Squeak-dev March 2012

squeak-dev@lists.squeakfoundation.org

63 participants
208 discussions

The Inbox: Monticello-nice.498.mcz
by commits＠source.squeak.org 09 Mar '12

09 Mar '12

A new version of Monticello was added to project The Inbox: http://source.squeak.org/inbox/Monticello-nice.498.mcz ==================== Summary ==================== Name: Monticello-nice.498 Author: nice Time: 9 March 2012, 11:17:06.384 pm UUID: 395becce-f60f-4242-a937-d1fc6ab4f462 Ancestors: Monticello-bf.496 Add a Monticello preference in order to #ignoreSourceCodeFormattingDifferences in method definitions. This will affect the diff/merge behavior of Monticello. When ignoring the differences, the diff is performed on the scanned tokens and _ := also compare equal. It's generally a bad idea to ignore those diffs, but does really help when browsing diffs with distant cousins like Etoys or Pharo. This preference must be considered as a quick hack. A better idea would be to implement such filter in the Monticello UI, but this ain't gonna be as easy. Bis and don't forget hash this time... =============== Diff against Monticello-bf.496 =============== Item was removed: - SystemOrganization addCategory: #Monticello! - SystemOrganization addCategory: #'Monticello-Base'! - SystemOrganization addCategory: #'Monticello-Chunk Format'! - SystemOrganization addCategory: #'Monticello-Loading'! - SystemOrganization addCategory: #'Monticello-Merging'! - SystemOrganization addCategory: #'Monticello-Modeling'! - SystemOrganization addCategory: #'Monticello-Patching'! - SystemOrganization addCategory: #'Monticello-Repositories'! - SystemOrganization addCategory: #'Monticello-Storing'! - SystemOrganization addCategory: #'Monticello-UI'! - SystemOrganization addCategory: #'Monticello-Versioning'! - SystemOrganization addCategory: #'Monticello-Mocks'! - SystemOrganization addCategory: #'Monticello-Utilities'! Item was added: + SystemOrganization addCategory: #'Monticello-Base'! + SystemOrganization addCategory: #'Monticello-Chunk Format'! + SystemOrganization addCategory: #'Monticello-Loading'! + SystemOrganization addCategory: #'Monticello-Merging'! + SystemOrganization addCategory: #'Monticello-Modeling'! + SystemOrganization addCategory: #'Monticello-Patching'! + SystemOrganization addCategory: #'Monticello-Repositories'! + SystemOrganization addCategory: #'Monticello-Storing'! + SystemOrganization addCategory: #'Monticello-UI'! + SystemOrganization addCategory: #'Monticello-Versioning'! Item was changed: MCDefinition subclass: #MCMethodDefinition + instanceVariableNames: 'classIsMeta source category selector className timeStamp tokens' + classVariableNames: 'IgnoreSourceCodeFormattingDifferences' - instanceVariableNames: 'classIsMeta source category selector className timeStamp' - classVariableNames: '' poolDictionaries: '' category: 'Monticello-Modeling'! MCMethodDefinition class instanceVariableNames: 'definitions'! MCMethodDefinition class instanceVariableNames: 'definitions'! Item was added: + ----- Method: MCMethodDefinition class>>ignoreSourceCodeFormattingDifferences (in category 'preferences') ----- + ignoreSourceCodeFormattingDifferences + "Accessor for the system-wide preference" + + <preference: 'Ignore source code formatting differences' + category: 'Monticello' + description: 'If enabled, the diff and merge tools will ignore formatting differences.' + type: #Boolean> + ^IgnoreSourceCodeFormattingDifferences ifNil: [ false ]! Item was added: + ----- Method: MCMethodDefinition class>>ignoreSourceCodeFormattingDifferences: (in category 'preferences') ----- + ignoreSourceCodeFormattingDifferences: aBoolean + "Accessor for the system-wide preference" + + IgnoreSourceCodeFormattingDifferences := aBoolean! Item was changed: ----- Method: MCMethodDefinition>>= (in category 'comparing') ----- + = t1 + self class ignoreSourceCodeFormattingDifferences + ifTrue: [^ super = t1 + and: [t1 tokens = self tokens]]. + ^ super = t1 + and: [t1 source = self source + and: [t1 category = self category + and: [t1 timeStamp = self timeStamp]]]! - = aDefinition - ^(super = aDefinition) - and: [aDefinition source = self source - and: [aDefinition category = self category - and: [aDefinition timeStamp = self timeStamp]]]! Item was changed: ----- Method: MCMethodDefinition>>hash (in category 'comparing') ----- hash + | t1 | + self class ignoreSourceCodeFormattingDifferences + ifTrue: [t1 := classIsMeta asString hashWithInitialHash: self tokens hash. + t1 := className hashWithInitialHash: t1. + ^ t1]. + t1 := classIsMeta asString hashWithInitialHash: 0. + t1 := source hashWithInitialHash: t1. + t1 := category hashWithInitialHash: t1. + t1 := className hashWithInitialHash: t1. + ^ t1! - | hash | - hash := classIsMeta asString hashWithInitialHash: 0. - hash := source hashWithInitialHash: hash. - hash := category hashWithInitialHash: hash. - hash := className hashWithInitialHash: hash. - ^ hash! Item was changed: ----- Method: MCMethodDefinition>>initializeWithClassName:classIsMeta:selector:category:timeStamp:source: (in category 'serializing') ----- initializeWithClassName: classString classIsMeta: metaBoolean selector: selectorString category: catString timeStamp: timeString source: sourceString className := classString asSymbol. selector := selectorString asSymbol. category := catString asSymbol. timeStamp := timeString. classIsMeta := metaBoolean. source := sourceString withSqueakLineEndings. + tokens := nil ! Item was added: + ----- Method: MCMethodDefinition>>tokens (in category 'accessing') ----- + tokens + ^tokens ifNil: [tokens := (Scanner new scanTokens: self source) replaceAll: #'_' with: #':=']!

