all the method versions in it, or should it be condenced first and rely on changes from MC to show the history of how the methods evolved? -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

Oct. 22, 2008

Re: [Pharo-project] Merging one-click and pharo-dev

by John M McIntosh

That is preserving any resource fork, and meta-data from hfs file attributes. Also so retains the hidden file that is used by the finder to format the folder view. It doesn't need to be present I believe in order for the re-zipped app to run, however I'd just keep it for now. After you replace the image/changes and other files, just rezip it all. On the mac unzipping it and the unarchive will do the right thing with the __MACOS__ files/folders On Oct 22, 2008, at 12:05 PM, Lukas Renggli wrote: >> However if you are using linux then you should be fine, the >> Squeak.app is >> just a folder with a bunch of files/folders. > > But there is a whole bunch of hidden files in these directories. I > guess these must be preserved? > > Also the ZIP archive contains a strange hidden __MACOS__ folder. > What is it for? > > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

Oct. 22, 2008

[Pharo-project] Returning something when debugging

by Damien Cassou

I've just discovered something today: when you are debugging you can always return from the current method by right-clicking on the stack frame and selecting 'return entered value'. You then get an input field where you can enter any expression and the result of that expression will be evaluated as the return value of the stack frame. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu)

Oct. 22, 2008

Re: [Pharo-project] Merging one-click and pharo-dev

by Keith Hodges

Lukas Renggli wrote: >> However if you are using linux then you should be fine, the Squeak.app is >> just a folder with a bunch of files/folders. >> It would be good to have a SakeTask for generating one click images, that can be incorporated into Bob. Keith

Oct. 22, 2008

Re: [Pharo-project] Merging one-click and pharo-dev

by Lukas Renggli

> However if you are using linux then you should be fine, the Squeak.app is > just a folder with a bunch of files/folders. But there is a whole bunch of hidden files in these directories. I guess these must be preserved? Also the ZIP archive contains a strange hidden __MACOS__ folder. What is it for? Lukas -- Lukas Renggli http://www.lukas-renggli.ch

Oct. 22, 2008

Re: [Pharo-project] Merging one-click and pharo-dev

by Damien Cassou

On Wed, Oct 22, 2008 at 9:29 AM, Marcus Denker <denker(a)iam.unibe.ch> wrote: > In general it would be nice to be able to build the one-click > programmatically. > There should be a "one-click" package-generator in the image. It should be easy in Bash, I can try. However, I don't know how it can be done in Squeak. Why do you need to do it in Squeak? -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu)

Oct. 22, 2008