Squeak-dev April 2020

squeak-dev@lists.squeakfoundation.org

39 participants
235 discussions

[BUG] Text drag and drop to workspace unreliable
by Jakob Reschke 24 Aug '20

24 Aug '20

Hi, I just noticed that the newly introduced text drag and drop sometimes does not work when dragging text between workspaces. It has always worked until today so I suspect the culprit is among the more recent changes. Problem: When dragging the selected text from one workspace to another, sometimes an object reference is copied to the bindings of the workspace instead of pasting the text. In the target text, the newly created variable name appears instead (e. g. text3997962) as if I had dropped another non-text object. At least the text is not removed from the source workspace when this happens... Problem #2: It does not always happen. When dragging from a third workspace it suddenly worked and then also from the second again. Then I reset my texts back to the old state and now it is broken again. Expected behavior: The selected text from the source workspace is cut from there, and pasted at the drop hand location in the target workspace. Currently I have three workspaces #1, #2, #3. If I drag the selection from #1 to #2, it works. If I drag from #1 to #3 it does not. Dragging from #2 to #3 does not work. Dragging from #2 to #1 does work. Now dragging from #1 to #3 suddenly does work... I dragged to the start of the text this time. Dragging back from #3 to #1 works. Dragging from #1 to #3 to the end of it also works. Dragging back from #1 to #3. Resetting (Cmd+l, lowercase L) text of #3. Dragging from #1 to the end of #3 does not work again. Variable pasted at the start of the text. Repeated dragging of the same text from #1 to #3 always pastes the variable. Dragging to the middle of #3 works! Dragging the pasted text inside of #3 to the end removed the text and pasted a variable instead. Hypothesis: it does not work to drop the text beyond the TextMorphForEditView. And so far I might just have been lucky to always have enough blank lines at the end of my target workspace. Workspaces #1 and #2 from above contain much more text than #3 and fill the whole editing area of the workspace, so this would be consistent. Squeak5.3beta latest update: #19300 Kind regards, Jakob

2 3

TrueTypeFont and Unicode Characters
by Eric Gade 05 Aug '20

05 Aug '20

Hello, I am trying to load a particular TrueTypeFont that represents the Cuneiform signs in the Cuneiform Unicode block. You can find some information on this page <https://www.hethport.uni-wuerzburg.de/cuneifont/>. Specifically, I'm attempting to work with the Neo-Assyrian font on that page ('Assurbanipal.ttf' in the zip file). I have also tried the Old Babylonian fonts with similar results. I have "imported" this font using the FileList tool and it does appear as a TextStyle. However, if I attempt to run: ``` m := TextMorph new beAllFont: ((TextStyle named: 'Assurbanipal') fontOfSize: 36); backgroundColor: Color white. signString := (Unicode value: 16r12038) asString. m contents: signString. m openInHand. ``` I get a TextMorph that appears blank. Either I am doing something incorrectly or the system cannot render the glyph at that code point (which should be the sign "AŠ"). Any ideas? -- Eric

5 9

Support an updated image format number for Sista
by David T. Lewis 19 Jul '20

19 Jul '20

Hi Eliot, This is a belated follow up on earlier private discussion during a Squeak board meeting a few months ago, in which we agreed that identifying "Sista in use" in the image format number would be a useful thing to do. I previously added the documentation for this in ImageFormat-dtl.39.mcz (as well as the updated ckformat.c program to match). Attached here is a change set that can be loaded into a VMMaker image to add VM support for setting and reading the updated image format numbers. Here I used a primitive to initially set or clear the "sista bit". There are other ways this could be done, but this at least gets us started with a working implementation. By way of explanation, I am copying below both the change set preamble and the earlier ImageFormat commit notice: === Change Set: VMM-Sista-bit-dtl Date: 12 January 2020 Author: David T. Lewis Let the image inform the interpreter that alternate bytecodes either are or are not in use, and remember that setting when writing or reading the image format number in a shapshot file header. Provide primitiveMultipleBytecodesActive to allow the setting to be updated from the image. Expected usage is that the image can call this primitive after recompiling with Systa bytecodes, or after reverting back. Note that an alternative implemenation would be to use a VM parameter. The primitive was chosen here because it can perform parameter checking and does not require coordinated update the the VM parameter usage. Expected usage is that this feature will remain unused for some grace period (perhaps a few months) during which an updated VM can be widely circulated. After that period, images using Sista bytecodes may be safely saved with the updated image format number indicating that multiple bytecodes are in use. This change set includes helper methods in SmalltalkImage for testing or setting the multipleBytecodeSetsActive setting in the VM. Note the change set postscript which is necessary for initializing a class variable prior to code generation. === Name: ImageFormat-dtl.37 Author: dtl Time: 20 July 2019, 5:23:59.004 pm UUID: 52b464d5-0bf7-4326-bde3-4f18ad70b239 Ancestors: ImageFormat-dtl.36 Identify extended bytecodes for Sista in the image format number. Bit 10 identifies an image that contains extended bytecodes. Thus a 32 bit Spur image is 6521, with Sista it is 7033, and a 64 bit Spur image is 68021, with Sista it is 68533. It is expected that additional bytecode sets can be identified by an additional field in the image header, probably 32 bits containing two 16 bit fields, if both zero then Sista. Per discussion with Eliot and Bert in a Squeak oversight board meeting 2019-05-15. Dave

1 3

Error in Promises/A+ reject implementation?
by Jakob Reschke 05 Jun '20

05 Jun '20

Hello all, I think our Promise behaves wrong, but I would like a second opinion. We have this test case: testifRejectedRunsBlockIfPromiseFails | p q error | error := nil. p := Promise new. q := p ifRejected: [:e | error := e]. p rejectWith: KeyNotFound new. self assert: q isRejected. self assert: KeyNotFound equals: error class. But the JavaScript behavior is actually different: x = {}; p = Promise.reject("Rejected"); q = p.then(null, error => error); q.then(value => { x.value = value; }, error => { x.error = error }); x // <- { value: "Rejected" } The difference: the chained promise q is resolved instead of rejected. Otherwise x would look like this: { error: "Rejected" }. I think the behavior in our Promise is also in violation with the Promises/A+ specification it claims to implement (https://promisesaplus.com/): [...] 2.2.7 then must return a promise. promise2 = promise1.then(onFulfilled, onRejected); 2.2.7.1 If either onFulfilled or onRejected returns a value x, run the Promise Resolution Procedure [[Resolve]](promise2, x). 2.2.7.2 If either onFulfilled or onRejected throws an exception e, promise2 must be rejected with e as the reason. [...] So if I understand this correctly, since in the test case Promise p gets rejected and the ifRejected: block evaluates to e without being curtailed by an exception, Promise q should be resolved, not rejected, with e, according to §2.2.7.1. Do you agree and should we fix this? On an only slightly related note, I am unhappy that I cannot debug errors in promise handler blocks (then: [...] ifRejected: [...]). It catches all those errors and just rejects the promise. In a server environment that would probably be ok, but at development time at least I want to have the chance to investigate the problem in a debugger. Especially those programmer's errors resulting in MessageNotUnderstood, which I could fix on the fly and let the program proceed. In JavaScript you cannot resume exceptions, but in Smalltalk you can. It might be nice to turn the on: Error do: [:e | p rejectWith: e] code in Promise into an ifCurtailed: [p rejectWith: e] to allow exceptions/errors to be resumed before rejecting the promise. The promise would still be rejected if you press Abandon in the debugger. Problem: further errors in the rejected handlers would result in "Unwind error during termination". This could in turn be avoided by rejecting the promise in the "future", that is the next cycle of the world. Opinions? Kind regards, Jakob

4 5

porting Croquet to Squeak6.0 alpha...
by gettimothy 29 May '20

29 May '20

Okey dokey, Poking along, there is a stray glyph in OGLUnix openGLLibraryName after openGLLibraryName ^Smalltalk osVersion = 'linux' ifTrue: ['libGL.so.1'] ifFalse: ['GL'] I removed it in my install and got past that error. Working exclusively with Croquet(Master)... My next error is in OGLUnixX11LE(OpenGL)>>glMatrixMode: glMatrixMode: mode "This method was automatically generated." "void glMatrixMode(GLenum mode);" <apicall: void 'glMatrixMode' (ulong) module: '#openGLLibraryName'> ^self externalCallFailed The <apicall:...> fails How to think about this? Is Croquet behind OpenGL latest? Would teaching myself OpenGL programming be of use to the Croquet project? cheers, tty

6 16

Broken 5.3 downloads need fixing
by tim Rowledge 15 May '20

15 May '20

The VM included in the 5.3 package for the ARM linux systems is fatally broken due to a code generator bug; we had various discussions about the issue in mid-March. We also noticed a problem with the bitBLT code for all platforms that was fixed by VMMaker.oscog-nice.2732 We really need to push the magic buttons to replace the ARM package in order for it to run at all; updating the other platforms to fix bitBLT would be nice. A quick test of the next available VM on bintray suggests that the https://bintray.com/opensmalltalk/vm/download_file?file_path=squeak.cog.spu… version is functional - but maybe we want to use the latest? What needs doing to solve this for the download page? tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Useful random insult:- On permanent leave of absence from his senses.

4 15

Please try out | Inspector Refactoring =)
by Marcel Taeumel 11 May '20

11 May '20

Hi all! :-) Christoph (ct) and I did finally finish reviewing his propsed changes and refactorings to our Inspector framework. Please find attached a changeset to try out those updates during the next 1-2 days. Of course, we will continue working on related issues even after merging it into Trunk. You should not expect any major UX changes. ;-) We have now more than 200 tests to check all of the specialized inspector flavors. First, here is some history and context on this proposal: - http://forum.world.st/The-Inbox-Tools-ct-900-mcz-td5104860.html [http://forum.world.st/The-Inbox-Tools-ct-900-mcz-td5104860.html] - http://forum.world.st/Refactoring-Inspectors-td5111788.html [http://forum.world.st/Refactoring-Inspectors-td5111788.html] Besides some structural clean-up, which you can browse through the message categories of the Inspector class, there are a few new things, Christoph wanted to push out to the Squeak community to enjoy: 1. More Clear Extension Points Inspectors show fields that expose certain characteristics of the object-under-inspection. InspectorField holds the specification for each field. Starting point is Inspector >> #streamFieldsOn:. A simple example on how to build a field is in #fieldAllInstVars. Take a look at how MorphInspector, CompiledCodeInspector, BitsetInspector refine the basic Inspector code. Also explore how CollectionInspector adds features to browse and modify collection elements. 2. Custom Fields You can now add custom fields during your inspector session. Invoke the field-list menu and choose "add field..." at the bottom. In its current form, simple expressions such as "self color darker" can be used to specify a field's value. If you want to also let the user change a field's value, simple blocks such as "[:newValue | self color: newValue]" can be added, too. Modifications through the value pane should work for all fields. Just type a valid Smalltalk expression and accept the changes through the menu or by pressing [cmd]+[s]. 3. "Copy expression" It is now possible to retrieve a piece of Smalltalk code for each field that tells you how to retrieve that field's value to be used in you own code. While fields are usually specified with blocks in a generic form like this .... ... valueGetter: [:object | object instVarNamed: name]; ... The "copy expression" feature, which you can access from the field-list menu, gives you a more precise answer: self instVarNamed: 'fullBounds' Of course, you can then follow up to figure out whether there is an accessor available. But in case you wondered where the contents for "all inst vars" is coming from, "copy expression" gives you this: self longPrintString Who of you did already know that? :-) Note that there are still some quirks here and there. We need volunteers to improve it further. ;-) Take a look at InspectorField class >> #generateExpressionFrom:argumentNames:. Well, it is basically an "inline block/method" refactoring where the temps are hopefully literals to be replaced inline. Maybe this can be part of FullBlockClosure in the future. 4. BasicInspector It is now possible to safely inspect subclasses of ProtoObject, such as proxies, in the BasicInspector. In an inspector's field-list menu, just click "basic inspect" and compare the differences. The basic inspector uses the VM's mirror primitives, which you can find in Context under "mirror primitives". This also meant changing some "anObject basicInspect" to "ToolSet basicInspect: anObject" to not send too many messages to the object-under-inspection. --- As mentioned above, please try out the attached changeset during the next 1-2 days. On file-in, that changeset will close all existing inspectors on re-open them again after loading the code. So please take care of any important information stored in an inspector's code-expression field at the bottom. Best, Christoph (ct) & Marcel (mt)

5 21

The Trunk: Help-Squeak-Project-nice.80.mcz
by commits＠source.squeak.org 05 May '20

05 May '20

Nicolas Cellier uploaded a new version of Help-Squeak-Project to project The Trunk: http://source.squeak.org/trunk/Help-Squeak-Project-nice.80.mcz ==================== Summary ==================== Name: Help-Squeak-Project-nice.80 Author: nice Time: 29 April 2020, 10:36:37.48125 pm UUID: e5ced11f-73d5-1d4f-a9ac-5af520cc1abe Ancestors: Help-Squeak-Project-mt.79 Let Help survive to Inspector refactorings. =============== Diff against Help-Squeak-Project-mt.79 =============== Item was changed: ----- Method: SqueakMessageCategoriesHelp>>menu (in category 'menu') ----- menu "This category includes all methods related to menus, often context menus in particular. It contains methods defining the menu as well as often callbacks for the labels of menu items or the actions to be executed. Sometimes the menus are split up in different categories if there are multiple menus or submenus." + ^ {Inspector>>#mainFieldListMenu: . SystemWindow>>#makeClosable}! - ^ {DictionaryInspector>>#mainFieldListMenu: . SystemWindow>>#makeClosable}!

3 3

The Inbox: Chronology-Core-cmm.53.mcz
by commits＠source.squeak.org 02 May '20

02 May '20

Chris Muller uploaded a new version of Chronology-Core to project The Inbox: http://source.squeak.org/inbox/Chronology-Core-cmm.53.mcz ==================== Summary ==================== Name: Chronology-Core-cmm.53 Author: cmm Time: 29 April 2020, 6:51:53.53317 pm UUID: b930b208-c85e-4c6b-8254-2837493384bd Ancestors: Chronology-Core-nice.52 - Extend the API to allow an integral representation of DateAndTime's, by introducing DateAndTime class>>#fromMicroseconds: DateAndTime>>#asMicroseconds which go along with the existing API: DateAndTime class>>#fromSeconds: DateAndTime>>#asSeconds - Every Squeak with the substring "utcMicrosecond" in the selector name refers to the number of microseconds since the POSIX epoch (1/1/1971@00:00:00.000), __except for one__: Time utcMicroseconds. Therefore, to avoid astonishment when using it with the "utc" API of DateAndTime, rename it without the "utc" prefix. =============== Diff against Chronology-Core-nice.52 =============== Item was added: + ----- Method: DateAndTime class>>fromMicroseconds: (in category 'smalltalk-80') ----- + fromMicroseconds: anInteger + "Answer a DateAndTime anInteger microseconds since the Squeak epoch: 1 January 1901." + ^ self + utcMicroseconds: anInteger - (self epochOffset * 1000000) + offset: self localOffsetSeconds! Item was added: + ----- Method: DateAndTime>>asMicroseconds (in category 'smalltalk-80') ----- + asMicroseconds + "Return the number of microseconds since the Squeak epoch, 1/1/1901 @ 00:00:00." + ^ utcMicroseconds + self class epochOffsetMicros! Item was changed: ----- Method: Time class>>estimateHighResClockTicksPerMillisecond (in category 'clock') ----- estimateHighResClockTicksPerMillisecond | t0 t1 t2 t3 | "Count the ticks ellapsed during a 10ms busy loop" + t0 := Time microsecondClock + 200. + [Time microsecondClock >= t0] whileFalse. - t0 := Time utcMicrosecondClock + 200. - [Time utcMicrosecondClock >= t0] whileFalse. t1 := self highResClock. + [Time microsecondClock >= (t0 + 10000)] whileFalse. + t1 := self highResClock - t1 * 1000 // (Time microsecondClock - t0). - [Time utcMicrosecondClock >= (t0 + 10000)] whileFalse. - t1 := self highResClock - t1 * 1000 // (Time utcMicrosecondClock - t0). "Count the ticks ellapsed during a 20ms busy loop" + t0 := Time microsecondClock + 200. + [Time microsecondClock >= t0] whileFalse. - t0 := Time utcMicrosecondClock + 200. - [Time utcMicrosecondClock >= t0] whileFalse. t2 := self highResClock. + [Time microsecondClock >= (t0 + 20000)] whileFalse. + t2 := self highResClock - t2 * 1000 // (Time microsecondClock - t0). - [Time utcMicrosecondClock >= (t0 + 20000)] whileFalse. - t2 := self highResClock - t2 * 1000 // (Time utcMicrosecondClock - t0). "Count the ticks ellapsed during a 30ms busy loop" + t0 := Time microsecondClock + 200. + [Time microsecondClock >= t0] whileFalse. - t0 := Time utcMicrosecondClock + 200. - [Time utcMicrosecondClock >= t0] whileFalse. t3 := self highResClock. + [Time microsecondClock >= (t0 + 30000)] whileFalse. + t3 := self highResClock - t3 * 1000 // (Time microsecondClock - t0). - [Time utcMicrosecondClock >= (t0 + 30000)] whileFalse. - t3 := self highResClock - t3 * 1000 // (Time utcMicrosecondClock - t0). "Take the median of the 3 estimates as the best" ^ t1 <= t2 ifTrue: [t2 <= t3 ifTrue: [t2] ifFalse: [t1 <= t3 ifTrue: [t3] ifFalse: [t1]]] ifFalse: [t1 <= t3 ifTrue: [t1] ifFalse: [t2 <= t3 ifTrue: [t3] ifFalse: [t2]]]! Item was changed: ----- Method: Time class>>localMicrosecondClock (in category 'clock') ----- localMicrosecondClock "Answer the local microseconds since the Smalltalk epoch (January 1st 1901, the start of the 20th century). The value is derived from the current UTC wallclock time and the image's current notion of time zone." + ^self microsecondClock + (DateAndTime localOffset asSeconds * 1000000)! - ^self utcMicrosecondClock + (DateAndTime localOffset asSeconds * 1000000)! Item was added: + ----- Method: Time class>>microsecondClock (in category 'clock') ----- + microsecondClock + "Answer the UTC microseconds since the Smalltalk epoch (January 1st 1901, the start of the 20th century). + The value is derived from the Posix epoch with a constant offset corresponding to elapsed microseconds + between the two epochs according to RFC 868." + <primitive: 240> + ^0! Item was changed: ----- Method: Time class>>microsecondsToRun: (in category 'general inquiries') ----- microsecondsToRun: timedBlock "Answer the number of microseconds timedBlock takes to return its value." | startUsecs | (self useHighResClockForTiming and: [self highResClock ~= 0]) ifTrue: [ ^(self nanosecondsToRunHighRes: timedBlock) + 500 // 1000]. + startUsecs := self microsecondClock. - startUsecs := self utcMicrosecondClock. timedBlock value. + ^self microsecondClock - startUsecs! - ^self utcMicrosecondClock - startUsecs! Item was changed: ----- Method: Time class>>millisecondClockValue (in category 'general inquiries') ----- millisecondClockValue "Answer the value of the millisecond clock." + ^self microsecondClock // 1000! - ^self utcMicrosecondClock // 1000! Item was changed: ----- Method: Time class>>nanosecondsToRun: (in category 'general inquiries') ----- nanosecondsToRun: timedBlock "Answer the number of nanoseconds timedBlock takes to return its value. Use high resolution clock if available and preferred." | startUsecs | (self useHighResClockForTiming and: [self highResClock ~= 0]) ifTrue: [ ^(self nanosecondsToRunHighRes: timedBlock)]. "Fallback to microseconds clock" + startUsecs := self microsecondClock. - startUsecs := self utcMicrosecondClock. timedBlock value. + ^self microsecondClock - startUsecs * 1000! - ^self utcMicrosecondClock - startUsecs * 1000! Item was added: + ----- Method: Time class>>totalMicroSeconds (in category 'smalltalk-80') ----- + totalMicroSeconds + "Answer the total seconds since the Squeak epoch: 1 January 1901, in local time." + + ^self localMicrosecondClock // 1000000! Item was changed: ----- Method: Time class>>utcMicrosecondClock (in category 'clock') ----- utcMicrosecondClock "Answer the UTC microseconds since the Smalltalk epoch (January 1st 1901, the start of the 20th century). The value is derived from the Posix epoch with a constant offset corresponding to elapsed microseconds between the two epochs according to RFC 868." <primitive: 240> + self deprecated: 'Use #microsecondClock'. ^0!

4 6

Installer metacello
by Nicolas Cellier 01 May '20

01 May '20

Hi all, I wonder why we haven't an InstallerMetacello flavour? We could write things like: Installer metacello configurationOf: 'Xtreams'; version: #stable; load. It would pick the Configuration in default repository ( http://www.squeaksource.com//MetacelloRepository). Of course, we could change that with Installer metacello repository: 'http://foobar.com' etc... Any taker?

7 24

← Newer
1
2
3
4
5
6
7
8
9
...
24
Older →

Jump to page: