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

Re: please unsubscribe me

by Marcus Denker

Done ! Marcus > On 11 Mar 2024, at 16:27, Rui Viana Pereira <ruivianapereira(a)gmail.com> wrote: > > Pereira

March 11, 2024

Re: [Pharo-dev] Re: "Introducing Atlas" â New blog post on all: objects all: theTime

by Aik-Siong Koh

Atlas is indeed ambitious, and I am glad you want to push the boundaries. Over the decades, GUIs have introduced a bewildering array of widgets and designs that I think we are in a mess again. I feel we should restart again with a blank window, allow select and rightclick for context menu, and proceed for there to introduce the base minimum for a functional GUI. I believe "select and rightclick context menu" can solve almost every GUI need simply. I also think the single column menu can be updated to a two-column menu which is more compact and balanced. The cursor can be centered between the two columns instead of being at the top left corner. The cursor has to travel less for any selection. https://askoh.com/index.html#TwoColumnMenu All the best, Aik-Siong Koh On Mon, 11 Mar 2024 09:58:40 +0000, "Tim Mackinnon" <tim(a)testit.works> wrote: Â As browsers are a passion in the Smalltalk world it will be great to read your thoughts - as its certainly a hot potato, and we don't seem to have quite cracked it so far. Â I recall the presentations on Calypso from Pharo days (might have been recorded for review if you haven't seen them). I recall being won over at the time (and I was hesitant) Â - there was lots of flexibility that had been thought about, and many useful and tricky browsing patterns were covered - but over time I think its proved tricky to work with. In Pharo 11, the browsers don't seem to work as well as they should (lots of funny focus issues and loss of context that I don't recall in previous version - which I think is more down to understanding how it was supposed to work than technical flaws). Â It's definitely worth generating conversation and getting some consensus otherwise it will be a rise and fall scenario all over again. This said, continuing to find a good model that is both flexible and simple is useful. Â Tim Â p.s. On thing I recall from those Calypso presentations was that the model should have let us design browsers where weÂ have different navigation models Â (e.g. you in theory you could design something where a class has a path to methods which are both instance and class so you don't have to have a mode to swap between them - something I find distracting when designing the interface of a class and trying to figure out how you instantiate/initialize it and you want to jump between the 2 view - I just want to see all all the methods in a list, differentiated in some way vs. hiding them). Â On Sun, 10 Mar 2024, at 5:24 PM, Koen De Hondt wrote: Dear Pharo users and developers, Â Some people already know that I am working on a browser for Pharo. With this announcement, I make it official ð. InÂ my latest blog post, I introduce Atlas as an ambitious successor of Calypso. It is the first post in a series. Â Happy reading! Â Ciao, Koen Â

March 11, 2024

please unsubscribe me

by Rui Viana Pereira

Thank you *Rui Viana Pereira*

March 11, 2024

Re: [Pharo-dev] "Introducing Atlas" â New blog post on all: objects all: theTime

by Tim Mackinnon

As browsers are a passion in the Smalltalk world it will be great to read your thoughts - as its certainly a hot potato, and we don't seem to have quite cracked it so far. I recall the presentations on Calypso from Pharo days (might have been recorded for review if you haven't seen them). I recall being won over at the time (and I was hesitant) - there was lots of flexibility that had been thought about, and many useful and tricky browsing patterns were covered - but over time I think its proved tricky to work with. In Pharo 11, the browsers don't seem to work as well as they should (lots of funny focus issues and loss of context that I don't recall in previous version - which I think is more down to understanding how it was supposed to work than technical flaws). It's definitely worth generating conversation and getting some consensus otherwise it will be a rise and fall scenario all over again. This said, continuing to find a good model that is both flexible and simple is useful. Tim p.s. On thing I recall from those Calypso presentations was that the model should have let us design browsers where we have different navigation models (e.g. you in theory you could design something where a class has a path to methods which are both instance and class so you don't have to have a mode to swap between them - something I find distracting when designing the interface of a class and trying to figure out how you instantiate/initialize it and you want to jump between the 2 view - I just want to see all all the methods in a list, differentiated in some way vs. hiding them). On Sun, 10 Mar 2024, at 5:24 PM, Koen De Hondt wrote: > Dear Pharo users and developers, > > Some people already know that I am working on a browser for Pharo. With this announcement, I make it official ð. > In my latest blog post <https://all-objects-all-the-time.st/#/blog/posts/6>, I introduce Atlas as an ambitious successor of Calypso. It is the first post in a series. > > Happy reading! > > Ciao, > Koen

March 11, 2024

This week (10/2024) on the Pharo Issue Tracker

by Marcus Denker

# Pharo 10 - Fix for: UFFI/TFFI call for WebBrowser with specific URL length adds strange characters to the URL #15980 #16253 https://github.com/pharo-project/pharo/pull/16253 # Pharo 11 - Fix for: UFFI/TFFI call for WebBrowser with specific URL length adds strange characters to the URL #15980 #16284 https://github.com/pharo-project/pharo/pull/16284 # Fixes - do not spam selectPage: #1528 https://github.com/pharo-spec/Spec/pull/1528 - Restore original message String #16274 https://github.com/pharo-project/pharo/pull/16274 - Fix Method name editor dialog up/down buttons #16273 https://github.com/pharo-project/pharo/pull/16273 - Fix Bug: OrderedDictionary>>#at:at:put: not updating ordered keys #16260 https://github.com/pharo-project/pharo/pull/16260 - Minor improvement in TF test class #16257 https://github.com/pharo-project/pharo/pull/16257 - Fix #magnifyBy:smoothing: on OSSDL2ExternalForm #16225 https://github.com/pharo-project/pharo/pull/16225 - fix a failing test #714 https://github.com/pharo-spec/NewTools/pull/714 # Improvements - Make #reportTestCase:runBlock: on HDTestReport print the test run time on the progress log stream #16266 https://github.com/pharo-project/pharo/pull/16266 # Refactoring Engine - Fix: rename method shortcut #16251 https://github.com/pharo-project/pharo/pull/16251 - Remove unused instance variable in #ReRenameMethodRefactoring #16279 https://github.com/pharo-project/pharo/pull/16279 - Fix: rename method preconditions fail with DNU #16270 https://github.com/pharo-project/pharo/pull/16270 - Test: rename method driver #16269 https://github.com/pharo-project/pharo/pull/16269 - Cleanup: rename method refactoring driver to new architecture with breaking choices #16258 https://github.com/pharo-project/pharo/pull/16258 # Cleanups - Refactor various âworldMenuâ registration methods to use #iconName: instead of #icon: #711 https://github.com/pharo-spec/NewTools/pull/711 - Updated baseline of Dependency analyzer #16278 https://github.com/pharo-project/pharo/pull/16278 - 16246-CI-Failing-test-SystemDependenciesTesttestExternalIDEDependencies #16264 https://github.com/pharo-project/pharo/pull/16264

March 11, 2024

Re: Finding references to non existent classes in Pharo

by Tim Mackinnon

So it seems my expectation wasn't unreasonable (I expected Pharo to do the same as Smalltalk/X - given you can leave an undeclared reference). Perhaps Marcus might be able to comment on this, and I'll go and see if its been written up as an existing bug and if its also present in Pharo 12 now that its in stabilisation mode. Appreciate those that chipped in - these are features that we need to ensure aren't lost (of course I recall the days when you couldn't save anything broken - and that was a PITA). Tim On Sun, 10 Mar 2024, at 8:56 PM, Richard O'Keefe wrote: > I created a method referring to a non-existent "Snark" class. > To my astonishment, "Code Search|References to it" quietly failed to > do anything visible. > It didn't find it, and it didn't say that it couldn't find it. > However, "Code Search|Method source with it" DID find the reference. > Pharo 9 and Pharo 11 both do this, although the on-screen interface is > different. > > In Squeak, "References to it" worked fine, BUT when I created the > method I wasn't > offered the option of leaving "Snark" undeclared; the closest was > "Declare global". > > Smalltalk/X offered me the option of "Continue" with an undeclared variable, > and then "References to it" worked perfectly. > > > > On Mon, 11 Mar 2024 at 01:43, Tim Mackinnon <Tim(a)testit.works> wrote: >> >> Hi - I was convinced in earlier Pharoâs, if you had a code reference to a non existent class you could find it by searching for references to its symbol name eg #MyMissingClass allReferences (or find references in the UI). This doesnât seem to work in Pharo 11? I loaded a package with a missing class, and when running something it complained about the missing class (it was an announcement), but I couldn't find an easy way to find it in my code to correct it? I ended up creating the fake class to then find references to it (as I then had a class), which seems way over the top? >> >> I haven't had a chance to try this in Pharo 12, but shouldn't what I have done work? Or is there some new way to do this? I asked on Discord users, but didn't get a reply other than it rang a bell. >> >> I know there has been a lot of work in the area of how things are represented and I wonder if something has got broken by mistake? >> >> Tim

March 11, 2024

Re: "Introducing Atlas" â New blog post on all: objects all: theTime

by Sebastian Jordan Montano

Nice!! Sebastian > De: "Russ Whaley" <whaley.russ(a)gmail.com> > Ã: "Any question about pharo is welcome" <pharo-users(a)lists.pharo.org> > EnvoyÃ©: Dimanche 10 Mars 2024 19:14:40 > Objet: [Pharo-users] Re: "Introducing Atlas" â New blog post on all: objects > all: theTime > Great reading Koen. I'm looking forward to your efforts :) > On Sun, Mar 10, 2024 at 1:25 PM Koen De Hondt < [ > mailto:koen@all-objects-all-the-time.st | koen(a)all-objects-all-the-time.st ] > > wrote: >> Dear Pharo users and developers, >> Some people already know that I am working on a browser for Pharo. With this >> announcement, I make it official ð. >> In [ https://all-objects-all-the-time.st/#/blog/posts/6 | my latest blog post ] >> , I introduce Atlas as an ambitious successor of Calypso. It is the first post >> in a series. >> Happy reading! >> Ciao, >> Koen > -- > Russ Whaley > [ mailto:whaley.russ@gmail.com | whaley.russ(a)gmail.com ]

March 11, 2024

Re: Finding references to non existent classes in Pharo

by Richard O'Keefe

I created a method referring to a non-existent "Snark" class. To my astonishment, "Code Search|References to it" quietly failed to do anything visible. It didn't find it, and it didn't say that it couldn't find it. However, "Code Search|Method source with it" DID find the reference. Pharo 9 and Pharo 11 both do this, although the on-screen interface is different. In Squeak, "References to it" worked fine, BUT when I created the method I wasn't offered the option of leaving "Snark" undeclared; the closest was "Declare global". Smalltalk/X offered me the option of "Continue" with an undeclared variable, and then "References to it" worked perfectly. On Mon, 11 Mar 2024 at 01:43, Tim Mackinnon <Tim(a)testit.works> wrote: > > Hi - I was convinced in earlier Pharoâs, if you had a code reference to a non existent class you could find it by searching for references to its symbol name eg #MyMissingClass allReferences (or find references in the UI). This doesnât seem to work in Pharo 11? I loaded a package with a missing class, and when running something it complained about the missing class (it was an announcement), but I couldn't find an easy way to find it in my code to correct it? I ended up creating the fake class to then find references to it (as I then had a class), which seems way over the top? > > I haven't had a chance to try this in Pharo 12, but shouldn't what I have done work? Or is there some new way to do this? I asked on Discord users, but didn't get a reply other than it rang a bell. > > I know there has been a lot of work in the area of how things are represented and I wonder if something has got broken by mistake? > > Tim

March 10, 2024

Re: Finding references to non existent classes in Pharo

by Richard Sargent

*If* you can locate one example, you may be able to manually examine the compiled method itself to see what distinguishes it in terms of the reference to a missing object. e.g. what's in the literal pool for a method which references a non-existent class? You could create one such example explicitly and then use what you learn from that to track down the real ones. VW has a MethodCollector class which allows a variety of selection options for methods, including an arbitrary block of clode. I don't know if Pharo has something comparable. Worst case scenario, you use a poor mon's approach. It is easy enough to enumerate all classes and traverse their class and instance methods. That would pretty much be what you would want from a MethodCollector anyway when using an arbitrary block for selection. VW also has an Undeclared namespace. I know Pharo doesn't have namespaces, per se, but it may still track undeclared references in some manner. On Sun, Mar 10, 2024 at 11:04â¯AM Russ Whaley <whaley.russ(a)gmail.com> wrote: > Tim, > Another anomaly I found with the above, even with my code that does not > reference the missing classes - but is contained within the same package of > classes that do reference the missing classes (clear?)... my application > crashes with an error that the reference could not be found. This may be > something introduced after build #913 - where I wasn't getting these > errors... and by build #1258 they appeared (may have also appeared earlier, > but I don't recall). > > I had hoped to use the allReferences you listed above, but I had to run > through all my classes and methods to find missing classes - the methods > are highlighted, but lots of methods are highlighted for various reasons. > When I found the offending methods - I had to "comment out" the references > to get the method to then save... most of these were maintenance methods > that are only still around for reference/documentation... I need to find > another way :) > > It sure would be good to be able to see ANY missing references in one call > so I could run that periodically to keep things clean. > > On Sun, Mar 10, 2024 at 8:43â¯AM Tim Mackinnon <Tim(a)testit.works> wrote: > >> Hi - I was convinced in earlier Pharoâs, if you had a code reference to a >> non existent class you could find it by searching for references to its >> symbol name eg #MyMissingClass allReferences (or find references in the >> UI). This doesnât seem to work in Pharo 11? I loaded a package with a >> missing class, and when running something it complained about the missing >> class (it was an announcement), but I couldn't find an easy way to find it >> in my code to correct it? I ended up creating the fake class to then find >> references to it (as I then had a class), which seems way over the top? >> >> I haven't had a chance to try this in Pharo 12, but shouldn't what I have >> done work? Or is there some new way to do this? I asked on Discord users, >> but didn't get a reply other than it rang a bell. >> >> I know there has been a lot of work in the area of how things are >> represented and I wonder if something has got broken by mistake? >> >> Tim >> > > > -- > Russ Whaley > whaley.russ(a)gmail.com >

March 10, 2024

Re: "Introducing Atlas" â New blog post on all: objects all: theTime

by Russ Whaley

Great reading Koen. I'm looking forward to your efforts :) On Sun, Mar 10, 2024 at 1:25â¯PM Koen De Hondt < koen(a)all-objects-all-the-time.st> wrote: > Dear Pharo users and developers, > > Some people already know that I am working on a browser for Pharo. With > this announcement, I make it official ð. > In my latest blog post > <https://all-objects-all-the-time.st/#/blog/posts/6>, I introduce Atlas > as an ambitious successor of Calypso. It is the first post in a series. > > Happy reading! > > Ciao, > Koen > -- Russ Whaley whaley.russ(a)gmail.com

March 10, 2024