- Pharo-dev - lists.pharo.org

Deprecations on Smalltalk Image in Pharo 5
by Henrik Johansen Jan. 4, 2017

Jan. 4, 2017

It was my impression there was a procedure in place earlier for moving methods from SmalltalkImage to classes with a more singular responsibility that ensures a smooth upgrade experience; 1) create accessor method on Smalltalk (if none of the existing are appropriate) 2) implement the old methods on new class in backwards compatability category 3) Deprecate method on SmalltalkImage, on the form: doX: someThing self deprecated: 'Use Smalltalk specializedArea x: someThing instead' on: 'foo' in: 'bar'. ^self specializedArea x: someThing (Ref: #os, #vm) I've run into a few cases where this is not the case when moving to Pharo 5, should they be changed, or am I wrong, and the procedure above is no longer considered valuable? Case 1: SmalltalkImage removeFromShutDownList: aClass (and removeFromStartUpList:) self deprecated: 'Please use registration methods provided in SessionManager / registration category.', String cr, 'ex: SessionManager default unregisterClassNamed: self name'. There's a new #session accessor on SmalltalkImage, but... 1) SessionManager does not implement a backwards compatible protocol. 2) The deprecated message above is wrong, the Smalltalk session unregisterClassNamed: is preferrable to referencing the class directly. 3) The deprecated method does nothing, rather than redirect to the new place. Silent errors when deprecation warnings are turned of, yay. 4) The deprecation message uses the deprecated form which does not include the parameters needed to know when to expect it removed forever. Also, the corresponding addTo... methods have not been deprecated, but redirect to SessionManager directly rather than through #session... Case 2: EndianDetector: Used to be Smalltalk #endianness 1) The method on SmalltalkImage has simply been removed, no deprecation. 2) Accessor remains the same, but all references in the base image has been rewritten to use EndianDetector directly... 2) Would not Smalltalk platform be a reasonable place to put this responsiblity, letting it use an EndianDetector behind the scenes? Both the preexisting #endianness, as well as the new #isLittleEndian and #isBigEndian? Cheers, Henry

3 4

Redline: Talking Runtime basics ...
by James Ladd Jan. 3, 2017

Jan. 3, 2017

Hi Pharo People, I have continued work on Redline Smalltalk and I'm wanting to discuss what the absolute minimum set of Classes and method should be included in the Redline Runtime. Would anyone here like to participate in that discussion? - James. Redline Smalltalk <http://redline.st> -- View this message in context: http://forum.world.st/Redline-Talking-Runtime-basics-tp4928375.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

11 25

it's too quiet ... please ignore (except for stef:)
by Dale Henrichs Jan. 2, 2017

Jan. 2, 2017

No messages on list since yesterday morning? Dale

8 12

:: Separator in class names
by Torsten Bergmann Jan. 2, 2017

Jan. 2, 2017

Hi, due to the absence of a namespace solution we today often prefix classes with one, two or three letters. You all know prefixes like "Zn", "Zdc", "GLM", "Tx", "WA", ... to avoid conflicting names. I would like to have a "::" separator within the class name to better distinguish the class name from the prefix and increase readability at the same time. SmalltalkMT had this solution in alignment with C++ and I really liked working with it. So much that I would like to see it in Pharo too - maybe also as a base for future Namespace additions. So I modified the #isValidGlobalName in Pharo 6.0 like mentioned below to allow and support the :: notation for classes. After that I can create classes like: Seaside::Component Core::Boolean Model::Person GLM::BrickListModel Tx::FontAttribute so far without any problem. Do you think - there will be places where this conflicts or creates hazzles (tools, metacello, ...) - if it would be a good idea to add this to the default image? Thanks for feedback. Thx T. -------------------------------------------------------------------------------------------------- isValidGlobalName self ifEmpty: [ ^ false ]. "reserverd default names" self = 'NameOfSubclass' ifTrue: [ ^ false ]. self = 'TNameOfTrait' ifTrue: [ ^ false ]. (self occurrencesOf: $:) = 2 ifTrue: [ ^(self splitOn: '::') allSatisfy: [: part | part isValidGlobalName ] ]. ^ Character supportsNonASCII ifTrue: [ (self first isLetter and: [self first isUppercase]) and: [ self allSatisfy: [:character | character isAlphaNumeric or: [ character = $_ ]]]] ifFalse: [ (self first between: $A and: $Z) and: [ self allSatisfy: [:character | (character between: $a and: $z) or: [ (character between: $A and: $Z) or: [ (character between: $0 and: $9) or: [ character = $_]]]]]]

8 11

uFFI chapter needs more love
by Stephane Ducasse Jan. 1, 2017

Jan. 1, 2017

Hi FFI is important and the FFI chapter is still draft. It would be good that some good souls help. https://github.com/SquareBracketAssociates/PharoInProgress/tree/master/Unif… Stef

5 7

IDNA / punycode for Zinc?
by Max Leske Dec. 28, 2016

Dec. 28, 2016

Hi (Sven), Zinc canât currently handle unicode domain names (e.g. http://Ã¼ni.ch <http://xn--ni-wka.ch/>). Are there any plans to implement punycode / IDNA conversion for Zinc? Or is there an explicit reason not to support it? I see that #parseHostPort: expects the host portion to be percent escaped, what is the use case for this? I have never seen a percent escaped host portion. Usually the host portion is either pure ASCII, unicode or punycode (in my experience at least). Just curious, as I just added IDNA conversion to one of our applications (I just let python perform the conversion: https://docs.python.org/2/library/codecs.html#module-encodings.idna <https://docs.python.org/2/library/codecs.html#module-encodings.idna>). Cheers, Max

4 9

[ANN] Tarantalk - Smalltalk + Lua + Tuple DB
by Masashi UMEZAWA Dec. 28, 2016

Dec. 28, 2016

Hi all, I've developed a new database connectivity library for Pharo, called Tarantalk. https://github.com/mumez/Tarantalk Tarantalk is a connector for Tarantool. https://tarantool.org/ Tarantool is a Lua engine with tuple persistence. It is like a GemStone in Lua. Firstly, Tarantalk can be used as a tuple-centric DBMS. You can store arbitrary Smalltalk objects as tuples, and retrieve them by multiple indexes. tarantalk := TrTarantalk connect: 'taran:talk@localhost:3301'. space := tarantalk ensureSpaceNamed: 'bookmarks'. primaryIndex := space ensurePrimaryIndexSetting: [ :options | options tree; partsTypes: #(#string)]. space insert: #('Tarantool' 'https://tarantool.org' 'Tarantool main site'). space insert: #('Pharo' 'http://pharo.org' 'Pharo main site'). primaryIndex selectHaving: #('Pharo'). " => #(#('Pharo' 'http://pharo.org' 'Pharo main site'))" What is more interesting is that Tarantalk can call remote Lua functions and can evaluate Lua expressions. So you can extend your database system freely with Lua and Smalltalk! -- Lua function bookmarkUrls() local urls = {} for k,v in box.space.bookmarks:pairs() do table.insert(urls, v[2]) end return urls end box.schema.func.create('bookmarkUrls') "Smalltalk" tarantalk call: 'bookmarkUrls'. " => #(#('http://pharo.org' 'https://tarantool.org'))" (tarantalk evalWithReturn: '(require("digest").sha512_hex(...))' arguments: {'Hello, Smalltalk'}) value. " => #('ec599e128831b282f6f7a833834c90a3eb2e61453e5757ca3c2bc8a26e94d7c2f76bd6a7ce33df2427f3821e44a12d26781d39eac6782b59a649950ea59f9e13')" Moreover, Tarantalk is pretty fast compared to some other NoSQL databases. https://github.com/mumez/Tarantalk#performance A simple micro benchmark shows that Tarantalk is about 5x faster than RediStick Redis client. Enjoy! -- [:masashi | ^umezawa]

2 1

Minheadless VM flavour status update
by Ronie Salgado Dec. 28, 2016

Dec. 28, 2016

Hi all, I am happy to announce that my latest VM refactoring experiment is almost complete. I am only missing adding traditional style Makefiles to make Eliot happy ;), and making this work with MSVC. The minheadless is a VM variant that unifies a lot of the code for Windows, Linux and OS X. To the point that I am using the same CMake scripts for building on the three platforms. The minheadless VM is a completely headless by default VM, unless the configure script finds SDL2. If the configure script finds SDL2, and the legacy display interface is enabled(which is by default), then this VM will be built using SDL2 to implement the normal diplay primitives. The SDL2 backend work with normal Pharo and Squeak image. If SDL2 is not found by the configure script, or the legacy display primitives are disabled in the configure script, or the VM is started in headless mode, then the traditional display primitives are just going to be a /dev/null. Currently I am commiting this VM variant to the MinimalisticHeadless branch at: https://github.com/ronsaldo/opensmalltalk-vm/tree/MinimalisticHeadless For building using CMake it can be done in two ways: ------------------------------------ Calling CMake directly mkdir build cmake .. -DCMAKE_BUILD_TYPE=<The build type> (or cmake-gui ..) make The supported build types are: - Debug - Assert - RelWithDebInfo - Release If you are on an x86_64 system and you want to build 32 bits version, you have to pass the -DBUILD_I386_VERSION=True option to cmake. mkdir build cmake .. (or cmake-gui ..) make ------------------------------------ Using mvm style scripts Easy mode: cd build.minheadless.cmake/<x86 or x64>/vm_variant ./mvm A bit more complicated/flexible mode: This mvm will configure and build the vm for the four build types. This behavior was present on the mvm for Windows, so I am replicating it. In these same folders that are two other scripts: mvm_configure (calls cmake to only configure the build types), and mvm_configure_variant mvm_configure_variant is used to actually call cmake and configure the VM for building. The syntax of mvm_configure_variant is the following: mvm_configure_variant <BuildDirectory> <BuildType> Currently my workflow for using these scripts is the following with MSys/Cygwin/Linux: ./mvm_configure make -j8 The makefile at build.minheadless.cmake/<x86 or x64>/vm_variant is only used for delegating to the makefiles created by CMake. ------------------------------------ I am leaving some prebuilt Pharo VMs with the SDL2 backend for Windows for experimenting at: http://ronie.cl/files/vm/ Using OSWindow with this VM requires loading the latest version of OSWindow by reloading the ConfigurationOfOSWindow manually in Monticello, and then doing ConfigurationOfOSWindow load. With this VM, there are finally no more duplicated events. My TODO list for this includes adding plain Makefile. BTW: during the following week I will continue working with Lowcode. I really need it working again. Happy longest/shortest day of the year Ronie

5 5

Q: any work around ROS has been done on Pharo?
by Igor Stasenko Dec. 27, 2016

Dec. 27, 2016

Hi, i wonder, are there any projects to run/connect Pharo with ROS(robot operating system) i'm interested in any bits, starting from basic ones, since i'm gonna to work with it in closest future, so would be nice to use Pharo & stand on a shoulders of others, of course, to learn it fester and understand it better. -- Best regards, Igor Stasenko.

9 13

How stable 64-bit VM?
by Igor Stasenko Dec. 25, 2016

Dec. 25, 2016

Hi, sorry for my ignorance, i downloaded fresh build of 64-bit VM and has strange hiccups (crash on downloading/installing a package)... is it stable, or it is just me, too ignorant & stupid? -- Best regards, Igor Stasenko.

2 2