Squeak-dev September 2011

squeak-dev@lists.squeakfoundation.org

64 participants
208 discussions

new Cog VMs available...
by Eliot Miranda 03 Oct '11

03 Oct '11

...at http://www.mirandabanda.org/files/Cog/VM/VM.r2496/. CogVM binaries as per VMMaker.oscog-eem.128/r2496 Fix regression in object-as-method/cannot-interpret for single and polymorphic inline cache misses (lookup:for:methodAndErrorSelectorInto:). Fix formatting bugette in context printing. This fixes a regression in objects-as-methods which Nicholas was suffering from as a crash during Run Coverage in the Test Runner. -- best, Eliot

4 5

Circular definitions of size and do:
by Rodney Polkinghorne 03 Oct '11

03 Oct '11

Dear list I wanted an array that lazily initialized elements 1 to n when you asked for the nth one. So I ran something like: SequenceableCollection variableSubclass #Foo ... When I started a workspace, and tried Foo new: 10, my image hung. Eventually, I found the problem. The workspace tried to print the answered instance, which resulted in it being sent a do: message to print its elements. This ran the method SequenceableCollection>>do: aBlock 1 to: self size do: [:index | aBlock value: (self at: index)] which ran Collection>>size | tally | tally := 0. self do: [:each | tally := tally + 1]. ^ tally which sent do:, which sent size, which ... Is this considered a bug, or a learning experience for new players? ArrayedCollection avoids it by restoring the primitive method for size. Rodney

4 7

Now available: a database conectivity API and its MySQL implementation
by Anindya Haldar 01 Oct '11

01 Oct '11

I am extremely pleased to offer to the community a generic Smalltalk API for database connectivity API, along with a MySQL implementation written from ground up. Think of JDBC and DBI like interfaces in other languages and environments, and this one offers abstract facilities like connection, prepared statement, resultset, row, metdata and so on for the application. The API implementation uses driver delegates, and insulates the applications from the driver specific interfaces. As of now, a MySQL implementation is available, with other databases engine supports in the future plans. The API supports server side prepared statements and cursor based fetch implemented in the MySQL driver. The implementation is currently available in 3 parts: 1) StdbCore project: This is primarily the API definitions and the delegation logic URL: http://www.squeaksource.com/StdbCore.html 2) StdbMysqlImpl project: This is the MySQL implementation of the StdbCore API URL: http://www.squeaksource.com/StdbMysqlImpl.html 3) StdbMysqlProtocol project: This is the underlying MySQL protocol driver, and used by the MySQL API implementation URL: http://www.squeaksource.com/StdbMysqlProtocol.html There is a package load ordering dependency (documented in the MySQL implementation project). Please load the core and the protocol driver packages first, before you load the implementation package. Combined, the packages has around 200 tests (most of them in the protocol driver and the API implementation) and are running green on Pharo 1.1.1 one click, Squeak 4.2 one click, Squeak 3.10 (with Seaside 2-8-573 dev image) under Mac OS X 10.5.8 with MySQL server 5.0.67. This was in the making for some time; and I feel it is time to share it with the community. I will continue to post more documentation and code snippets in the project page Wiki sections. For now, looking up the tests may be your best guide. Happy programming! Best, Anindya Haldar

5 12

Re: [squeak-dev] The Inbox: Collections-nice.423.mcz
by Eliot Miranda 01 Oct '11

01 Oct '11

I disagree. size implemented in terms of do: is quite appropriate. The subclass responsibility is do:. If this doesn't suit SequenceableCollection then size should be implemented as a subclass responsibility in SequenceableCollection, /not/ in Collection. On Fri, Sep 30, 2011 at 1:12 PM, <commits(a)source.squeak.org> wrote: > A new version of Collections was added to project The Inbox: > http://source.squeak.org/inbox/Collections-nice.423.mcz > > ==================== Summary ==================== > > Name: Collections-nice.423 > Author: nice > Time: 30 September 2011, 10:12:20.333 pm > UUID: 7613573f-6d08-9d4c-9d9e-2eaacff799fe > Ancestors: Collections-ul.422 > > #size and #do: default implementation will cause an infinite loop in every > new subclass of SequenceableCollection without any form of warning to the > poor programmer. > > It's simple to avoid this bad behavior by: > - moving Collection>>size implementation to the sole place where it is used > (LinkedList) > - replacing it with a subclassResponsibility. > > =============== Diff against Collections-ul.422 =============== > > Item was changed: > ----- Method: Collection>>size (in category 'accessing') ----- > size > "Answer how many elements the receiver contains." > > + ^self subclassResponsibility! > - | tally | > - tally := 0. > - self do: [:each | tally := tally + 1]. > - ^ tally! > > Item was added: > + ----- Method: LinkedList>>size (in category 'accessing') ----- > + size > + "Answer how many elements the receiver contains." > + > + | tally | > + tally := 0. > + self do: [:each | tally := tally + 1]. > + ^ tally! > > > -- best, Eliot

1 0

Re: [Pharo-project] [squeak-dev] Started to post Videos of "Deep into Smalltalk"
by Marcus Denker 01 Oct '11

01 Oct '11

On Sep 30, 2011, at 9:52 AM, Alexander Lazarević wrote: > 2011/9/30 Marcus Denker <marcus.denker(a)inria.fr>: >> and the classic talk of Doru from ESUG 2008 showing how (not) to give ESUG talks. > > Saw it for the first time! Nice :) > > Even if this is most likely a very well known service, it can't hurt > to mention http://www.slideshare.net as a source of inspiration. > Yes! We put all ESUG slides there, too: http://www.slideshare.net/esug/ (335 Talks!) So the slides of the presentation are there, too: http://www.slideshare.net/esug/1-presentingatesug -- Marcus Denker -- http://marcusdenker.de

3 2

The Inbox: Collections-nice.423.mcz
by commits＠source.squeak.org 30 Sep '11

30 Sep '11

A new version of Collections was added to project The Inbox: http://source.squeak.org/inbox/Collections-nice.423.mcz ==================== Summary ==================== Name: Collections-nice.423 Author: nice Time: 30 September 2011, 10:12:20.333 pm UUID: 7613573f-6d08-9d4c-9d9e-2eaacff799fe Ancestors: Collections-ul.422 #size and #do: default implementation will cause an infinite loop in every new subclass of SequenceableCollection without any form of warning to the poor programmer. It's simple to avoid this bad behavior by: - moving Collection>>size implementation to the sole place where it is used (LinkedList) - replacing it with a subclassResponsibility. =============== Diff against Collections-ul.422 =============== Item was changed: ----- Method: Collection>>size (in category 'accessing') ----- size "Answer how many elements the receiver contains." + ^self subclassResponsibility! - | tally | - tally := 0. - self do: [:each | tally := tally + 1]. - ^ tally! Item was added: + ----- Method: LinkedList>>size (in category 'accessing') ----- + size + "Answer how many elements the receiver contains." + + | tally | + tally := 0. + self do: [:each | tally := tally + 1]. + ^ tally!

1 0

Started to post Videos of "Deep into Smalltalk"
by Marcus Denker 30 Sep '11

30 Sep '11

Hello, After a long wait... finally: http://www.youtube.com/user/esugboard#grid/user/843D1D545F9F52B6 This now has: - Extensive, great two part PetitParser Tutorial by Doru (3 hours) http://www.youtube.com/watch?v=vgiJG3Vd6KU If you ever wanted to write a Parser, this will give you a very understandable, step-by-step introduction. - Enhancing SqueakNOS. http://www.youtube.com/watch?v=u9RUmIY2Sz8 Crazy, these Argentinians... - NativeBoost: http://www.youtube.com/watch?v=Bd6mz6F167s Generate Binary code from the image. And a *very* fast FFI as an example. Guido and Javier of the SqueakNOS video above are right now busy here in Lille building something *very* interesting with NativeBoost... - FFI, an introduction by Igor: http://www.youtube.com/watch?v=fU8XRCO7MUU What is FFI and how do we use it? - and the first part of "A little Journey in a Virtual Machine" http://www.youtube.com/watch?v=FJQmjhOvjUs This has 5 (!) parts and together with the FFI presentation, this allows you to experience the very first lecture ever given by Igor Stasenko. More to come... but this already might keep interested people busy for some enternaining evenings :-) We will post all the rest step by step. The ESUG Youtube account has more stuff posted already, e.g. the talks of this years IWST Workshop and some talks of ESUG2011, and the classic talk of Doru from ESUG 2008 showing how (not) to give ESUG talks. All here: http://www.youtube.com/user/esugboard Marcus -- Marcus Denker -- http://marcusdenker.de

3 2

Daily Commit Log
by commits＠source.squeak.org 30 Sep '11

30 Sep '11

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2011-September/005006.… Name: Morphic-bf.581 Ancestors: Morphic-dtl.580 Fix keyDown/keyUp events. Must not be interpreted. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-September/005007.… Name: Collections-eem.460 Ancestors: Collections-ul.459 Fix String>numArgs for prefAllowUnderscoreSelectors regime. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-September/005008.… Name: Kernel-eem.632 Ancestors: Kernel-ul.631 Avoid work in common case of setting method's source pointer to zero when already zero. =============================================

1 0

Oddities with #keyDown and Sensor
by Lawson English 30 Sep '11

30 Sep '11

I've been playing with creating a simple 2D asteroids type game in Squeak, and when it came time to try to monitor how long the accelerator key for the rocket (up arrow in this case) had been pressed, I ran into all sorts of oddities. On an intel Mac OS X 10.6.8 box, running various Cog VMs, I found that the event passed to #keyDown: contained zero as the value, regardless of which key was pressed. On Linux, this problem seems not to exist. Again on the Mac, this test class http://paste.lisp.org/display/124816 doesn't seem to work. I was trying to grab the raw Sensor event buffer, but I can't seem to get more than a handful of events. #step is where the meat of the code is. It tries to grab events from Sensor and if they are keyboard events, put them into a buffer for examination. You evoke the test morph thusly: myMorph := MyMorph new. myMorph openInWorld. Transcript show: myMorph aCounter. Transcript show: myMorph eventBuffer. Obviously, I'm doing something fundamentally wrong (or the MacOS X bug for #keyDown manifests directly with Sensor). Help? THanks Lawson.

3 10

The Trunk: Kernel-eem.632.mcz
by commits＠source.squeak.org 29 Sep '11

29 Sep '11

Eliot Miranda uploaded a new version of Kernel to project The Trunk: http://source.squeak.org/trunk/Kernel-eem.632.mcz ==================== Summary ==================== Name: Kernel-eem.632 Author: eem Time: 29 September 2011, 2:09:31.751 pm UUID: 807c955e-061f-4006-b9c2-981ffd241b4c Ancestors: Kernel-ul.631 Avoid work in common case of setting method's source pointer to zero when already zero. =============== Diff against Kernel-ul.631 =============== Item was changed: ----- Method: CompiledMethod>>setSourcePointer: (in category 'source code management') ----- setSourcePointer: srcPointer + "We can't change the trailer of existing method, since it could have a + completely different format. Therefore we need to generate a copy + with new trailer, containing a srcPointer, and then #become it." - "We can't change the trailer of existing method, since - it could have completely different format. Therefore we need to - generate a copy with new trailer, containing an scrPointer, and then - #become it" | trailer copy start | + trailer := srcPointer = 0 + ifTrue: "catch the common case of setting the source pointer to 0 when already 0" + [self sourcePointer = 0 ifTrue: + [^self]. + CompiledMethodTrailer empty] + ifFalse: + [CompiledMethodTrailer new sourcePointer: srcPointer]. - trailer := CompiledMethodTrailer new sourcePointer: srcPointer. copy := self copyWithTrailerBytes: trailer. "ar 3/31/2010: Be a bit more clever since #become: is slow. If the old and the new trailer have the same size, just replace it." + (self trailer class == trailer class and:[self size = copy size]) + ifTrue: + [start := self endPC + 1. + self replaceFrom: start to: self size with: copy startingAt: start] + ifFalse: + [self becomeForward: copy]. - (self trailer class == trailer class and:[self size = copy size]) ifTrue:[ - start := self endPC + 1. - self replaceFrom: start to: self size with: copy startingAt: start. - ] ifFalse:[self becomeForward: copy]. ^self "will be copy if #become was needed" !

1 0

Jump to page: