Self-interest January 2014

self-interest@lists.selflanguage.org

13 participants
11 discussions

Unhooking Connectors
by sean＠clipperadams.com 01 Feb '14

01 Feb '14

When you pick up the = attached to the top left of a value outliner, and drop in in the middle of nowhere, the effect is unclear. After, if you move the target, the = re-attaches. I would think that unattaching would have broken the link (nil it out). Is there a use case for leaving it hanging out? If not, it might be good to have a drop-rejected animation that brings it back to its original target unless it is dropped on a valid new target.

2 1

Improving the Self website
by Russell Allen 28 Jan '14

28 Jan '14

Hi all, One of the main comments from the Mallard release was that the website was showing its age - it didn't give important information up front, assumed too much knowledge of what Self was and looked tired. I've tried to do a first pass at a better website. It should be good on mobiles as well as desktops, and hopefully is cleaner. We need I think short videos covering simple Self tasks to add. Does anyone have any suggestions how to go about this? eg what software to use etc? I assume we would just upload to YouTube and then embed in the site. The first draft of the website is online at: www.self.messaging.io Have a look and let me know what you think. For extra fun, I'm serving it from a Self webserver :) Russell

3 6

Two classes of Input Events?
by sean＠clipperadams.com 26 Jan '14

26 Jan '14

http://docs.selflanguage.org/4.5/morphic.html#handling-user-input mentions: KeyDown events... are dispatched using the “falling through the submorphs” However, playing with self 4.5.0, it seems that my morph only receives key down events if I subscribeKeyboard:. Did something change? Thanks! - Sean DeNigris

1 1

Demo-Snapshot?
by Michael Paap 26 Jan '14

26 Jan '14

Hello all, will there be a demo-snapshot for Mallard? I found it very useful in 4.4, especially the tutorials. Greetings, Michael

1 0

Avacado: JavaScript trying to be more like Self
by Casey Ransberger 23 Jan '14

23 Jan '14

http://avocadojs.com Interesting what they couldn't do. JS has insufficient reflection to see inside of JS functions. No debugger, which they said had to do with single-threadedness. No snapshots, because the browsers don't do that, so it relies on the Transporter. Still, it's kind of cool if you've got to write JavaScript. --Casey

4 4

trashcan problems
by Michael Paap 19 Jan '14

19 Jan '14

Hello list, I've got some really stupid problem. I am using the demo snapshot of Self 4.4 to play around with the system. And I repeatedly had the experience, that I dismissed some object by accident. You see: I need a trashcan. There is one in the clean snapshot, but not in the demo. I am sure, it /must/ be simple to get one an to integrate it in the world. Can someone give me a hint? BTW: If you would like to have some information, of which kind the problems are, which may prevent a programmer, familiar with Linux and Windows and state-of-the-art-IDEs, but not with exotic stuff like MacOS or Emacs (OMG!), I could give you a recapitulation of my experiences of the last weeks. ;-) Greetings, Michael

2 1

Re: [self-interest] [Ann][Quack] Self 4.5.0 ("Mallard") released
by Baltasar 14 Jan '14

14 Jan '14

Hi, everyone! > OK, so Self Mallard (4.5.0, named after the the majestic, awe inspiring Anas platyrhynchos) is now live at: > http://selflanguage.org/files/releases/4.5/Self-4.5.0.dmg > http://selflanguage.org/files/releases/4.5/Self-4.5.0.tar.gz First of all, this is an excellent work. Kudos for the effort! Okay, I've tried and installed it. Since it has been a very long time since I played with Self, it was more like a newbie experience than anything else. I hadn't read the blog yet, so I Hadn't learn about the -s option. After unpacking, the first thing I tried was: == $ ./Self bash: ./Self: Permission denied baltasarq@PC-baltasarq:~/Desktop/Self-4.5.0$ chmod +x Self == This is a little bit annoying, but anyway. == $ ./Self Self Virtual Machine Version 4.1.13, Sat 11 Jan 14 21:43:40 Linux i386 (4.4-272-g885323f) Copyright 1989-2003: The Self Group (type _Credits for credits) for I386: LogVMMessages = true for I386: PrintScriptName = true for I386: Inline = true for I386: SICDeferUncommonBranches = false (not implemented) for I386: SICReplaceOnStack = false (not implemented) for I386: SaveOutgoingArgumentsOfPatchedFrames = true #VM == Well, all right, let's try the magic word: == #VM help To begin using Self, you must read in the world of Self objects. To read in the world, type: 'worldBuilder.self' _RunScript When this process is complete, you will be at the Self prompt. At the Self prompt, you can start the user interface by typing: desktop open == So now, the next step is obvious: == #vm 'worldBuilder.self' _RunScript A lookup error happened while sending the message primitiveFailedError:Name: to 'worldBuilder.self'. Subsequently, the lookup error message undefinedSelector:Receiver:Type:Delegatee:MethodHolder:Arguments: was sent to <2>, and was also not understood, causing the process to be aborted by the Self VM. (...) == This is not good. I exited Self by hitting ^C many times... == VM# ^C^C^C^C^CSelf VM warning: press one more ^\ or ^C to force abort ^C Do you really want to force a 'hard' abort (y/n)? y Self VM fatal error (/home/russell/self/vm/src/any/os/sig.cpp, line 191): forced ^\ abort VM Version: 4.1.13, Sat 11 Jan 14 21:43:40 Linux i386 (4.4-272-g885323f) Self process 25883 on PC-baltasarq has crashed. Do you want to: 1) Quit Self (optionally attempting to write a snapshot) 2) Try to print the Self stack 3) Try to return to the Self prompt 4) Force a core dump 5) Print the interrupted context registers Your choice: 1 Enter snapshot name (hit return to omit snapshot) > No snapshot specified, will skip this step == Next thing I tried was to drop Clean.snap onto the Self executable (nothing happened), or call Self with the --help option (a standard option nowadays). This executed the Self console without any noticeable change. > I've lightly updated the website, and put a blog post up at: > http://blog.selflanguage.org/2014/01/12/self-mallard-4-5-0-released/ Thanks to that, I was able to finally fire up Self. == $ ./self -s Clean.snap == And the main problem comes when you close the window of the desktop. The console does not exit, and the prompt just shows: "Self 1". I tried the other magic word: == "Self 1" quit Save to Clean.snap before quitting? y => save, then quit n => quit without saving RET => cancel Response: y Snapshot file name? [RET => Clean.snap] Saving object code in snapshot is currently disabled. 1. Enable 2. Disable [Type the number of your selection] 1 Snapshot compression is currently enabled. 1. Enable 2. Disable [Type the number of your selection] 1 Code cache size (code only) ? [RET => 16777216] Change object heap sizes y/[n] ? n Starting: Thoroughly cleaning up memory...... Finished: Thoroughly cleaning up memory... Starting: Writing snapshot to Clean.snap... sigchild ignored Finished: Writing snapshot to Clean.snap == Save it or not? Enable compression? Cache size? Shouldn't the image just be saved automatically? Maybe copying the original one as Clean.snap.bak? The thing I want to state is that Self, in its current state, is not very friendly for beginners. Maybe just a "start.sh" with the "chmod +x Self && ./Self -s Clean.snap" would just solve the problem of beginning with Self. About finishing, I think it should be completely simplified. Another, completely different issue is how bad the font is seen in Ubuntu. Maybe is just that the system font is too small in the title window, and in the buttons: Get it, Do it and Close. Can that be somehow fixed? Finally, in the self page the tutorial in: http://selflanguage.org/_static/tutorial/index.html is recommended. The problem is how to start. Nowhere is to be found how to create a new, empty object, so you can start with the balance example. Hope this, somehow, helps. -- Baltasar García Perez-Schofield (jbgarcia(a)uvigo.es) Dpt. Informática, Universidad de Vigo, España http://webs.uvigo.es/jbgarcia/

8 12

Parent Priorities
by Michael Paap 13 Jan '14

13 Jan '14

Hello SELF-group, some weeks ago I did read about "Prioritized Parents" "Parents are Shared Parts of Objects: Inheritance and Encapsulation in SELF". Some unsuccessfull tryouts and some searching in the archive of this mailing list later I had to realize, that this feature is no longer supported in current releases of Self. And in "Programming as an Experience" I found some comprehensible arguments for the decision to remove it. But... in the handbook (Self Handbook Documentation, Release 2 for Self 4.4) in chapter 8.6 Lookup errors I found the following statement: "The lookup found two matching system slots which means the message is ambiguous. The error message also says where the matching slots were found. Ambiguities can often be resolved by changing parent priorities." Uuuh? May it be, that there is a relict of former times in the handbook? Or did I misunderstand something and the statement means something different? Is it possible to change paent priorities, ad if so: How is it done? Greetings, Michael

3 3

[Ann][Quack] Self 4.5.0 ("Mallard") released
by Russell Allen 13 Jan '14

13 Jan '14

Hi everyone, OK, so Self Mallard (4.5.0, named after the the majestic, awe inspiring Anas platyrhynchos) is now live at: http://selflanguage.org/files/releases/4.5/Self-4.5.0.dmg http://selflanguage.org/files/releases/4.5/Self-4.5.0.tar.gz I've lightly updated the website, and put a blog post up at: http://blog.selflanguage.org/2014/01/12/self-mallard-4-5-0-released/ Happy quacking, Russell

1 0

[Ann] Native
by Russell Allen 12 Jan '14

12 Jan '14

Hi guys, I've pushed a new branch called 'native' to github.com/russellallen/self which adds a new primitive based on the ideas in the Pharo NativeBoost package, but simpler and with fewer features. If you build a vm and image based on that branch then you should be able to play with it. I've pasted below a description. I've only tested it so far with MacOS X and not Linux. Cheers, Russell Native ====== Last updated 4 January 2014 Introduced in Self 4.5, the ``native`` module allows you to generate and run native machine code for your platform from within a Self World. It is inspired by the work of Igor Stasenko on NativeBoost for Pharo. Underlying Primitive -------------------- The core functionality of the ``native`` system is the primitive ``_RunNativePassing:With:With:With:With:With:With:With:``, which when run on a byteVector containing the machine code for a standard C function using the "cdecl" calling convention, and passing eight other byteVectors (from the machine code side seen as "char*") will run the machine code in byteVector. byteVectors inherit from ``traits byteVector`` the helper methods ``runNative``, ``runNativePassing:``, ``runNativePassing:With:`` etc. Marshalling ----------- The running machine code does not understand any of the Self World. It cannot access the virtual machine or Self objects. It can only see up to eight pointers to char arrays passed to it. It is the responsiblity of the calling Self code to marshal any objects into appropriate byteVectors to pass to the native code, and to unmarshal those byteVectors after the native code has run. The native code should be careful not to overflow the passed byteVector buffers. It goes without saying that any native code run through the ``runNative`` primitive can blow up the virtual machine in a number of exciting and profitable ways. There are a number of useful methods in ``traits byteVector`` to help with marshalling and unmarshalling, including ``cIntSize: bitSize Signed: bool At: index Put: anInt``, ``cIntSize: bitSize Signed: bool At: index`` etc in the "C Types" category. Coding ------ Pharo's NativeBoost implements a x86 and x64 assembler in Smalltalk. As yet Self doesn't have this, but there is a helper object ``native support nasm`` which covers calling out to the Nasm_ assembler if it is installed. .. _Nasm: http://www.nasm.us Native Modules -------------- The customary way to package native code is to create a object which delegates to ``native support nativeModuleParent``, and which has at least a single assignable slot ``currentPlatform``, and a single method ``adjustToPlatform``, which the VM will call at startup, and has the responsibility of ensuring that currentPlatform points to an appropriate platform specific object. It is also customary to put a wrapper object around this ensemble, with publicly accessible methods. See ``native example``. FFI --- It is possible to use ``native`` in conjunction with a ``fctProxy`` to dynamically call external C functions. In general, the method is to use ``foreignCodeDB at: pathToLib IfFail: failBlock`` to load the external library, ``lib lookupFunction: functionName`` to get the address of the function, ``fctProxy _PointerInByteVector: byteVectorSize4`` to get the C pointer, then pass that byteVector as a parameter to your native code. See ``native exampleBinding``. Limitations ----------- Callbacks aren't (yet) supported. Unlike NativeBoost, the machine code has no access to the virtual machine or to Self objects. Hints ----- If you wish to file out generated machine code stored in byteVectors, you will need to view the annotation of the byteVector and set the ``isComplete`` flag to ``Yes``, otherwise the ``transporter`` will complain.

4 15

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

Self-interest January 2014