Squeak-dev August 2017

squeak-dev@lists.squeakfoundation.org

42 participants
86 discussions

Morphic red box with big X - how to get the morph to redraw?
by Chris Cunningham 29 Aug '17

29 Aug '17

Hi. Normally when I have a morphic drawing failure (usually due to a bug in my drawing code) and the whole morph turns into the yellow X on the red box, I just delete the morph and make a new one. However, I'd like to get it running/drawing again instead. Is there a trick to get the morph to start drawing again? Thanks, cbc

3 3

Unable to open pull request via GitHub for squeak.org website repository
by tim Rowledge 26 Aug '17

26 Aug '17

I wanted to revise the Scratch related project section. I cloned the squeak.org repository, opened in githubdesktop, made a branch, edited the relevant md file, saved it, committed it, then tried to open a pull request. Authentication fails. Apparently I’m logged in ok. What next? tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim What passes for common sense is always revisable

8 15

FFI pinning examples?
by Phil B 25 Aug '17

25 Aug '17

I remember reading a while back that Spur added memory pinning and was curious if any Squeak packages has made use of it yet? (Looking for an example or two to get me started)

4 7

Adding a NuSCratch zip to ftp.squeak.org or similar
by tim Rowledge 24 Aug '17

24 Aug '17

I’d like to out a zip of a (hopefully) useful release of NuScratch in a suitable place; ftp.squeak.org seems a plausible option. So far as I can see I don't have any write permissions there (and probably shouldn’t have) so hints as to how to achieve this, or indeed suggestions for other good servers to use, would be welcomed. tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Strange OpCodes: L: Lie!

6 11

New VM tech (update/survey/beta information)
by Zak Fenton 22 Aug '17

22 Aug '17

Hi everyone! Some of you may recall my earlier posts about starting a business around a new VM technology. I'm sorry I left some of you hanging and promised updates... But the good news is, after much perseverance, it looks like the business will be going ahead - I've secured enough funding for the short-term, subject to the outcome of some feasibility studies. To be clear, the technology is not specific to Smalltalk users, but it does include a Smalltalk VM as well as many other features which may be of interest to many of you, so I'd like to make sure Smalltalk users are represented in my initial market research. It's a very short survey and should take less than five minutes to complete. In return, you'll get a chance to play with the first beta release as it becomes available over the coming months (however places in the beta program are limited, so availability cannot be guaranteed). The survey is here (the last page allows you to opt in to the beta program): https://www.surveymonkey.com/r/HTLLWTB Thanks, -Zak.

1 0

Caffeine :: Livecode the Web!
by Craig Latta 20 Aug '17

20 Aug '17

Hi folks-- I've released Caffeine, an in-browser IDE for livecoding JS frameworks from SqueakJS. It effectively turns a webpage into an interim dynabook. :) There are also new demos of... - Snowglobe (Smalltalk app streaming and remote GUI access) - HTML5 Morphic (each Morphic window on its own HTML5 canvas, relieving much of the computation burden on SqueakJS) - morphic.js (Jens Mönig's great standalone implementation of Morphic, which is much faster than the current SqueakJS Morphic) - three.js (WebGL 3D graphics) See the announcement on my blog at [1]. I would love any feedback you might have, and to collaborate on new use cases. thanks! -C *** [1] https://thiscontext.com/2017/06/22/caffeine-livecode-the-web -- Craig Latta Black Page Digital Amsterdam :: San Francisco craig(a)blackpagedigital.com +31 6 2757 7177 (SMS ok) + 1 415 287 3547 (no SMS)

5 6

Re: [squeak-dev] [Pharo-dev] Can we have subclasses of Process?
by Eliot Miranda 17 Aug '17

17 Aug '17

Hi Denis, as others have pointed out one can subclass Process. One can also add instance variables to Process. The constraint is that the first four (*) instance variables must be left unaltered. The VM makes use of the first four: Object subclass: #Link instanceVariableNames: 'nextLink' Link subclass: #Process instanceVariableNames: 'suspendedContext priority myList... (*five when we have the multithreaded FFI, the fifth being threadId used to bind processes to specific OS threads when required) The same goes for ProcessorScheduler - the VM uses quiescentProcessLists & activeProcess, and Semaphore - the VM uses firstLink lastLink & excessSignals. On Thu, Aug 17, 2017 at 5:03 AM, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: > Hi. > > Is it allowed to subclass Process? I remember that there were some > problems with it but I could not find any information. > > Generally I think about system processes which can be represented by > concrete subclasses. For example explicit UIProcess can be used to remove > many #isUIProcess conditions from system. > > Best regards, > Denis > -- _,,,^..^,,,_ best, Eliot

2 1

The Inbox: ShoutCore-tpr.59.mcz
by commits＠source.squeak.org 17 Aug '17

17 Aug '17

tim Rowledge uploaded a new version of ShoutCore to project The Inbox: http://source.squeak.org/inbox/ShoutCore-tpr.59.mcz ==================== Summary ==================== Name: ShoutCore-tpr.59 Author: tpr Time: 16 August 2017, 3:31:37.498742 pm UUID: 88fc7c09-2574-451d-817f-2b1289f62ae7 Ancestors: ShoutCore-ul.58 #sharedPools was making a Set with included nils, which means that SHParserST80>resolvePartial: breaks when 'c sharedPools do:[.... is used. This change prevents the obscene spectacle of a Set with deliberately included nils. =============== Diff against ShoutCore-ul.58 =============== Item was changed: ----- Method: SHMCClassDefinition>>sharedPools (in category 'act like a class') ----- sharedPools | d | d := Set new. classDefinition poolDictionaries do:[:each | + Smalltalk at: each asSymbol ifPresent:[:v| d add: v] ]. - d add: (Smalltalk at: each asSymbol ifAbsent:[nil]) ]. ^d!

2 1

SHParserST80>resolvePartial: and Sets with nil elements
by tim Rowledge 17 Aug '17

17 Aug '17

Whilst using Monticello’s ‘browse’ option to look at some code I tripped over a problem I just knew was going to be annoying someday - Sets with actual nils included as elements. That means that any enumeration has to explicitly handle the possibility of an item being a nil, instead of the normal and proper expectation that nils mean nothing and don’t get involved in Set>do:. So as a result in SHParserST80>resolvePartial: we have c sharedPools do: [:p | (p hasBindingThatBeginsWith: aString) ifTrue: [^#incompleteIdentifier]]. where c is an SHMCClassDefinition and the result of the #sharedPools should be a properly empty set. However, SHMCClassDefinition>sharedPools | d | d := Set new. classDefinition poolDictionaries do:[:each | d add: (Smalltalk at: each asSymbol ifAbsent:[nil]) ]. ^d … carefully adds a nil to the Set. So we get ‘nil hasBinding…’ and kaboom. *If* I assume that the real expectation is for the shared pools list to be like those elsewhere and not include nils, then a simple fix to | d | d := Set new. classDefinition poolDictionaries do:[:each | Smalltalk at: each asSymbol ifPresent:[:v| d add: v] ]. ^d … appears to make things happier. However, I don’t know anything about the intended working of Shout and I’d rather not simply stomp around in here. Also the code as-was has been around since 2010 so maybe something else has more recently changed and this is not the right thing to fix. Committed to inbox in the hope someone au fait with the code can look into it. tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Don't document the program; program the document.

2 1

File & Socket Handle Access
by Eliot Miranda 16 Aug '17

16 Aug '17

Hi All, recently Denis Kudriashov wanted to access the OS handle of a Socket from the image to pass through the FFI. David Lewis pointed out that the AioPlugin provides a primitive to do this. I asked David why he didn't add it to the SocketPlugin in the first place and he discussed Andreas Rabb's security concerns. It strikes me a) that accessing the OS handle of a file or a socket is extremely useful in some circumstances and invalid in others b) having handle access in a secondary plugin is inconvenient to say the least c) having access only to Socket handles and not to file handles is also inconvenient. So I propose adding a SecurityPlugin facility to control handle access, which would be granted by default, and to add primitives to both the SocketPlugin and FilePlugin to answer the native OS handle when the SecurityPlugin grants permission. I also propose adding a compilation flag to the Windows VM to allow compiling out Andreas' handle management, which consists merely of maintaining a table of the file handles (not the socket handles) that have been created by the VM. It strikes me that a much better solution is merely to encrypt the handle via xoring with a value created at random. I don't see how Andreas' solution adds value, especially since it is not implemented on other platforms, and merely adds overhead. Thoughts? Objections? _,,,^..^,,,_ best, Eliot

5 4

← Newer
1
2
3
4
5
6
7
8
9
Older →