[Cuis-dev] Debug action for buttons

[Juan Vuletich]

Hi Facundo,

On 11/13/2023 8:16 PM, Facundo Javier Gelatti via Cuis-dev wrote:
> Hello folks!
> I include some code to add an option to debug the action of 
> MenuItemMorphs and PluggableButtonMorphs. See demo video 
> <https://drive.google.com/file/d/1cVMwcMS56EgpYOiQg5MawVF6W_RaEzX3/view> 
> and screenshot below:
>
> image.png
> Cheers!
> Facu.
>
> PS: I leave additional details about the implementation, that could be 
> useful during code-review 👇
>
> ------------------------------------------------------------------------
> 📝 The changes include support for MenuItemMorphs and 
> PluggableButtonMorphs (for convenience, in this email, I'll refer to 
> both of them as "buttons").
> It'd be nice to have a more general implementation¹, but I started 
> only with those because I usually debug the actions of those kinds of 
> morph types (manually, until now :P).
>
> The way it's implemented is by opening a debugger on the morph's 
> action, and making it go forward² until it reaches the appropriate 
> method on the morph's target/model (see #debugAction).
> Notice that it also shows a confirmation message in case you try to 
> debug —and therefore run— the action of a disabled button³.
>
> To execute the button's action, I reused some code from 
> PluggableButtonMorph>>#performAction and MenuItemMorph>>#invokeWithEvent:.
> To do that, I extracted a couple of methods; I include here the 
> reasons that motivated the way I chose to do it:
>
>    1. To be able to skip the check to see if the button was enabled or
>       not, I had to extract the code that runs the action without
>       making any precondition checks.
>    2. In the case of MenuItemMorph>>#invokeWithEvent:, the
>       #numArgs message is sent to the corresponding selector. During
>       manual testing, I discovered that this caused the
>       String>>#numArgs method to run in the debugger (stepping through
>       it as part of the mechanism mentioned before), and it was /very
>       slow/. So, to avoid running that part of the code in the
>       debugger, I extracted a method that does all the pre-processing
>       and returns a block that's ready to run the action
>       (MenuItemMorph>>#actionBlockForEvent:). Then, I run /that/ block
>       on the debugger.
>    3. The issue mentioned in point (2) did not happen for
>       PluggableButtonMorph>>#performAction. Nonetheless, I made a
>       similar implementation, just for consistency's sake
>       (PluggableButtonMorph>>#actionBlock). The different name has to
>       do with the different vocabulary used for the
>       PluggableButtonMorph vs. the MenuItemMorph, and with the fact
>       that the MenuItemMorph needs an event parameter.
>    4. As the MenuItemMorph needs an event, we have to simulate it. I
>       wasn't sure what was the best way to do it, so I'm currently
>       passing nil. While trying it out, I noticed that all of the menu
>       items I tested the option with were /not/ using the event
>       object. At any rate, if there's some menu option that uses it,
>       it'll fail with a "does not understand" error.
>
> Additionally, as a refactor, to be able to open the debugger on a 
> block without doing anything else, I extracted the method Debugger 
> class>>#openDebugging:label: from Debugger 
> class>>#openDebugging:to:label:.
> _____
> ¹ e.g. detecting the events supported by any morph and letting you 
> simulate and debug the one of your choosing.
> ² i.e. "step into", by sending the #send message.
> ³ Since the action is not run by a normal interaction with the button, 
> and if it's disabled maybe running the action could leave the system 
> in an unexpected state. It does, though, let you debug the action if 
> you really want to do it.

This is very nice! Thank you.

It is now at GitHub.

Cheers,

-- 
Juan Vuletich
cuis.st
github.com/jvuletich
researchgate.net/profile/Juan-Vuletich
independent.academia.edu/JuanVuletich
patents.justia.com/inventor/juan-manuel-vuletich
linkedin.com/in/juan-vuletich-75611b3
twitter.com/JuanVuletich

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20231121/26c5a546/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 92698 bytes
Desc: not available
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20231121/26c5a546/attachment-0001.png>