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

Re: [Pharo-project] RPackage registration mess

by StÃ©phane Ducasse

> My bad :/. > > BTW, we have to introduce one method to restore the system subscriptions to a good state whatever was the state before. And that way we do not mess it so easily again. YES!

July 26, 2012

Re: [Pharo-project] markdown to latext and html

by StÃ©phane Ducasse

> The Pier Book offers this pretty much out of the box. I know, I wrote the seaside book with it and I would like to stick with svn and my default latex tools. For example sven wants to have a copy of his text on his server and I go much slower on a pier book when I cannot work in the train or not connected. Stef

July 26, 2012

Re: [Pharo-project] [ANN] Android Dr. Geo 12.07 release

by StÃ©phane Ducasse

Excellent!!!! Stef On Jul 26, 2012, at 9:11 AM, Hilaire Fernandes wrote: > Dmitry and Hilaire are happy to announce you the Android Dr. Geo 12.07 > release. > It comes with a working Freetype plugin for Android. > > http://www.drgeo.eu/home/dr-geo-news/androiddrgeo1207summertimerelease > > Enjoy, > > Thanks > > >

July 26, 2012

Re: [Pharo-project] How to filter out method extensions in Ring?

by StÃ©phane Ducasse

No since this is not available in the image. Now if you implement it somewhere we could set the package information somewhere. On Jul 25, 2012, at 8:25 PM, HernÃ¡n Morales Durand wrote: > Thanks VerÃ³nica, now it works fine. > > I have one more question :) it is possible with Ring to query if a package is a "system" package? (included in the downloaded image). We've been talking about that feature in the list http://lists.gforge.inria.fr/pipermail/pharo-project/2009-December/018168.h… but no solution was proposed. > > HernÃ¡n > > 2012/7/25 Veronica Isabel Uquillas Gomez <vuquilla(a)vub.ac.be> > Hi Hernan, > > Sorry, I have been totally disconnected because of my thesis. > > Ok I quickly reply... > You need to instantiate methods with #realClass:selector: to manipulate objects from the image.. > otherwise, they are static (within a ring source model). The category is not assigned then is not possible to know if they are from a class extension. > > ArrayedCollection selectors asSortedCollection collect: [ : each | > RGMethodDefinition > realClass: ArrayedCollection > selector: each ] > > then you can use: aRGMethod isExtension > > Cheers, > Veronica > > On 23 Jul 2012, at 13:45, HernÃ¡n Morales Durand wrote: > > > Hello all, > > You may help me on this one. I want to filter out method selectors which are extensions of a particular class. For example, in Pharo 1.4 the ArrayedCollection class has these selectors: > > ArrayedCollection selectors asSortedCollection > > > > I would want all of them except #writeOnGZIPByteStream: which is in "Compression" package. I did a try to Ring > > > > ArrayedCollection selectors asSortedCollection collect: [ : each | > > RGMethodDefinition > > class: ArrayedCollection > > selector: each ] > > > > with MethodReference I did : > > > > ... ( each category beginsWith: '*' ) not ... > > > > but it seems not supported in Ring? Any hints? > > > > Cheers, > > > > HernÃ¡n > > > >

July 26, 2012

Re: [Pharo-project] a little pearl from SystemProgressMorph

by StÃ©phane Ducasse

And I removed the strip behavior. In 2.0 it is not there. Stef On Jul 25, 2012, at 4:47 PM, Tudor Girba wrote: > Just for fun read the below piece of code :) > > > SystemProgressMorph>>label: shortDescription min: startMinValue max: > startMaxValue > "Answer the block that updates the progress bar." > > "some fun stuff added (by kph) > > - bar value: #label. - tell me my current label. > - bar value: 'newLabel'. - enable changing the label from within the workBlock > - bar value: #increment. - enable progress by one without keeping a counter > - bar value: #decrement. - go backwards (if you really have to, > useful for an abort, or rollback)! > > - bar value: newBigNum. - change the max on the fly - when you find > there is more to do. > - bar value: (bar value: #setMax) + 20 - change the max on the fly - > when you find there is more/less to do. > - bar value: (bar value: #setMin) - 20 - change the min on the fly - > not sure why you would want to. > " > | slot range barSize lastRefresh maxValue minValue bar| > maxValue := startMaxValue. > minValue := startMinValue. > ((range := maxValue - minValue) <= 0 or: [(slot := self nextSlotFor: > shortDescription) = 0]) > ifTrue: [^[:barVal| 0 ]]. > self recenter. > self openInWorld. > barSize := -1. "Enforces a inital draw of the morph" > lastRefresh := Time millisecondClockValue. > bar := bars at: slot. > ^[:barValArg | | barVal return newBarSize | > barVal := barValArg. > return := nil. > bar := bars at: slot. > "new fun stuff here" > barVal == #current ifTrue: [ return := barSize ]. > barVal == #label ifTrue:[ return := (labels at: slot) contents ]. > barVal == #setMax ifTrue: [ return := maxValue. maxValue := minValue ]. > barVal == #setMin ifTrue: [ return := minValue. minValue := maxValue ]. > barVal == #increment ifTrue: [return := barVal := barSize + 1]. > barVal == #decrement ifTrue: [ return := barVal := barSize - 1]. > (barVal isString and: [barVal isSymbol not]) ifTrue: [ > (labels at: slot) contents: barVal. > barVal := #refresh]. > barVal == #refresh ifTrue: [self currentWorld displayWorld. return := true]. > (barVal == SmallInteger maxVal or: [ barVal == #finished ]) ifTrue: > [return := slot]. > return ifNil: [ > barVal > maxValue ifTrue: [return := maxValue := barVal]. > barVal < minValue ifTrue: [return := minValue := barVal]. > (barVal between: minValue and: maxValue) > ifTrue: [newBarSize := (barVal - minValue / range * BarWidth) truncated. > newBarSize = barSize > ifFalse: [barSize := newBarSize. > (Time millisecondsSince: lastRefresh) > 25 > ifTrue: [barVal := #refresh ]]]. > barVal == #refresh ifTrue: [ > ((bar valueOfProperty: #useStripes) ifNil: [false]) > ifTrue: [bar fillStyle origin: bar position - ((Time > millisecondClockValue // 50 \\ 30) @ 0)]. > bar barSize: barSize. > self currentWorld displayWorld. > lastRefresh := Time millisecondClockValue]]. > return] > > Cheers, > Doru > > -- > www.tudorgirba.com > > "Every thing has its own flow" >

July 26, 2012

Re: [Pharo-project] Way to see diffs between different builds?

by StÃ©phane Ducasse

We should have a database with ring objects and perform queries. I know that hernan and ezequiel are working on that. Stef >> Hi, >> >> Is there a way to see the diffs (in terms of changes to image) between one >> 2.0 build number and another? I don't see any obvious link on the CI server >> to it, and the github export seems to be done in a way that makes the diffs >> enourmous. > > you can use the git repos and filter out non-st methods: > git diff $VERSION_A $VERSION_B -- $(git diff --name-only $VERSION_A $VERSION_B | grep .st) > > Git even has a utility to find the version that introduced a certain bug :) > http://incitecode.com/2009/10/13/finding-bugs-with-git-bisect/ > > >> If not, is there a way to do it in image? I can only see changes for >> individual MC packages, not a more system level diff to see exactly what >> changed from one build to another. > > We should have a global Pharo MC Package which has all other packages as > dependencies, that would simplify quite a few things here... > > Otherwise you're stuck with downloading the image for a specific version > which might be corrupted, and thus leaves you without any chance of > inspecting the changes... > > >> I was starting to dig into the metacello bootstrapping problem in 2.0 and >> became clear to me it'd help me a lot if I could more easily follow the >> diffs. >> >> Thanks, >> Patrik > >

July 26, 2012

Re: [Pharo-project] assertion failed using NBFFI call

by Lawson English

Doh, as you point out, I renamed the parameter... Thanks. L On 7/26/12 9:47 AM, Igor Stasenko wrote: > On 26 July 2012 18:22, Lawson English <lenglish5(a)cox.net> wrote: >> >> I'm getting an assertion failed error in NBFFICallout >> >> The debugger is point to this line: >> >> self assert: type loader notNil. >> <==========AssertionFailure:AssertionFailed >> ^ type >> > check that argument name is one of the method's arguments or instance variables. > this error happens when NB cannot associate an argument name in > function signature with any variable in method's scope (or it's class > scope) > >> when I run this code: >> >> >> >> +++++++++++++++ >> myBytes := ByteArray new: 20. >> (GMPfrWrapper new) test4: myBytes withLength: 20. >> >> >> ++++++++++++++++ >> GMPfrWrapper>> test4:bytes withLength: length >> >> >> <primitive: #primitiveNativeCall module: #NativeBoostPlugin> >> >> ^ NBFFICallout cdecl: #( int _ByteTest(ulong length, NBByteArrayPtr >> byteArray) ) >> module: >> '/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib' >> ++++++++++++++++++ >> >> I'ved checked with nm >> "/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib" >> >> and get: >> >> 00000e30 T _ByteTest >> ... >> >> >> I can load the file by doing: >> >> string := FileStream >> readOnlyFileNamed:'/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib' >> do: [:stream| >> stream binary. >> stream contents]. >> >> >> >> >> Suggestions? >> >> >> Thanks, >> >> >> L >> >> >> >> >> >> -- >> Squeak from the very start (introduction to Squeak and Pharo Smalltalk for >> the (almost) complete and compleate beginner). >> https://www.youtube.com/playlist?list=PL6601A198DF14788D&feature=view_all >> > > -- Squeak from the very start (introduction to Squeak and Pharo Smalltalk for the (almost) complete and compleate beginner). https://www.youtube.com/playlist?list=PL6601A198DF14788D&feature=view_all

July 26, 2012

[Pharo-project] assertion failed using NBFFI call

by Lawson English

I'm getting an assertion failed error in NBFFICallout The debugger is point to this line: self assert: type loader notNil. <==========AssertionFailure:AssertionFailed ^ type when I run this code: +++++++++++++++ myBytes := ByteArray new: 20. (GMPfrWrapper new) test4: myBytes withLength: 20. ++++++++++++++++ GMPfrWrapper>> test4:bytes withLength: length <primitive: #primitiveNativeCall module: #NativeBoostPlugin> ^ NBFFICallout cdecl: #( int _ByteTest(ulong length, NBByteArrayPtr byteArray) ) module: '/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib' ++++++++++++++++++ I'ved checked with nm "/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib" and get: 00000e30 T _ByteTest ... I can load the file by doing: string := FileStream readOnlyFileNamed:'/Users/lawsonenglish/Library/Developer/Xcode/DerivedData/Mandelbrot_Set-budpkxrcvoxdiocqojopadbgogzi/Build/Products/Debug/libMandelDraw.dylib' do: [:stream| stream binary. stream contents]. Suggestions? Thanks, L -- Squeak from the very start (introduction to Squeak and Pharo Smalltalk for the (almost) complete and compleate beginner). https://www.youtube.com/playlist?list=PL6601A198DF14788D&feature=view_all

July 26, 2012

Re: [Pharo-project] ODBC on W7 and Pharo 1.4

by Friedrich Dominicus

"Torsten Bergmann" <astares(a)gmx.de> writes: >>So to me it seems this is a problem in the ODBC Driver vor Access, but I >>could not imagine that no-one ever have treated Doubles while trying to >>access MS Access with Pharo. > > Maybe you should try the newest runtime for Access 2010: I don't think you can expect that an Access runtime 2010 can handle an access 2003 application. That's not very likely. > > to see if it is really a ODBC driver problem. > > According to [1] and [2] this should be 15 instead of 53 > for SQL_DOUBLE. Correct? Yes. I tested this result also with an C programm and yes indeed it is 53 and this is wront. It probably is the number of bit sin the mantissa of a Double. A work around may be not to use this query of the column but "hack" it that it just picks 15 bytes. I guess it tries to load all the 53 bits and well, who knows what bits pattern there may be around a Double... Hardly anything meaningful.

July 26, 2012

[Pharo-project] ODBC on W7 and Pharo 1.4

by Torsten Bergmann

>So to me it seems this is a problem in the ODBC Driver vor Access, but I >could not imagine that no-one ever have treated Doubles while trying to >access MS Access with Pharo. Maybe you should try the newest runtime for Access 2010: http://www.microsoft.com/en-us/download/details.aspx?id=10910 (for instance in a www.virtualbox.org installation). to see if it is really a ODBC driver problem. According to [1] and [2] this should be 15 instead of 53 for SQL_DOUBLE. Correct? Bye T. [1] http://msdn.microsoft.com/en-us/library/windows/desktop/ms716289(v=vs.85).a… [2] http://msdn.microsoft.com/en-us/library/windows/desktop/ms711786(v=vs.85).a…

July 26, 2012