1 0

The Inbox: Monticello-nice.497.mcz
by commits＠source.squeak.org 09 Mar '12

09 Mar '12

A new version of Monticello was added to project The Inbox: http://source.squeak.org/inbox/Monticello-nice.497.mcz ==================== Summary ==================== Name: Monticello-nice.497 Author: nice Time: 9 March 2012, 10:42:27.071 pm UUID: 8537ad03-c3e3-4bf6-b318-896c3efee5f7 Ancestors: Monticello-bf.496 Add a Monticello preference in order to #ignoreSourceCodeFormattingDifferences in method definitions. This will affect the diff/merge behavior of Monticello. When ignoring the differences, the diff is performed on the scanned tokens and _ := also compare equal. It's generally a bad idea to ignore those diffs, but does really help when browsing diffs with distant cousins like Etoys or Pharo. This preference must be considered as a quick hack. A better idea would be to implement such filter in the Monticello UI, but this ain't gonna be as easy. =============== Diff against Monticello-bf.496 =============== Item was removed: - SystemOrganization addCategory: #Monticello! - SystemOrganization addCategory: #'Monticello-Base'! - SystemOrganization addCategory: #'Monticello-Chunk Format'! - SystemOrganization addCategory: #'Monticello-Loading'! - SystemOrganization addCategory: #'Monticello-Merging'! - SystemOrganization addCategory: #'Monticello-Modeling'! - SystemOrganization addCategory: #'Monticello-Patching'! - SystemOrganization addCategory: #'Monticello-Repositories'! - SystemOrganization addCategory: #'Monticello-Storing'! - SystemOrganization addCategory: #'Monticello-UI'! - SystemOrganization addCategory: #'Monticello-Versioning'! - SystemOrganization addCategory: #'Monticello-Mocks'! - SystemOrganization addCategory: #'Monticello-Utilities'! Item was added: + SystemOrganization addCategory: #'Monticello-Base'! + SystemOrganization addCategory: #'Monticello-Chunk Format'! + SystemOrganization addCategory: #'Monticello-Loading'! + SystemOrganization addCategory: #'Monticello-Merging'! + SystemOrganization addCategory: #'Monticello-Modeling'! + SystemOrganization addCategory: #'Monticello-Patching'! + SystemOrganization addCategory: #'Monticello-Repositories'! + SystemOrganization addCategory: #'Monticello-Storing'! + SystemOrganization addCategory: #'Monticello-UI'! + SystemOrganization addCategory: #'Monticello-Versioning'! Item was changed: MCDefinition subclass: #MCMethodDefinition instanceVariableNames: 'classIsMeta source category selector className timeStamp' + classVariableNames: 'IgnoreSourceCodeFormattingDifferences' - classVariableNames: '' poolDictionaries: '' category: 'Monticello-Modeling'! MCMethodDefinition class instanceVariableNames: 'definitions'! MCMethodDefinition class instanceVariableNames: 'definitions'! Item was added: + ----- Method: MCMethodDefinition class>>ignoreSourceCodeFormattingDifferences (in category 'preferences') ----- + ignoreSourceCodeFormattingDifferences + "Accessor for the system-wide preference" + + <preference: 'Ignore source code formatting differences' + category: 'Monticello' + description: 'If enabled, the diff and merge tools will ignore formatting differences.' + type: #Boolean> + ^IgnoreSourceCodeFormattingDifferences ifNil: [ false ]! Item was added: + ----- Method: MCMethodDefinition class>>ignoreSourceCodeFormattingDifferences: (in category 'preferences') ----- + ignoreSourceCodeFormattingDifferences: aBoolean + "Accessor for the system-wide preference" + + IgnoreSourceCodeFormattingDifferences := aBoolean! Item was changed: ----- Method: MCMethodDefinition>>= (in category 'comparing') ----- = aDefinition + IgnoreSourceCodeFormattingDifferences ifTrue: [ + ^(super = aDefinition) + and: [| t1 t2 | + t1 := Scanner new scanTokens: aDefinition source. + t2 := Scanner new scanTokens: self source. + t1 size = t2 size and: [(1 to: t1 size) allSatisfy: [:i | + (t1 at: i) = (t2 at: i) + or: [#( #( #'_' #':=') #( #':=' #'_')) includes: {t1 at: i. t2 at: i} ]]]]]. ^(super = aDefinition) and: [aDefinition source = self source and: [aDefinition category = self category and: [aDefinition timeStamp = self timeStamp]]]!

1 0

Testing Morphic Designer
by Hernán Morales Durand 09 Mar '12

09 Mar '12

Does someone would mind to share a Morphic Designer working image? [*] I've tried to install it in both Squeak 4.3 and Pharo 1.3 without success. Thanks in advance, Hernán [*] https://www.hpi.uni-potsdam.de/hirschfeld/trac/SqueakCommunityProjects/wiki…

2 2

SM, scripts, and application names (I forgot something)
by Chris Cunnington 09 Mar '12

09 Mar '12

Actually, I forgot another way. This is the install script for Magma: SMSqueakMap default loadUpdates. (Smalltalk hasClassNamed: #MaObjectRepository) ifFalse: [ SMSqueakMap default installPackageNamed: 'Magma' version: '1.3 server' ]. #('Magma-Tester-cmm.395' 'Magma-Squeak-Tester-cmm.2') do: [ : each | (Installer repository: 'http://www.squeaksource.com/Magma') install: each ] That is short, doesn't list all packages individually, and installs lots and lots of packages. It comes as a downloadable script from here: http://wiki.squeak.org/squeak/2657 Drop that on the World and step back, while it loads everything. This, also, could go in a script deployed by a button in SM. Chris

1 0

AW: [squeak-dev] SM, scripts, and application names
by Tim Felgentreff 09 Mar '12

09 Mar '12

I agree in all points with Frank. Metacello offers a system to do it, SM a system to run the DoIt. If we just standardize on Metacello ( at least for the default case) you don't even need to specify versions and install scripts in SM anymore. Just point to a ConfigurationOf package. SM can strip the prefix to get the name, parse the data to get the versions, generate the install script for each version. Couldn't be easier to add sth to the map.

1 0

Smalltalk Implementation of the Viable System Model
by Robert Carleton 09 Mar '12

09 Mar '12

I've been slowly piecing together an implementation of what is called the Viable System Model, or VSM. The model was developed by operations research theorist and cybernetician Stafford Beer in the early 70's. One of the best known projects to use the VSM was an attempted real-time computer-controlled planned economy in the years 1970 to 1973. It was called Project Cybersyn. The system used a telex network and applications running on an IBM 360 computer. I've been working on a Smalltalk implementation of the model, that I'm calling the Viable System Agent. It's come far enough along, that I think it is ready to show. There is a small web page for the project: http://home.rbcarleton.com/rbc/software/smalltalk/VSA/ There is a presentation, along with file outs of the present code. See the unit tests for how it might be used. I have to add that it is just a beginning. If you have previously seen Beers ideas in operations research and cybernetics, you might find it interesting. Best, --Bruce

2 4

SM, scripts, and application names
by Chris Cunnington 09 Mar '12

09 Mar '12

-inter-project dependencies -names -why Metacello configurations shouldn't be the underlying mechanism to put projects up on SM? Thank you for the feedback. I'd like to try something that may show my point of view. I may be misinformed on some things. Let's combine two ideas, one yours and one mine. We'll take two postulates and combine them. That should make it clearer in my mind and perhaps answer your question. Or at least you'll see where I'm coming from. Your postulate will be the one above. I'll reduce that down to "only use Metacello". That's my interpretation of the question above. My postulate will be "install with a button". To combine the two, then, is to ask "how do I install a Metacello configuration with a button"? I haven't used Metacello Browser, because I tried to load it into Pharo with a Metacello configuration and it wouldn't. The idea sounds great, and I'm not sure why Pharo doesn't use it. Who doesn't like buttons? Well, Pharo and Squeak have been doing without buttons just fine. The standard is to cut and paste a script. Maybe you get it from a mailing list. Maybe you get it from SqueakSource to download Moose with a Metacello configuration. It's a poor solution only supported by social proof and precedent. Installing with a button is better. Take a Metacello configuration like this: Installer ss project: 'MetacelloRepository'; install: 'FooFooFoo'; ((Smalltalk at: #FooFooFoo) project version: #stable) load. We have that in a Workspace. In SM we put that in a file called, say, Foo.st. We save that on the SM server. Then we have a URL that installs with a button. And if you don't like files, you can't really be for SM, because it is only a storehouse of URLs with files at the end (mcz, sar, st). So, from the point of view of things in a Workspace that become URL accessible files, Metacello is a subset of the tools we use to capture the context -- the file. You can't install Metacello without either Gofer or Installer. If we have a file at the end of a URL, then you can have Installer or Gofer install MC configuration files (like this one http://www.squeaksource.com/VMMaker/update-dtl.mcm), or package names, or Metacello Repository configurations. They can all go in a file. The only difference is whether the dependencies information is in the Metacello configuration or in the file itself like this: Installer ss project: 'KomHttpServer'; install: 'DynamicBindings'; install: 'KomServices'; install: 'KomHttpServer'. That is a load script with the dependencies specified. In the Metacello related script above the dependency resolution is delegated from the script to Metacello. So, I don't really understand your question, because you cannot make Metacello the underlying mechanism, because the underlying mechanism is taking something that is executed in a Workspace and making it a script, a file, that goes at the end of a URL. What you put into a Workspace is never confined to Metacello, and neither would any install need to be. Oh yea, one more thing. I was advocating application names yesterday. I don't see anything in Metacello that does what I'm talking about. It has names for configurations like ConfigurationOfFoo. I just want Foo. Give me Foo. If I'm SmalltalkDirectons and someone gives me a tip on a great application called Foo, then I want to look for Foo. I don't want a "ConfigurationOfFoo" and I don't want to guess that the package name MCFoo or FFFoo or QSQuasiFoo is the package that may be the root. I want to see a list. I want to see Foo. And I want to press a button. Thanks, Chris

2 1

AW: [squeak-dev] SM, scripts, and application names
by Tim Felgentreff 09 Mar '12

09 Mar '12

Regarding inter-project dependencies and names: Metacello offers that. Why not just use it? I don't understand why Metacello configurations shouldn't be the underlying mechanism to put projects up on SM. It works, it already has versions, dependencies and all that Jazz. SM as a better GUI to Metacello than the MetacelloBrowser to me feels like an ideal solution. -Tim

1 0

Daily Commit Log
by commits＠source.squeak.org 09 Mar '12

09 Mar '12

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2012-March/005210.html Name: Monticello-bf.496 Ancestors: Monticello-bf.495 - keep branch name when saving new version =============================================

1 0

SOB minutes - 3/6/12
by Chris Cunnington 08 Mar '12

08 Mar '12

- all members attending - Election procedure was discussed: * since the Squeak People website is defunct a page listing all the eligible voters needs to be created and posted on squeak.org * might be similar to Sugar Labs (another SFC member project) http://wiki.sugarlabs.org/go/Sugar_Labs/Members/List * a person would be eligible to vote by having shown they'd made a contribution * as there is a conflict of interest, not the board, but the election team should approve contributors * possibly this could be automated, e.g. by adding a Squeak People-like network of trust to Squeak Map * this discussion needs to be public on Squeak-Dev - A beta version of Etoys 5 has been released (http://squeakland.org/download <http://squeakland.org/download/>) - Colin discussed creating remote tools for work on a mini Squeak image: * While Smalltalk offers reflection, the ideal is to have a program as unaffected by its tools as a program is by a text editor * Colin uses a miniature image for the Atom feed of his website wiresong.ca. He built the image from Pavel's core using the Mason build tool. He is planning to post a link to Squeak-Dev to allow people to download this image - All members of the board are running for re-election with the exception of Jecel, who described his position at the last meeting - Armed with a FreeBSD vm, Randal is in the process of creating a booking engine application or Insight Cruises and offered to post a link to the website when it's done

9 13

← Newer
1
...
14
15
16
17
18
19
20
21
Older →

Jump to page: