June 2024 - Pharo-users - lists.pharo.org

Re: How can I remove an item on the main Pharo menu left over by a Roassal installation?

by vfclists

On Fri, 28 Jun 2024 at 19:00, Renaud de Villemeur via Pharo-users < pharo-users(a)lists.pharo.org> wrote: > The <worldMenu> pragma is automatically scanned by Pharo World, and > identified item are then picked and displayed. This makes it easy to add > custom menu item while loading external project. The downside is that > unless you want to disable completely this mechanism, you need to remove > this simple pragma from Roassal library to get thie menu entry removed (or > remove Roassal from your image). > > This won't have any impact on Roassal code or code in your image. It'll > simply remove this method from the list for <worldMenu> item. > > Renaud > > > > 28 juin 2024, 10 h 49 de vfclists(a)gmail.com: > > > > On Fri, 28 Jun 2024 at 12:34, Noury Bouraqadi <bouraqadi(a)gmail.com> wrote: > > Remove the <worldMenu> pragma form the method in charge of creating your > menu. > > Looking at senders of worldMenu lists all methods that create such menus. > > Noury > > > On Jun 28 2024, at 12:07 pm, vfclists <vfclists(a)gmail.com> wrote: > > After an attempt to install Roassal a persistent menu item has been to the > top level of the image's menu. > > The menu label is `Load full version` and after searching the image it is > coming from a `RSWorldMenu menu06LoadFullVersionOn` method which triggers > the `RSWorldMenu loadFullVersion` event. > > How do I remove this element from the main menu object which I suspect > might fix the problem? > > -- > Frank Church > > Shouldn't there be some setting that causes the Roassal menu to be > triggered? > > Isn't there some option to stop that menu from being called. > > My assumption is that there is some variable or setting persisting between > sessions that was created by the Roassal code during installation or > execution and simply has to be unset. > > Modifying library code is not something I'm keen on. > > -- > Frank Church > > > I have seen another example when its execution is made conditional by having a flat adde to the menu code - https://stackoverflow.com/questions/43021482/how-can-i-edit-the-worldmenu-i… . >From what I understand pragmas are somewhat related to compiled methods. Could it be that the menu is out of date but needs recompilation to remove this? I can't see the logic of this menu item still being present if the necessary libraries have been loaded. Roassal seems to be working fine. Do all Roassal users experience this menu item? -- Frank Church ======================= http://devblog.brahmancreations.com

June 28, 2024

Re: How can I remove an item on the main Pharo menu left over by a Roassal installation?

by Renaud de Villemeur

TheÂ <worldMenu> pragma is automatically scanned by Pharo World, and identified item are then picked and displayed. This makes it easy to add custom menu item while loading external project. The downside is that unless you want to disable completely this mechanism, you need to remove this simple pragma from Roassal library to get thie menu entry removed (or remove Roassal from your image). This won't have any impact on Roassal code or code in your image. It'll simply remove this method from the list for <worldMenu> item.Â Renaud 28 juin 2024, 10 h 49 de vfclists(a)gmail.com: > > > On Fri, 28 Jun 2024 at 12:34, Noury Bouraqadi <> bouraqadi(a)gmail.com> > wrote: > >> Remove the >> <worldMenu>>> Â pragma form the method in charge of creating your menu. >> >> Looking at senders of >> worldMenu>> Â lists all methods that create such menus. >> >> Noury >> Â >> >> On Jun 28 2024, at 12:07 pm, vfclists <>> vfclists(a)gmail.com>> > wrote: >> >>> After an attempt to install Roassal a persistent menu item has been to the top level of the image's menu. >>> >>> The menu label is `Load full version` and after searching the image it is coming from a `RSWorldMenu menu06LoadFullVersionOn` method which triggers the `RSWorldMenu loadFullVersion` event. >>> >>> How do I remove this element from the main menu object which I suspect might fix the problem? >>> >>> -- >>> Frank Church >>> > Shouldn't there be some setting that causes the Roassal menu to be triggered? > > Isn't there some option to stop that menu from being called. > > My assumption is that there is some variable or setting persisting between sessions that was created by the Roassal code during installation or execution and simply has to be unset. > > Modifying library code is not something I'm keen on. > > -- > Frank Church >

June 28, 2024

Re: How can I remove an item on the main Pharo menu left over by a Roassal installation?

by vfclists

On Fri, 28 Jun 2024 at 12:34, Noury Bouraqadi <bouraqadi(a)gmail.com> wrote: > Remove the <worldMenu> pragma form the method in charge of creating your > menu. > > Looking at senders of worldMenu lists all methods that create such menus. > > Noury > > > On Jun 28 2024, at 12:07 pm, vfclists <vfclists(a)gmail.com> wrote: > > After an attempt to install Roassal a persistent menu item has been to the > top level of the image's menu. > > The menu label is `Load full version` and after searching the image it is > coming from a `RSWorldMenu menu06LoadFullVersionOn` method which triggers > the `RSWorldMenu loadFullVersion` event. > > How do I remove this element from the main menu object which I suspect > might fix the problem? > > -- > Frank Church > > Shouldn't there be some setting that causes the Roassal menu to be triggered? Isn't there some option to stop that menu from being called. My assumption is that there is some variable or setting persisting between sessions that was created by the Roassal code during installation or execution and simply has to be unset. Modifying library code is not something I'm keen on. -- Frank Church

June 28, 2024

Re: How can I remove an item on the main Pharo menu left over by a Roassal installation?

by Noury Bouraqadi

Remove the <worldMenu> pragma form the method in charge of creating your menu. Looking at senders of worldMenu lists all methods that create such menus. Noury On Jun 28 2024, at 12:07 pm, vfclists <vfclists(a)gmail.com> wrote: > After an attempt to install Roassal a persistent menu item has been to the top level of the image's menu. > > The menu label is `Load full version` and after searching the image it is coming from a `RSWorldMenu menu06LoadFullVersionOn` method which triggers the `RSWorldMenu loadFullVersion` event. > How do I remove this element from the main menu object which I suspect might fix the problem? > -- > Frank Church > >

June 28, 2024

How can I remove an item on the main Pharo menu left over by a Roassal installation?

by vfclists

After an attempt to install Roassal a persistent menu item has been to the top level of the image's menu. The menu label is `Load full version` and after searching the image it is coming from a `RSWorldMenu menu06LoadFullVersionOn` method which triggers the `RSWorldMenu loadFullVersion` event. How do I remove this element from the main menu object which I suspect might fix the problem? -- Frank Church

June 28, 2024

This week (26/2024) on the Pharo Issue Tracker

by Marcus Denker

# Fixes - Fix/Patch production build zip error #16806 https://github.com/pharo-project/pharo/pull/16806 - File Browser: Add "disable preview" #788 https://github.com/pharo-spec/NewTools/pull/788 - Add #restore to the SpWindowPresenter API #1563 https://github.com/pharo-spec/Spec/pull/1563 # Refactoring Engine / Rules - Migrate Refactoring for Generating printOn: method #16789 https://github.com/pharo-project/pharo/pull/16789 - Add indentation rules #16785 https://github.com/pharo-project/pharo/pull/16785 # Cleanup - Move #isRoot to fast table #16800 https://github.com/pharo-project/pharo/pull/16800 - Cut dependencies of Iceberg on UI #1829 https://github.com/pharo-vcs/iceberg/pull/1829 # Website - Update 2024-06-07-MoocOO.md: make clear that Mooc Enrollment stays open #76 https://github.com/pharo-project/pharo-website/pull/76

June 28, 2024

Re: Does the Pharo and Squeak IDEs have some kind of screen and keyboard macro system?

by stephane ducasse

Yes! In pharo you can control the SDL2.0 driver. S > On 25 Jun 2024, at 21:14, vfclists <vfclists(a)gmail.com> wrote: > > > Does Pharo and Squeak have some screen activity recording system that can record your keystrokes and mouse clicks and convert them to UI commands for playback? > > > -- > Frank Church > StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 27, 2024

Re: [Pharo-dev] [ANN] New MOOC "Advanced Object-Oriented Design"

by Marcus Denker

Hello, Even though the course started, it stays open for enrollment (till June 10, 2025): https://www.fun-mooc.fr/en/courses/advanced-object-oriented-design-and-deve… > On 7 Jun 2024, at 10:23, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > [ANN] New MOOC "Advanced Object-Oriented Designâ > > <ð.png> Get ready for a new #MOOC "Advanced Object-Oriented Design"! > Understand the key mechanisms and the essence of OOD with the following topics: > > <â.png>Understanding messages > <â.png>Test-Driven Design > <â.png>Hooks: support for evolution > <â.png>Objects: state and behavior > <â.png>Elementary design patterns > <â.png>Double dispatch > <â.png>Object creation > <â.png>Sharing objects > <â.png>Inversion of control > <â.png>About types > > Trailer: https://www.youtube.com/watch?v=9HMYS-mzDKQ&t=1s > > You will find: > <ð¹.png> Videos to explain main concepts & tools > <ð¡.png>Quizzes and exercises that we've packaged together in companion exercises book > <ð.png>Lectures with command transcripts & explanations > > <ðï¸.png> Course starts on June 17 > <ð.png> Enroll for free: https://urlr.me/M6hSr <https://pharo.us11.list-manage.com/track/click?u=6f667565c2569234585a7be77&…> > <â±ï¸.png> Pace: Self-Paced, 1-year access > <ð£ï¸.png> English > > >

June 26, 2024

Does the Pharo and Squeak IDEs have some kind of screen and keyboard macro system?

by vfclists

Does Pharo and Squeak have some screen activity recording system that can record your keystrokes and mouse clicks and convert them to UI commands for playback? -- Frank Church

June 25, 2024

Re: A test message to see if subscription is active.

by vfclists

Yes :) On Mon, 24 Jun 2024 at 07:53, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > Hi Frank, > > you are still here :) > > > El 23 jun 2024, a las 7:21â¯p. m., vfclists <vfclists(a)gmail.com> > escribiÃ³: > > > > A test message to see if subscription is active. > > > > Confirmation took longer than I expected. > > > > -- > > Frank Church > > > > ======================= > > > -- Frank Church ======================= http://devblog.brahmancreations.com

June 25, 2024

Object-centric breakpoint experiment during ESUG'24

by Steven Costiou

Dear community, First, I would like to thanks again all the participants to our experiment with object-centric breakpoints in Pharo. Many people participated, and we plan to present our results in an informal manner at ESUG. However, from the scientifical point of view it would strenghten our analysis if we had more participants. The more participants we have, the more we can trust our results and take enlightened decisions about how to further develop the object-centric debugger. Therefore, the experiment remains open until July 10th (included). If you're interested you can still do it, the instructions are here https://github.com/Pharo-XP-Tools/XPImageGeneration/tree/exp-ocd Many people also declined to participate because of a lack of time, which we totally understand. This is why we also propose to you, if time is an issue, to participate to the experiment during the first days of esug. We can find a calm room in which we can supervise the experiment and collect your data (reminder: once in our server the data is anonymized). Please let us know if you would be potentially interested to participate in such setting during ESUG. Thanks in advance, Steven and Valentin, EVREF team.

June 24, 2024

Re: A test message to see if subscription is active.

by Guillermo Polito

Hi Frank, you are still here :) > El 23 jun 2024, a las 7:21â¯p. m., vfclists <vfclists(a)gmail.com> escribiÃ³: > > A test message to see if subscription is active. > > Confirmation took longer than I expected. > > -- > Frank Church > > ======================= >

June 24, 2024

A test message to see if subscription is active.

by vfclists

A test message to see if subscription is active. Confirmation took longer than I expected. -- Frank Church =======================

June 23, 2024

Method comparison

by Davide Varvello

In older Pharo versions it was possible to compare two methods selecting them via UI.Â I couldn't find this opportunity in Pharo10, Â so digging into the source code I found SpDiffPresenter, if you pass to it a couple of selectors it can compare them. Example: SpDiffPresenter new showOptions: true; leftText: (MyObject >> #firstSelector) sourceCode; rightText: (MyObject >> #secondSelector) sourceCode; contextClass: True; openÂ CheersDavide

June 21, 2024

This week (25/2024) on the Pharo Issue Tracker

by Marcus Denker

This week we backported some fixes to Pharo12, mostly related to high resolution rendering and Traits # Pharo 12 - Backport change of default value for setting #autoSetCanvasScaleFactor to true #16783 https://github.com/pharo-project/pharo/pull/16783 - Remove (wrong) primitive usage #16781 https://github.com/pharo-project/pharo/pull/16781 - [Backport] Port fixes on traits and metaclasses to Pharo 12 #16776 https://github.com/pharo-project/pharo/pull/16776 - Backport use of FormSet for #withCloseBox and #withPinBox on MenuTitleMorph #16774 https://github.com/pharo-project/pharo/pull/16774 - Backport selective removal of entries in #shrinkTo: on FreeTypeCache #16773 https://github.com/pharo-project/pharo/pull/16773 - [P12][Backport] fix #needRequestorScope of StDebuggerContextInteractionModel so that its bindings can be used when evaluating code in the debugger #784 https://github.com/pharo-spec/NewTools/pull/784 # VM - New harmonize rule #817 https://github.com/pharo-project/pharo-vm/pull/817 - Fixing warnings that are errors in newer versions of clang #813 https://github.com/pharo-project/pharo-vm/pull/813 - Adding an implementation of the aio.c using EPOLL in Linux. #805 https://github.com/pharo-project/pharo-vm/pull/805 - Ignoring EAGAIN in epoll_wait #818 https://github.com/pharo-project/pharo-vm/pull/818 # Features - Update roassal to v1.06f #16770 https://github.com/pharo-project/pharo/pull/16770 - Propose a DarkBlueTheme #16777 https://github.com/pharo-project/pharo/pull/16777 - Make setting pages less dense without comments for each setting #772 https://github.com/pharo-spec/NewTools/pull/772 # Fixes - test file dialogs returns the selected file #778 https://github.com/pharo-spec/NewTools/pull/778 - Revert "Use taskbar icon name for inspector and playground" #779 https://github.com/pharo-spec/NewTools/pull/779 - Fix issue when opening custom scope on calypso #783 https://github.com/pharo-spec/NewTools/pull/783 - Fix Open Playground menu item hanging the image #787 https://github.com/pharo-spec/NewTools/pull/787 - Add back old playground shortcut with deprecation #16788 https://github.com/pharo-project/pharo/pull/16788 - Remove (wrong) primitive usage #16780 https://github.com/pharo-project/pharo/pull/16780 - add missing accessor #1558 https://github.com/pharo-spec/Spec/pull/1558 - adding test to ensure it is possible to recompile code within a bloc with a dead home context twice #723 https://github.com/pharo-spec/NewTools/pull/723 # Refactoring Engine - Migrate Refactoring for Generating Equal and Hash Methods #16786 https://github.com/pharo-project/pharo/pull/16786 # Cleanup - Remove call to deprecated Author #83 https://github.com/pharo-vcs/libgit2-pharo-bindings/pull/83 - Remove deprecated classes and methods #758 https://github.com/pharo-spec/NewTools/pull/758 - Removing UIMnagaer dependencies from the debugger infrastructure #16015 https://github.com/pharo-project/pharo/pull/16015 - The debugger now delegate UI decisions to the backend provided by its application #667 https://github.com/pharo-spec/NewTools/pull/667 - Fix deprecated usage of StInspector #16769 https://github.com/pharo-project/pharo/pull/16769 - taskbarIcon moved to SystemWindow #16792 https://github.com/pharo-project/pharo/pull/16792 - Rename RBNode into ASTNode #16596 https://github.com/pharo-project/pharo/pull/16596

June 21, 2024

Re: Spec2 - Morph query & move window to another position

by Haroldo Stenger

hi Esteban I'm very sorry for the delay, and very grateful for your help. El lun, 17 jun 2024 a la(s) 2:39â¯a.m., Esteban Lorenzano via Pharo-users ( pharo-users(a)lists.pharo.org) escribiÃ³: > you need to position points/morphs in a canvas ? > in the WorldMorph. > In that case you need a) a canvas and b) the points. > The World Morph I see it as the canvas, the points are the windows (SpPresenter's) coordinates understood by moveTo: > This cannot (and should not) be achieved with different widgets (spec) but > with just one : a morph there acts as a canvas. > Yes, I try to understand. > So, instead of going to the window of the presenter (which,, by the way, > may or may not exist) I would just add a morph, e.g. : > > (this example is just a script, in real life you would do things > differently) > > presenter := SpPresenter new. > presenter layout: (SpBoxLayout newVertical > add: (morphPresenter := presenter newMorph); > yourself). > > morph := Morph new color: Color white. > morphPresenter morph: morph. > morph color: Color white. > 1 to: 10 do: [ :index | > | pointMorph | > pointMorph := Morph new > color: Color random; > extent: 10@10; > yourself. > pointMorph position: 300 atRandom @ 200 atRandom. > morph addMorph: pointMorph ]. > > presenter open. > got thru the example which is very nice. > getting the real size of a window is not really provided by the api except > for its initial status (after that, you can suscribe to the resize event if > you want to know the new size), but you shouldn't needed that (and getting > the SpWindow which is a morph is not the way to go, since this is not > portable). > Now I understand what subscriptions are ! > Spec is a framework to make and place **widgets**, as soon as you need > something that is not a widget, you need to go to a lower level than it > (hence, rendering a morph). > Correct. Nontheles, the SpPresenter subclassed objects answer the moveTo: x@y and move the window. Is the window itself considered a widget in this context? Just asking to learn and orient myself. I could more or less do what I originally wanted. A methiod of mine goes thru all opened Survey windows, and reposition themselves in a circle. I like the result , at least as a "variant" of the default positioning, and as an excercise of querying and moving. thanks again and best regards looking forward meeting you soon in Buenos Aires or wherever. Haroldo

June 18, 2024

Re: Spec2 - Morph query & move window to another position

by Esteban Lorenzano

I do not understand. you need to position points/morphs in a canvas ? In that case you need a) a canvas and b) the points. This cannot (and should not) be achieved with different widgets (spec) but with just one : a morph there acts as a canvas. So, instead of going to the window of the presenter (which,, by the way, may or may not exist) I would just add a morph, e.g. : (this example is just a script, in real life you would do things differently) presenter := SpPresenter new. presenter layout: (SpBoxLayout newVertical Â Â Â add: (morphPresenter := presenter newMorph); Â Â Â yourself). morph := Morph new color: Color white. morphPresenter morph: morph. morph color: Color white. 1 to: 10 do: [ :index | Â Â Â | pointMorph | Â Â Â pointMorph := Morph new Â Â Â Â Â Â color: Color random; Â Â Â Â Â Â extent: 10@10; Â Â Â Â Â Â yourself. Â Â Â pointMorph position: 300 atRandom @ 200 atRandom. Â Â Â morph addMorph: pointMorph ]. presenter open. this will give you something like this: getting the real size of a window is not really provided by the api except for its initial status (after that, you can suscribe to the resize event if you want to know the new size), but you shouldn't needed thatÂ (and getting the SpWindow which is a morph is not the way to go, since this is not portable). Spec is a framework to make and place **widgets**, as soon as you need something that is not a widget, you need to go to a lower level than it (hence, rendering a morph). Esteban On 17/06/2024 00:08, Haroldo Stenger wrote: > hi all there , I'm playing around with this example from the nice Spec > book, which I try to get right for spec2. That's all right. I modify > the example for having two counters, one bound to the class of the > Surveys and then one for each survey. That was a good learning > experience. Then I added a button to spawn further a Survey which > participates of the class-bound count of votes. That's all right too. > Then I started to write a method that given a Rectangle, and a number > n, gives an array of n Points that are in the enclosed circle of the > rectangle, more or less. Waht I want to achieve es that for each added > Survey, its Spec2 - Morph window gets positioned in one of these > points, but before rendering it, re position all the already existing > Spec2 - Morphs Survey windows in the rest of the points of the array > (being n the total number of Surveys after adding one). I'm facing two > challenges: (a) getting the list of current positions of Spec2 - Morph > Surveys already existing. (b) moving a Spec2 - Morph window to another > position. I'm pretty stuck trying to make my way for this two goals, > so any kind help will be really welcome. Yo can import the files into > a Pharo12 image with the File manager, then Changelist Browser and > selecting all the lines and "file in". > > thanks in advance ! > Haroldo >

June 17, 2024

Re: Spec2 - Morph query & move window to another position

by Haroldo Stenger

thanks to HernÃ¡n Morales Durand , he pointed that *CustomerSatisfactionPresenter being a subclass of **SpPresenter* , answers the message *SpPresenter* window which renders a *SpWindowPresenter* object and *SpWindowPresenter* window renders a *SpWindow* object, which answers the message *SpWindow* bounds and *SpWindowPresenter* moveTo: x@y moves the window to the point. any other suggestions are welcome too best regards Haroldo El dom, 16 jun 2024 a la(s) 7:08â¯p.m., Haroldo Stenger ( haroldo.stenger(a)gmail.com) escribiÃ³: > hi all there , I'm playing around with this example from the nice Spec > book, which I try to get right for spec2. That's all right. I modify the > example for having two counters, one bound to the class of the Surveys > and then one for each survey. That was a good learning experience. Then I > added a button to spawn further a Survey which participates of the class-bound > count of votes. That's all right too. Then I started to write a method > that given a Rectangle, and a number n, gives an array of n Points that > are in the enclosed circle of the rectangle, more or less. Waht I want to > achieve es that for each added Survey, its Spec2 - Morph window gets > positioned in one of these points, but before rendering it, re position > all the already existing Spec2 - Morphs Survey windows in the rest of the > points of the array (being n the total number of Surveys after adding one). > I'm facing two challenges: (a) getting the list of current positions of > Spec2 - Morph Surveys already existing. (b) moving a Spec2 - Morph window > to another position. I'm pretty stuck trying to make my way for this two > goals, so any kind help will be really welcome. Yo can import the files > into a Pharo12 image with the File manager, then Changelist Browser and > selecting all the lines and "file in". > > thanks in advance ! > Haroldo > >

June 17, 2024

Spec2 - Morph query & move window to another position

by Haroldo Stenger

hi all there , I'm playing around with this example from the nice Spec book, which I try to get right for spec2. That's all right. I modify the example for having two counters, one bound to the class of the Surveys and then one for each survey. That was a good learning experience. Then I added a button to spawn further a Survey which participates of the class-bound count of votes. That's all right too. Then I started to write a method that given a Rectangle, and a number n, gives an array of n Points that are in the enclosed circle of the rectangle, more or less. Waht I want to achieve es that for each added Survey, its Spec2 - Morph window gets positioned in one of these points, but before rendering it, re position all the already existing Spec2 - Morphs Survey windows in the rest of the points of the array (being n the total number of Surveys after adding one). I'm facing two challenges: (a) getting the list of current positions of Spec2 - Morph Surveys already existing. (b) moving a Spec2 - Morph window to another position. I'm pretty stuck trying to make my way for this two goals, so any kind help will be really welcome. Yo can import the files into a Pharo12 image with the File manager, then Changelist Browser and selecting all the lines and "file in". thanks in advance ! Haroldo

June 16, 2024

Pharo VM v10.2.1

by tesonep＠gmail.com

Hi, there is a new release of the VM ## v10.2.1 * Adding the missing tty.c file in the packaging. by @tesonep in https://github.com/pharo-project/pharo-vm/pull/771 * Do not allow comparing objects of different types by @guillep in https://github.com/pharo-project/pharo-vm/pull/772 * Prepare release 10.2.1 by @guillep in https://github.com/pharo-project/pharo-vm/pull/773 * Constant conversion warning by @guillep in https://github.com/pharo-project/pharo-vm/pull/777 * Fix tautological-pointer-compare warnings by @guillep in https://github.com/pharo-project/pharo-vm/pull/775 * Removed Cogit>>#voidNSSendCache: by @jordanmontt in https://github.com/pharo-project/pharo-vm/pull/776 * Re-enable incompatible-function-pointer-types warning by @guillep in https://github.com/pharo-project/pharo-vm/pull/778 * Removed unused functions from the C written by hand by @jordanmontt in https://github.com/pharo-project/pharo-vm/pull/780 * Fix function pointer comparison warning by @guillep in https://github.com/pharo-project/pharo-vm/pull/783 * Fix -return-type warnings by @doste in https://github.com/pharo-project/pharo-vm/pull/781 * Fix shift-negative-value warning and reenable warning by @guillep in https://github.com/pharo-project/pharo-vm/pull/785 * Remove warnings c compiler by @PalumboN in https://github.com/pharo-project/pharo-vm/pull/789 * Remove StackInterpreter interruptPending instance variable by @jordanmontt in https://github.com/pharo-project/pharo-vm/pull/793 * Do not use asserta: in a statement by @guillep in https://github.com/pharo-project/pharo-vm/pull/788 * Bump to MacOS 11 and above by @guillep in https://github.com/pharo-project/pharo-vm/pull/792 * Moving pharo.signatures files for OSX Bundle to Resources folder by @tesonep in https://github.com/pharo-project/pharo-vm/pull/770 * Do not retain any selectors in CogAbstractInstruction by @RenaudFondeur in https://github.com/pharo-project/pharo-vm/pull/798 * remove warning because of type and unsigned shift issues by @RenaudFondeur in https://github.com/pharo-project/pharo-vm/pull/800 * add tests for type harmonization by @RenaudFondeur in https://github.com/pharo-project/pharo-vm/pull/807 * Primitive format by @doste in https://github.com/pharo-project/pharo-vm/pull/802 * Adding an implementation of the aio.c using EPOLL in Linux. by @tesonep in https://github.com/pharo-project/pharo-vm/pull/805 * Fixing warnings that are errors in newer versions of clang by @tesonep in https://github.com/pharo-project/pharo-vm/pull/813 ## New Contributors * @RenaudFondeur made their first contribution in https://github.com/pharo-project/pharo-vm/pull/798 **Full Changelog**: https://github.com/pharo-project/pharo-vm/compare/v10.2.0...v10.2.1 More Information: https://github.com/pharo-project/pharo-vm/releases/tag/v10.2.1 -- Pablo Tesone. tesonep(a)gmail.com

June 14, 2024

This week (24/2024) on the Pharo Issue Tracker

by Marcus Denker

# Pharo 12 - Fix Issue: Backport Pharo12: Wrong source for CompiledBlocks due to incorrect bytecode to AST nodes mapping #16756 https://github.com/pharo-project/pharo/pull/16756 # Improvemetns - Adds a simple MoveTo taskbar menu item #16544 https://github.com/pharo-project/pharo/pull/16544 - replace JPEGReadWriter by PluginBasedJPEGReadWriter and deprecate theâ¦ #16535 https://github.com/pharo-project/pharo/pull/16535 - return the result of ok action and simplify file selection #776 https://github.com/pharo-spec/NewTools/pull/776 - withWindowDo: should return the result of the block #1555 https://github.com/pharo-spec/Spec/pull/1555 - Update after selection in Color Picker #766 https://github.com/pharo-spec/NewTools/pull/766 - New dialog to select multiple items using checkboxes #1548 https://github.com/pharo-spec/Spec/pull/1548 # Fixes - fix preLoading.st #1557 https://github.com/pharo-spec/Spec/pull/1557 - fixing recompilation of bloc in the debugger with dead home context whose method has already been recompiled in the debugger before #16319 https://github.com/pharo-project/pharo/pull/16319 - Fix comparison of two Deprecation objects #16735 https://github.com/pharo-project/pharo/pull/16735 - Updated text to new shortcut #764 https://github.com/pharo-spec/NewTools/pull/764 # Tools - Added new subclass with unified view based on SpDiffPresenter #1452 https://github.com/pharo-spec/Spec/pull/1452 - Add StProtocolNameChooserPresenter #759 https://github.com/pharo-spec/NewTools/pull/759 - Spotter split case #773 https://github.com/pharo-spec/NewTools/pull/773 - Use new StProtocolNameChooserPresenter to replace morphic tool #16690 https://github.com/pharo-project/pharo/pull/16690 - Icon for debugger in the toolbar #16661 https://github.com/pharo-project/pharo/pull/16661 - ProfilerUI: Added a warning when trying to compile and the variables are not declared #741 https://github.com/pharo-spec/NewTools/pull/741 - Update playground shortcut #738 https://github.com/pharo-spec/NewTools/pull/738 # Cleanups - Remove deprecated code #1545 https://github.com/pharo-spec/Spec/pull/1545 - Remove references to SystemAnnouncer singleton #742 https://github.com/pharo-spec/NewTools/pull/742 - Move some debugger related code of Process to debugger packages #15799 https://github.com/pharo-project/pharo/pull/15799 - Moved pragma declaration to top of methods #16728 https://github.com/pharo-project/pharo/pull/16728 - Categorization rules tests #16767 https://github.com/pharo-project/pharo/pull/16767

June 14, 2024

Re: [ANN] New MOOC "Advanced Object-Oriented Design"

by stephane ducasse

Thanks Some people here used it to assemble new lectures and they cherrypicked some topics. And this is also a material that teachers can use in their lectures. S > On 12 Jun 2024, at 22:08, Offray Vladimir Luna CÃ¡rdenas <offray.luna(a)mutabit.com> wrote: > > Congrats for this MOOC. > > We shared it on the Grafoscopio community, and hopefully some of us will take it. > > Cheers, > > Offray > > On 7/06/24 3:23, Marcus Denker wrote: >> [ANN] New MOOC "Advanced Object-Oriented Designâ >> >> <ð.png> Get ready for a new #MOOC "Advanced Object-Oriented Design"! >> Understand the key mechanisms and the essence of OOD with the following topics: >> >> <â.png>Understanding messages >> <â.png>Test-Driven Design >> <â.png>Hooks: support for evolution >> <â.png>Objects: state and behavior >> <â.png>Elementary design patterns >> <â.png>Double dispatch >> <â.png>Object creation >> <â.png>Sharing objects >> <â.png>Inversion of control >> <â.png>About types >> >> Trailer: https://www.youtube.com/watch?v=9HMYS-mzDKQ&t=1s >> >> You will find: >> <ð¹.png> Videos to explain main concepts & tools >> <ð¡.png>Quizzes and exercises that we've packaged together in companion exercises book >> <ð.png>Lectures with command transcripts & explanations >> >> <ðï¸.png> Course starts on June 17 >> <ð.png> Enroll for free: https://urlr.me/M6hSr <https://pharo.us11.list-manage.com/track/click?u=6f667565c2569234585a7be77&…> >> <â±ï¸.png> Pace: Self-Paced, 1-year access >> <ð£ï¸.png> English >> >> >> StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 12, 2024

Re: [ANN] New MOOC "Advanced Object-Oriented Design"

by Offray Vladimir Luna CÃ¡rdenas

Congrats for this MOOC. We shared it on the Grafoscopio community, and hopefully some of us will take it. Cheers, Offray On 7/06/24 3:23, Marcus Denker wrote: > > > [ANN] New MOOC "Advanced Object-Oriented Designâ > > > ð.pngÂ Get ready for a new #MOOCÂ "Advanced Object-Oriented Design"! > Understand the key mechanisms and the essence of OOD with the > following topics: > > â.pngUnderstanding messages > â.pngTest-Driven Design > â.pngHooks: support for evolution > â.pngObjects: state and behavior > â.pngElementary design patterns > â.pngDouble dispatch > â.pngObject creation > â.pngSharing objects > â.pngInversion of control > â.pngAbout types > > Trailer: https://www.youtube.com/watch?v=9HMYS-mzDKQ&t=1s > <https://www.youtube.com/watch?v=9HMYS-mzDKQ&t=1s> > > You will find: > ð¹.pngÂ Videos to explain main concepts & tools > ð¡.pngQuizzes and exercises that we've packaged together in companion > exercises book > ð.pngLectures with command transcripts & explanations > > ðï¸.pngÂ Course starts on *June 17* > ð.pngÂ Enroll for free: https://urlr.me/M6hSr > <https://pharo.us11.list-manage.com/track/click?u=6f667565c2569234585a7be77&…> > â±ï¸.pngÂ Pace: Self-Paced, 1-year access > > ð£ï¸.pngÂ English > >

June 12, 2024

Re: [Pharo-dev] [ANN] New MOOC "Advanced Object-Oriented Design"

by Noury Bouraqadi

Well done guys! ð This Mooc is must! I liked the video and forwarded to some people that would benefit from the Mooc. On Jun 7 2024, at 10:23 am, Marcus Denker <marcus.denker(a)inria.fr> wrote: > [ANN] New MOOC "Advanced Object-Oriented Designâ > > English

June 10, 2024

This week (23/2024) on the Pharo Issue Tracker

by Marcus Denker

# Pharo 12 - Fix/Update Iceberg #16752 https://github.com/pharo-project/pharo/pull/16752 - 16733-Backport-Pharo12-DebugPoint-on-class-side-method-16682 #16734 https://github.com/pharo-project/pharo/pull/16734 - fix bad label before first progress event. #1551 https://github.com/pharo-spec/Spec/pull/1551 - Updating LibGit2 #1827 https://github.com/pharo-vcs/iceberg/pull/1827 # Features - Add newTenured allocation methods #16729 https://github.com/pharo-project/pharo/pull/16729 - Fix ZnBufferedWriteStream does not support #reset #16727 https://github.com/pharo-project/pharo/pull/16727 - Primitive format #802 https://github.com/pharo-project/pharo-vm/pull/802 # Fixes - fix bad label before first progress event. #1552 https://github.com/pharo-spec/Spec/pull/1552 - add missing code from P12 #763 https://github.com/pharo-spec/NewTools/pull/763 - 16495-Wrong-source-for-CompiledBlocks-due-to-incorrect-bytecode-to-AST-nodes-mapping #16664 https://github.com/pharo-project/pharo/pull/16664 - Calypso cmd-v method versions shortcut #16742 https://github.com/pharo-project/pharo/pull/16742 - Fix Random seed with float value #16738 https://github.com/pharo-project/pharo/pull/16738 # Refactoring Engine - Add subclass driver #16353 https://github.com/pharo-project/pharo/pull/16353 - Refactor: inline method cleaning #16745 https://github.com/pharo-project/pharo/pull/16745 # Rules - Add setting rules feature #16750 https://github.com/pharo-project/pharo/pull/16750 - Move rules matadatas to class side #16705 https://github.com/pharo-project/pharo/pull/16705 # ClassBuilder / Traits - Fix multiple bugs in metaclass management of ShiftClassBuilder #16722 https://github.com/pharo-project/pharo/pull/16722 - Fix bug causing classes to lose their traits. #16739 https://github.com/pharo-project/pharo/pull/16739 - Use #update:to: instead of #make: and #fillFor: #16746 https://github.com/pharo-project/pharo/pull/16746 # Cleanups - Remove deprecated aliases of P12 #16659 https://github.com/pharo-project/pharo/pull/16659

June 7, 2024

[ANN] New MOOC "Advanced Object-Oriented Design"

by Marcus Denker

[ANN] New MOOC "Advanced Object-Oriented Designâ ï¿¼ Get ready for a new #MOOC "Advanced Object-Oriented Design"! Understand the key mechanisms and the essence of OOD with the following topics: ï¿¼Understanding messages ï¿¼Test-Driven Design ï¿¼Hooks: support for evolution ï¿¼Objects: state and behavior ï¿¼Elementary design patterns ï¿¼Double dispatch ï¿¼Object creation ï¿¼Sharing objects ï¿¼Inversion of control ï¿¼About types Trailer: https://www.youtube.com/watch?v=9HMYS-mzDKQ&t=1s You will find: ï¿¼ Videos to explain main concepts & tools ï¿¼Quizzes and exercises that we've packaged together in companion exercises book ï¿¼Lectures with command transcripts & explanations ï¿¼ Course starts on June 17 ï¿¼ Enroll for free: https://urlr.me/M6hSr <https://pharo.us11.list-manage.com/track/click?u=6f667565c2569234585a7be77&…> ï¿¼ Pace: Self-Paced, 1-year access ï¿¼ English

June 7, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by stephane ducasse

> On 5 Jun 2024, at 10:22, Tim Mackinnon <tim(a)testit.works> wrote: > > Hi Stef - I saw that blog post - which looks like its essentially the help text that comes up in the rewrite tool when you press the help button. I was going to submit to add to the help text as I think a better example that pulls all the ideas together is helpful. However, I have realised the Rewrite tool is not the main pharo-project repo and comes under NewTools? How do I find out where that repo is to submit a pr? > > I think the help button on that tool should launch the normal doc browser with the text in the normal help format vs a seperate window? The help browser could probably use some work but at least if help is one place it all will get updated together? Yes it can be nice to have a single documentation so that when we improve it will benefit directly to the tools. > > I was going to propose the following example which helped me a lot - if I wanted to rewrite how logging was done in the following example (and separate the Transcript line into 2 lines): > > appLog: anObject > "Log anObject for debugging" > | date | > date := Date today. > Transcript traceCr: 'debug: ', date printString, anObject printString. > date isLeapYear ifTrue: [ ^nil ]. > ^anObject > > I could write a match expression like this: > > appLog: `arg > | `@temps | > `.@before. > Transcript traceCr: `#str , `@exp1, `@exp2. > `.@after > > and a rewrite rule like this: > > appLog: `arg > | `@temps | > `.@before. > Transcript cr; traceCr: `#str, `@exp1. > `arg traceCr. > `.@after > > I think this demonstrates most of the concepts that caught me out - namely: you need to call out the variable matching explicitly -= the .@before (list statement matching doesn't match variables - and using @temps will match 0 or more vars, so its safer to include it if matching expressions that may have any otherwise nothing will match. > > If you want to match a string you need to match it as a literal e.g. `#str. > > You may need to match 0 or more statements before and after your target expression hence `.@before and `.@after but its important to include the trailing "." on the @before. to cut off before the expression you are trying to match next. > > I didn't fully understand in my example is why I can't match the remaining statements after the transcript string using just `.@exp1. I seem to have to call out both expressions which I now realise (and worth documenting) is because the AST for that line is actually VariableNode(transcript) + MsgNode(MsgNode(Literal+MsgNode(exp1)), MsgNode(exp2) and so you have to call them all out to match things or match the entire sequence with @stmnt (in my case I wanted to pick them apart). I think the top tip in all of this is to use the handy source code ast inspector to help you reason on the structure. > Sometimes this is arcane to me. I think that I would have to concentrate and do a full review of code using tons of new tests :) > The above is probably the gist of the kinds of docs we would need, and happy to submit a pr with this in it. please do. > > On Wed, 5 Jun 2024, at 6:35 AM, stephane ducasse wrote: >> Hi Tim >> >> This is on my todo to have a serious look at the pattern matcher >> Now one of the few documentation that we have in the >> blog and I ported it in the blog post in /doc folder. >> >> S >> >> >> >> >>> On 4 Jun 2024, at 01:35, Tim Mackinnon <tim(a)testit.works> wrote: >>> >>> Hey Gabriel - thanks for chipping in - you've made me realise that its perhaps easier to give a simpler example (for the multiple replacement)., and while trying to explain in simpler terms I managed to solve the problem myself - but it does demonstrate the docs around this need a bit more work as I didn't find it easy to reason on. >>> >>> I am still confused what the setting "this rule is for a method" does - so am keen for someone to shed light. >>> >>> Anyway for my example it seems that I have to be very specific in how I match things for example I had to use the search pattern: >>> >>> taskbarButtonMenu: `arg >>> | `@temps | >>> `.@before. >>> >>> moveSubmenu >>> addToggle: 'Move right' translated >>> target: self >>> selector: #taskbarMoveRight >>> getStateSelector: nil >>> enablementSelector: #canBeMovedToRight. >>> >>> `.@after >>> >>> And it took me ages to realise that I have to specify the temp vars matching, and also understand that the @before and @after both needed the "." to represent lists of expressions both before and after and then the really subtle "." after @before which I guess prevents greedy matching. >>> >>> With all of that in place, could then have the write rule add an extra statement after the moveSubmenu (or as many as I want) >>> >>> e.g. >>> >>> taskbarButtonMenu: `arg >>> | `@temps | >>> `.@before. >>> >>> moveSubmenu >>> addToggle: 'Move right' translated >>> target: self >>> selector: #taskbarMoveRight >>> getStateSelector: nil >>> enablementSelector: #canBeMovedToRight. >>> >>> moveSubmenu >>> addToggle: 'Move to...' translated >>> target: self >>> selector: #taskbarMoveTo >>> getStateSelector: nil. >>> >>> `.@after >>> >>> >>> I'm curious now whether there is a more efficient way to do what I have done? Possibly the {} syntax could help now i've figured out the basic matching. >>> >>> I really find it easy to confuse that if you don't specify the @temps section that can leave out a lot of matches (it essentialy means 0 or more temps vs. leaving it out means no temps at all). This is a trap that the docs should call out better I think. >>> >>> Anyway - now I just need to figure out how to do this programatically - and possibly I can write a few more useful refactorings I've also thought it was odd that were missing. >>> >>> On Mon, 3 Jun 2024, at 12:19 PM, Gabriel Cotelli wrote: >>>> I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: >>>> >>>> self doSomething: ( >>>> moveSubmenu >>>> addToggle: 'Move right' translated >>>> target: self >>>> selector: #taskbarMoveRight >>>> getStateSelector: nil >>>> enablementSelector: #canBeMovedToRight) >>>> >>>> and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. >>>> >>>> Regards, Gabriel >>>> >>>> On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >>>> >>>> I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). >>>> >>>> I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). >>>> >>>> This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). >>>> >>>> I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. >>>> >>>> e.g. >>>> >>>> moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } >>>> >>>> but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? >>>> >>>> Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: >>>> >>>> and I entered this: >>>> moveSubmenu >>>> addToggle: 'Move right' translated >>>> target: self >>>> selector: #taskbarMoveRight >>>> getStateSelector: nil >>>> enablementSelector: #canBeMovedToRight. >>>> >>>> Which matches exactly - then I want to replace the single statement with 2 statements e.g. >>>> >>>> moveSubmenu >>>> addToggle: 'Move right' translated >>>> target: self >>>> selector: #taskbarMoveRight >>>> getStateSelector: nil >>>> enablementSelector: #canBeMovedToRight. >>>> >>>> moveSubmenu >>>> addToggle: 'Move to' translated >>>> target: self >>>> selector: #taskbarMoveTo >>>> getStateSelector: nil. >>>> >>>> >>>> However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). >>>> >>>> But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. >>>> >>>> If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. >>>> >>>> This seems an odd choice - but am I missing something obvious? >>>> >>>> Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? >>>> >>>> I'm just curious on this - as it seemed time to learn this properly? >>>> >>>> Tim >> >> >> StÃ©phane Ducasse >> http://stephane.ducasse.free.fr >> 06 30 93 66 73 >> >> "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 5, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by Tim Mackinnon

Hi Stef - I saw that blog post - which looks like its essentially the help text that comes up in the rewrite tool when you press the help button. I was going to submit to add to the help text as I think a better example that pulls all the ideas together is helpful. However, I have realised the Rewrite tool is not the main pharo-project repo and comes under NewTools? How do I find out where that repo is to submit a pr? I think the help button on that tool should launch the normal doc browser with the text in the normal help format vs a seperate window? The help browser could probably use some work but at least if help is one place it all will get updated together? I was going to propose the following example which helped me a lot - if I wanted to rewrite how logging was done in the following example (and separate the Transcript line into 2 lines): appLog: anObject "Log anObject for debugging" | date | date := Date today. Transcript traceCr: 'debug: ', date printString, anObject printString. date isLeapYear ifTrue: [ ^nil ]. ^anObject I could write a match expression like this: appLog: `arg | `@temps | `.@before. Transcript traceCr: `#str , `@exp1, `@exp2. `.@after and a rewrite rule like this: appLog: `arg | `@temps | `.@before. Transcript cr; traceCr: `#str, `@exp1. `arg traceCr. `.@after I think this demonstrates most of the concepts that caught me out - namely: you need to call out the variable matching explicitly -= the .@before (list statement matching doesn't match variables - and using @temps will match 0 or more vars, so its safer to include it if matching expressions that may have any otherwise nothing will match. If you want to match a string you need to match it as a literal e.g. `#str. You may need to match 0 or more statements before and after your target expression hence `.@before and `.@after but its important to include the trailing "." on the @before. to cut off before the expression you are trying to match next. I didn't fully understand in my example is why I can't match the remaining statements after the transcript string using just `.@exp1. I seem to have to call out both expressions which I now realise (and worth documenting) is because the AST for that line is actually VariableNode(transcript) + MsgNode(MsgNode(Literal+MsgNode(exp1)), MsgNode(exp2) and so you have to call them all out to match things or match the entire sequence with @stmnt (in my case I wanted to pick them apart). I think the top tip in all of this is to use the handy source code ast inspector to help you reason on the structure. The above is probably the gist of the kinds of docs we would need, and happy to submit a pr with this in it. On Wed, 5 Jun 2024, at 6:35 AM, stephane ducasse wrote: > Hi Tim > > This is on my todo to have a serious look at the pattern matcher > Now one of the few documentation that we have in the > blog and I ported it in the blog post in /doc folder. > > S > > > > >> On 4 Jun 2024, at 01:35, Tim Mackinnon <tim(a)testit.works> wrote: >> >> Hey Gabriel - thanks for chipping in - you've made me realise that its perhaps easier to give a simpler example (for the multiple replacement)., and while trying to explain in simpler terms I managed to solve the problem myself - but it does demonstrate the docs around this need a bit more work as I didn't find it easy to reason on. >> >> I am still confused what the setting "this rule is for a method" does - so am keen for someone to shed light. >> >> Anyway for my example it seems that I have to be very specific in how I match things for example I had to use the search pattern: >> >> taskbarButtonMenu: `arg >> | `@temps | >> `.@before. >> >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> `.@after >> >> And it took me ages to realise that I have to specify the temp vars matching, and also understand that the @before and @after both needed the "." to represent lists of expressions both before and after and then the really subtle "." after @before which I guess prevents greedy matching. >> >> With all of that in place, could then have the write rule add an extra statement after the moveSubmenu (or as many as I want) >> >> e.g. >> >> taskbarButtonMenu: `arg >> | `@temps | >> `.@before. >> >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> moveSubmenu >> addToggle: 'Move to...' translated >> target: self >> selector: #taskbarMoveTo >> getStateSelector: nil. >> >> `.@after >> >> >> I'm curious now whether there is a more efficient way to do what I have done? Possibly the {} syntax could help now i've figured out the basic matching. >> >> I really find it easy to confuse that if you don't specify the @temps section that can leave out a lot of matches (it essentialy means 0 or more temps vs. leaving it out means no temps at all). This is a trap that the docs should call out better I think. >> >> Anyway - now I just need to figure out how to do this programatically - and possibly I can write a few more useful refactorings I've also thought it was odd that were missing. >> >> On Mon, 3 Jun 2024, at 12:19 PM, Gabriel Cotelli wrote: >>> I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: >>> >>> self doSomething: ( >>> moveSubmenu >>> addToggle: 'Move right' translated >>> target: self >>> selector: #taskbarMoveRight >>> getStateSelector: nil >>> enablementSelector: #canBeMovedToRight) >>> >>> and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. >>> >>> Regards, Gabriel >>> >>> On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >>>> __ >>>> I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). >>>> >>>> I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). >>>> >>>> This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). >>>> >>>> I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. >>>> >>>> e.g. >>>> >>>> moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } >>>> >>>> but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? >>>> >>>> Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: >>>> >>>> and I entered this: >>>> moveSubmenu >>>> addToggle: 'Move right' translated >>>> target: self >>>> selector: #taskbarMoveRight >>>> getStateSelector: nil >>>> enablementSelector: #canBeMovedToRight. >>>> >>>> Which matches exactly - then I want to replace the single statement with 2 statements e.g. >>>> >>>> moveSubmenu >>>> addToggle: 'Move right' translated >>>> target: self >>>> selector: #taskbarMoveRight >>>> getStateSelector: nil >>>> enablementSelector: #canBeMovedToRight. >>>> >>>> moveSubmenu >>>> addToggle: 'Move to' translated >>>> target: self >>>> selector: #taskbarMoveTo >>>> getStateSelector: nil. >>>> >>>> >>>> However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). >>>> >>>> But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. >>>> >>>> If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. >>>> >>>> This seems an odd choice - but am I missing something obvious? >>>> >>>> Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? >>>> >>>> I'm just curious on this - as it seemed time to learn this properly? >>>> >>>> Tim > > StÃ©phane Ducasse > http://stephane.ducasse.free.fr > 06 30 93 66 73 > > "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes > >

June 5, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by stephane ducasse

Hi Tim This is on my todo to have a serious look at the pattern matcher Now one of the few documentation that we have in the blog and I ported it in the blog post in /doc folder. S > On 4 Jun 2024, at 01:35, Tim Mackinnon <tim(a)testit.works> wrote: > > Hey Gabriel - thanks for chipping in - you've made me realise that its perhaps easier to give a simpler example (for the multiple replacement)., and while trying to explain in simpler terms I managed to solve the problem myself - but it does demonstrate the docs around this need a bit more work as I didn't find it easy to reason on. > > I am still confused what the setting "this rule is for a method" does - so am keen for someone to shed light. > > Anyway for my example it seems that I have to be very specific in how I match things for example I had to use the search pattern: > > taskbarButtonMenu: `arg > | `@temps | > `.@before. > > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > `.@after > > And it took me ages to realise that I have to specify the temp vars matching, and also understand that the @before and @after both needed the "." to represent lists of expressions both before and after and then the really subtle "." after @before which I guess prevents greedy matching. > > With all of that in place, could then have the write rule add an extra statement after the moveSubmenu (or as many as I want) > > e.g. > > taskbarButtonMenu: `arg > | `@temps | > `.@before. > > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > moveSubmenu > addToggle: 'Move to...' translated > target: self > selector: #taskbarMoveTo > getStateSelector: nil. > > `.@after > > > I'm curious now whether there is a more efficient way to do what I have done? Possibly the {} syntax could help now i've figured out the basic matching. > > I really find it easy to confuse that if you don't specify the @temps section that can leave out a lot of matches (it essentialy means 0 or more temps vs. leaving it out means no temps at all). This is a trap that the docs should call out better I think. > > Anyway - now I just need to figure out how to do this programatically - and possibly I can write a few more useful refactorings I've also thought it was odd that were missing. > > On Mon, 3 Jun 2024, at 12:19 PM, Gabriel Cotelli wrote: >> I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: >> >> self doSomething: ( >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight) >> >> and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. >> >> Regards, Gabriel >> >> On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >> >> I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). >> >> I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). >> >> This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). >> >> I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. >> >> e.g. >> >> moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } >> >> but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? >> >> Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: >> >> and I entered this: >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> Which matches exactly - then I want to replace the single statement with 2 statements e.g. >> >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> moveSubmenu >> addToggle: 'Move to' translated >> target: self >> selector: #taskbarMoveTo >> getStateSelector: nil. >> >> >> However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). >> >> But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. >> >> If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. >> >> This seems an odd choice - but am I missing something obvious? >> >> Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? >> >> I'm just curious on this - as it seemed time to learn this properly? >> >> Tim StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 5, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by stephane ducasse

Hi Tim This is on my todo to have a serious look at the pattern matcher Now one of the few documentation that we have in the blog and I ported it in the blog post in /doc folder. S > On 4 Jun 2024, at 01:35, Tim Mackinnon <tim(a)testit.works> wrote: > > Hey Gabriel - thanks for chipping in - you've made me realise that its perhaps easier to give a simpler example (for the multiple replacement)., and while trying to explain in simpler terms I managed to solve the problem myself - but it does demonstrate the docs around this need a bit more work as I didn't find it easy to reason on. > > I am still confused what the setting "this rule is for a method" does - so am keen for someone to shed light. > > Anyway for my example it seems that I have to be very specific in how I match things for example I had to use the search pattern: > > taskbarButtonMenu: `arg > | `@temps | > `.@before. > > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > `.@after > > And it took me ages to realise that I have to specify the temp vars matching, and also understand that the @before and @after both needed the "." to represent lists of expressions both before and after and then the really subtle "." after @before which I guess prevents greedy matching. > > With all of that in place, could then have the write rule add an extra statement after the moveSubmenu (or as many as I want) > > e.g. > > taskbarButtonMenu: `arg > | `@temps | > `.@before. > > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > moveSubmenu > addToggle: 'Move to...' translated > target: self > selector: #taskbarMoveTo > getStateSelector: nil. > > `.@after > > > I'm curious now whether there is a more efficient way to do what I have done? Possibly the {} syntax could help now i've figured out the basic matching. > > I really find it easy to confuse that if you don't specify the @temps section that can leave out a lot of matches (it essentialy means 0 or more temps vs. leaving it out means no temps at all). This is a trap that the docs should call out better I think. > > Anyway - now I just need to figure out how to do this programatically - and possibly I can write a few more useful refactorings I've also thought it was odd that were missing. > > On Mon, 3 Jun 2024, at 12:19 PM, Gabriel Cotelli wrote: >> I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: >> >> self doSomething: ( >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight) >> >> and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. >> >> Regards, Gabriel >> >> On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >> >> I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). >> >> I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). >> >> This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). >> >> I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. >> >> e.g. >> >> moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } >> >> but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? >> >> Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: >> >> and I entered this: >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> Which matches exactly - then I want to replace the single statement with 2 statements e.g. >> >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> moveSubmenu >> addToggle: 'Move to' translated >> target: self >> selector: #taskbarMoveTo >> getStateSelector: nil. >> >> >> However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). >> >> But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. >> >> If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. >> >> This seems an odd choice - but am I missing something obvious? >> >> Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? >> >> I'm just curious on this - as it seemed time to learn this properly? >> >> Tim StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 5, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by Tim Mackinnon

Hey Gabriel - thanks for chipping in - you've made me realise that its perhaps easier to give a simpler example (for the multiple replacement)., and while trying to explain in simpler terms I managed to solve the problem myself - but it does demonstrate the docs around this need a bit more work as I didn't find it easy to reason on. I am still confused what the setting "this rule is for a method" does - so am keen for someone to shed light. Anyway for my example it seems that I have to be very specific in how I match things for example I had to use the search pattern: taskbarButtonMenu: `arg | `@temps | `.@before. moveSubmenu addToggle: 'Move right' translated target: self selector: #taskbarMoveRight getStateSelector: nil enablementSelector: #canBeMovedToRight. `.@after And it took me ages to realise that I have to specify the temp vars matching, and also understand that the @before and @after both needed the "." to represent lists of expressions both before and after and then the really subtle "." after @before which I guess prevents greedy matching. With all of that in place, could then have the write rule add an extra statement after the moveSubmenu (or as many as I want) e.g. taskbarButtonMenu: `arg | `@temps | `.@before. moveSubmenu addToggle: 'Move right' translated target: self selector: #taskbarMoveRight getStateSelector: nil enablementSelector: #canBeMovedToRight. moveSubmenu addToggle: 'Move to...' translated target: self selector: #taskbarMoveTo getStateSelector: nil. `.@after I'm curious now whether there is a more efficient way to do what I have done? Possibly the {} syntax could help now i've figured out the basic matching. I really find it easy to confuse that if you don't specify the @temps section that can leave out a lot of matches (it essentialy means 0 or more temps vs. leaving it out means no temps at all). This is a trap that the docs should call out better I think. Anyway - now I just need to figure out how to do this programatically - and possibly I can write a few more useful refactorings I've also thought it was odd that were missing. On Mon, 3 Jun 2024, at 12:19 PM, Gabriel Cotelli wrote: > I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: > > self doSomething: ( > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight) > > and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. > > Regards, Gabriel > > On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >> __ >> I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). >> >> I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). >> >> This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). >> >> I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. >> >> e.g. >> >> moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } >> >> but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? >> >> Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: >> >> and I entered this: >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> Which matches exactly - then I want to replace the single statement with 2 statements e.g. >> >> moveSubmenu >> addToggle: 'Move right' translated >> target: self >> selector: #taskbarMoveRight >> getStateSelector: nil >> enablementSelector: #canBeMovedToRight. >> >> moveSubmenu >> addToggle: 'Move to' translated >> target: self >> selector: #taskbarMoveTo >> getStateSelector: nil. >> >> >> However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). >> >> But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. >> >> If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. >> >> This seems an odd choice - but am I missing something obvious? >> >> Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? >> >> I'm just curious on this - as it seemed time to learn this properly? >> >> Tim

June 3, 2024

Re: Can the rewrite tool replace a single statement with multiple statements?

by Gabriel Cotelli

I've used the rewrite tool extensively but never tried to create more than one expression from the original one. I think that this probably won't work because in your example the code is just statements, but the matching expression can be in code like this: self doSomething: ( moveSubmenu addToggle: 'Move right' translated target: self selector: #taskbarMoveRight getStateSelector: nil enablementSelector: #canBeMovedToRight) and in this case, you cannot replace this with two statements. So I'm tempted to say that what you want is not supported. Regards, Gabriel On Mon, Jun 3, 2024 at 5:44â¯AM Tim Mackinnon <tim(a)testit.works> wrote: > I've never really learned the rewrite tool but while waiting for changes > to propagate through P12, I thought maybe I could use it to safely patch > changes until there are official updates (and it was a good chance to get > more familiar with it). > > I thought there was more documentation available for how it works > (interested in any old articles - but the online help in P12 is ok - > although it doesn't display in markdown - in fact give there is markdown > support now - I am a bit confused what should appear more readable these > days - and how to fix/contribute changes to help). > > This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and > add my fix to move tabs to any position). > > I started writing the following search expression - the find the last call > to #addToggle:... method call that has a literal referencing > #canBeMovedRight - as I want to add my new menu item after it. > > e.g. > > moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = > #canBeMovedToRight] } > > but I'm not sure my {} filter block is ever called, as the above matches > both addToggle:... calls, and not the one I'm after? I don't understand > this - any have ideas on the proper syntax for generic matching? > > Anyway - I decided that in my case there is only one - so I can match it > explicitly and so I search on the full call: > > and I entered this: > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > Which matches exactly - then I want to replace the single statement with 2 > statements e.g. > > moveSubmenu > addToggle: 'Move right' translated > target: self > selector: #taskbarMoveRight > getStateSelector: nil > enablementSelector: #canBeMovedToRight. > > moveSubmenu > addToggle: 'Move to' translated > target: self > selector: #taskbarMoveTo > getStateSelector: nil. > > > However this gives an error in the rewrite tool (actually it results in a > zero change, that then gives an error for EpMonitor). > > But my question is - why can't you rewrite a single statement to multiple > ones? I have noticed that in this case I can cascade the message send - and > this then rewrites properly - so it looks like it wants to replace a single > node with a single node. > > If I couldn't cascade - It seems I would have to rewrite with something > like: [ "put multiple expressions here" ] value. > > This seems an odd choice - but am I missing something obvious? > > Do you have to do what I'm proposing in 2 steps - rewrite to a block (or > some equivalent) and then a 2nd rewrite to simplify the expression (like a > refactor step) - so its more provably correct? > > I'm just curious on this - as it seemed time to learn this properly? > > Tim >

June 3, 2024

Can the rewrite tool replace a single statement with multiple statements?

by Tim Mackinnon

I've never really learned the rewrite tool but while waiting for changes to propagate through P12, I thought maybe I could use it to safely patch changes until there are official updates (and it was a good chance to get more familiar with it). I thought there was more documentation available for how it works (interested in any old articles - but the online help in P12 is ok - although it doesn't display in markdown - in fact give there is markdown support now - I am a bit confused what should appear more readable these days - and how to fix/contribute changes to help). This aside - I wanted to patch the SystemWindow>>taskbarButtonMenu: (and add my fix to move tabs to any position). I started writing the following search expression - the find the last call to #addToggle:... method call that has a literal referencing #canBeMovedRight - as I want to add my new menu item after it. e.g. moveSubmenu `@addToggle: `{ :node | node isLiteralNode and: [ node value = #canBeMovedToRight] } but I'm not sure my {} filter block is ever called, as the above matches both addToggle:... calls, and not the one I'm after? I don't understand this - any have ideas on the proper syntax for generic matching? Anyway - I decided that in my case there is only one - so I can match it explicitly and so I search on the full call: and I entered this: moveSubmenu addToggle: 'Move right' translated target: self selector: #taskbarMoveRight getStateSelector: nil enablementSelector: #canBeMovedToRight. Which matches exactly - then I want to replace the single statement with 2 statements e.g. moveSubmenu addToggle: 'Move right' translated target: self selector: #taskbarMoveRight getStateSelector: nil enablementSelector: #canBeMovedToRight. moveSubmenu addToggle: 'Move to' translated target: self selector: #taskbarMoveTo getStateSelector: nil. However this gives an error in the rewrite tool (actually it results in a zero change, that then gives an error for EpMonitor). But my question is - why can't you rewrite a single statement to multiple ones? I have noticed that in this case I can cascade the message send - and this then rewrites properly - so it looks like it wants to replace a single node with a single node. If I couldn't cascade - It seems I would have to rewrite with something like: [ "put multiple expressions here" ] value. This seems an odd choice - but am I missing something obvious? Do you have to do what I'm proposing in 2 steps - rewrite to a block (or some equivalent) and then a 2nd rewrite to simplify the expression (like a refactor step) - so its more provably correct? I'm just curious on this - as it seemed time to learn this properly? Tim

June 3, 2024

Re: Recent Pharo's seem to lose first keystroke when you swap to them to paste?

by Tim Mackinnon

Hi Russ (and others) - there is an update in the ticket, it looks like the meta key's state isn't been properly tracked due to how cocoa (macos) propogates events to the sdlc library (or something like that). Pablo has spotted something he thinks might fix it - so fingers crossed on this. And thanks for everyone piping up - it makes the system better. Tim see. https://github.com/pharo-project/pharo/issues/16700 On Sat, 1 Jun 2024, at 3:45 PM, Russ Whaley wrote: > I've paid specific attention to my pastes when an (outside of Pharo) window is active... such as a web browser and I'm copying a URL, etc. I find I must select the Pharo window, THEN select the widget for the paste. If I just select the widget, it appears to be giving the window focus only, and if I then selecting the widget, the paste works fine. > > Hope this helps. > Russ > > On Wed, May 29, 2024 at 9:03â¯AM Tim Mackinnon <tim(a)testit.works> wrote: >> __ >> It's very stange - and it seems many people experience it from comments on the submitted but - although it seems an OSX problem (so far) and one that has happened possibly in P9 or P10. >> >> Perhaps we can narrow it down enough to figure out the cause (and it occurs to me to see if GT has the same issue). >> >> Tim >> >> On Wed, 29 May 2024, at 12:29 PM, stephane ducasse wrote: >>> Tx >>> I did not know that pasting twice was workingâ¦ >>> How a strange bug. >>> >>> S >>> >>>> On 27 May 2024, at 22:32, Tim Mackinnon <tim(a)testit.works> wrote: >>>> >>>> I've submitted the issue https://github.com/pharo-project/pharo/issues/16700 and noted other comments >>>> >>>> On Mon, 27 May 2024, at 5:02 PM, Tim Mackinnon wrote: >>>>> >>>>> It would be handy to know if itâs just macOS - but Iâll write up an issue this evening. >>>>> >>>>> Thanks for confirming. >>>>> >>>>> Tim >>>>> >>>>> >>>>>> On 27 May 2024, at 16:01, Koen De Hondt <koen(a)all-objects-all-the-time.st> wrote: >>>>>> ï»¿I have suffered from copy & paste issues like this for a long time, in P11 and P12. >>>>>> I always have trouble with copying text from an image. I am already used to pressing Command-C twice ð >>>>>> >>>>>> I use a M2 Mac with MacOS 14.5, but it happened on earlier versions of MacOS as well. >>>>>> >>>>>> Koen >>>>>> >>>>>>> On 27 May 2024, at 15:48, Noury Bouraqadi <bouraqadi(a)gmail.com> wrote: >>>>>>> >>>>>>> Same bug on P11 with MacOS. >>>>>>> >>>>>>> On May 27 2024, at 3:44 pm, Tim Mackinnon <tim(a)testit.works> wrote: >>>>>>>> For a while I've noticed that every time I try and copy some text from a web browser or email and then paste it into pharo the first time it fails - and then thinking I didn't copy, I try it again and then realise that I have to paste twice - as the first one doesn't work. >>>>>>>> >>>>>>>> This didn't seem to do this in Pharo 9 (maybe 10) but its definitely the case in Pharo 11 and Pharo 12. This is on OSX (14.x) so I'm wondering if the same occurs in Windows or Linux or if its just my laptop? Its quite annoying and I'm wondering if it might be a side effect from the gradual UI changes in Pharo? >>>>>>>> >>>>>>>> I've looked in the issues tracker - but its a tricky thing to search for - and so I wanted to canvas experience and then will report it. >>>>>>>> >>>>>>>> Tim >>> >>> StÃ©phane Ducasse >>> http://stephane.ducasse.free.fr >>> 06 30 93 66 73 >>> >>> "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes >>> >>> >> > > > -- > Russ Whaley > whaley.russ(a)gmail.com

June 1, 2024

Re: Recent Pharo's seem to lose first keystroke when you swap to them to paste?

by Russ Whaley

I've paid specific attention to my pastes when an (outside of Pharo) window is active... such as a web browser and I'm copying a URL, etc. I find I must select the Pharo window, THEN select the widget for the paste. If I just select the widget, it appears to be giving the window focus only, and if I then selecting the widget, the paste works fine. Hope this helps. Russ On Wed, May 29, 2024 at 9:03â¯AM Tim Mackinnon <tim(a)testit.works> wrote: > It's very stange - and it seems many people experience it from comments on > the submitted but - although it seems an OSX problem (so far) and one that > has happened possibly in P9 or P10. > > Perhaps we can narrow it down enough to figure out the cause (and it > occurs to me to see if GT has the same issue). > > Tim > > On Wed, 29 May 2024, at 12:29 PM, stephane ducasse wrote: > > Tx > I did not know that pasting twice was workingâ¦ > How a strange bug. > > S > > On 27 May 2024, at 22:32, Tim Mackinnon <tim(a)testit.works> wrote: > > I've submitted the issue > https://github.com/pharo-project/pharo/issues/16700 and noted other > comments > > On Mon, 27 May 2024, at 5:02 PM, Tim Mackinnon wrote: > > > It would be handy to know if itâs just macOS - but Iâll write up an issue > this evening. > > Thanks for confirming. > > Tim > > > On 27 May 2024, at 16:01, Koen De Hondt <koen(a)all-objects-all-the-time.st> > wrote: > > ï»¿I have suffered from copy & paste issues like this for a long time, in > P11 and P12. > I always have trouble with copying text from an image. I am already used > to pressing Command-C twice ð > > I use a M2 Mac with MacOS 14.5, but it happened on earlier versions of > MacOS as well. > > Koen > > On 27 May 2024, at 15:48, Noury Bouraqadi <bouraqadi(a)gmail.com> wrote: > > Same bug on P11 with MacOS. > > On May 27 2024, at 3:44 pm, Tim Mackinnon <tim(a)testit.works> wrote: > > For a while I've noticed that every time I try and copy some text from a > web browser or email and then paste it into pharo the first time it fails - > and then thinking I didn't copy, I try it again and then realise that I > have to paste twice - as the first one doesn't work. > > This didn't seem to do this in Pharo 9 (maybe 10) but its definitely the > case in Pharo 11 and Pharo 12. This is on OSX (14.x) so I'm wondering if > the same occurs in Windows or Linux or if its just my laptop? Its quite > annoying and I'm wondering if it might be a side effect from the gradual UI > changes in Pharo? > > I've looked in the issues tracker - but its a tricky thing to search for - > and so I wanted to canvas experience and then will report it. > > Tim > > > StÃ©phane Ducasse > http://stephane.ducasse.free.fr > 06 30 93 66 73 > > "If you knew today was your last day on earth, what would you do > differently? ....ESPECIALLY if, by doing something different, today > might not be your last day on earth.â Calvin & Hobbes > > > > -- Russ Whaley whaley.russ(a)gmail.com

June 1, 2024

12 Month engineer position at Toulouse in Pharo

by stephane ducasse

Hi people Can you distribute this job position around you? Yannick Chevalier has a position to develop a tool to detect abnomalies DeÌveloppement de lâoutil iCANt pour la deÌtection dâanomalies <mailto:yannick.chevalier@gmail.com> Who: yannick.chevalier(a)gmail.com <mailto:yannick.chevalier@gmail.com> Where: Toulouse Salary: 45 K Euros a more experienced dev could be hired for a shorter period and get a better salary. S. StÃ©phane Ducasse http://stephane.ducasse.free.fr 06 30 93 66 73 "If you knew today was your last day on earth, what would you do differently? ....ESPECIALLY if, by doing something different, today might not be your last day on earth.â Calvin & Hobbes

June 1, 2024