Unable to access constructor value in interactive mode

[4e6]

Issue

In the interactive mode, when trying to create a node with a type getter, it returns an error that the method is not available. The issue only affects the nodes created after the project is started.

How to reproduce

• Open the Unnamed project in the IDE Unnamed.zip

  from Standard.Base import all
  from Standard.Table import all
  import Standard.Visualization
  
  type Student
     Value id region
  
  main =
     operator1 = Student.Value 1 'EAST'
     operator2 = operator1.id
     operator2

• Create a new node operator1.id

Expected result

Node evaluates to a value

Actual result

Node evaluates to an error:

Method `id` of type Student could not be found

issue-type-accessor.mp4

[radeusgd]

This looks like duplicate of #5249, but good that you reported it because it apparently got unscheduled.

[hubertp]

This was demoted for now, since users cannot define their own types (intuitively) in the project currently (other than the editor but that is not what we want).

[radeusgd]

Hitting that issue again when doing Advent of Code - I tried defining my custom types and I cannot really define functions on them as that simply does not work:

I also cannot rely on customizing to_display_text on such types, because it is not picked up either.

[jdunkerley]

Possibly related to #8505

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-15):

Progress: Started working on the task. Was finishing the task with the compiler error on a duplicate self argument. Started working on the test case reproducing the scenario. It should be finished by 2023-12-21.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for yesterday (2023-12-18):

Progress: continue working on the task. Was trying to reproduce the issue with the test scenario. Was not able to prepare the test. Although the issue can be reproduced in gui with the same scenario. It should be finished by 2023-12-21.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-19):

Progress: continue working on the task. Started debugging the case in gui. Found that after the text edit, the runtime contains a type constructor with the wrong reference to the type. Might be related to caching. It should be finished by 2023-12-21.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-20):

Progress: continue working on the task. It has to do something with how atoms and atom constructors in particular are registered in the scope. It feels like they are some kind of singletons, and their scope is not updated during the program re-execution. It should be finished by 2023-12-21.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for yesterday (2023-12-21):

Progress: continue working on the task. Figured out that the method resolution fails because the resulting atom is cached. The cached atom contains a reference to an old type. This type is then used to resolve a method in an updated scope and the resolution fails. It should be finished by 2023-12-21.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new 🔴 DELAY for today (2023-12-22):

Summary: There is 5 days delay in implementation of the Unable to access constructor value in interactive mode (#7184) task.
It will cause 5 days delay for the delivery of this weekly plan.

Delay Cause: This issue was quite tricky to identify and debug. Now the cause is clear. It will take a couple more days to implement a fix and test everything.

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-22):

Progress: Continue working on the task. Implemented the fix when the type information is retained between the compilations. Started working on the case when the type is changed and the engine state needs to be properly invalidated. It should be finished by 2023-12-26.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-25):

Progress: Continue working on the task. Implemented the tests to check the caches. Updated the cache invalidation logic during the interactive compilation to invalidate the nodes that depend on the resolution errors. Started testing It should be finished by 2023-12-26.

Next Day: Next day I will be working on the #7184 task. Continue working on the task

[enso-bot]

Dmitry Bushev reports a new STANDUP for today (2023-12-26):

Progress: Continue working on the task. Finished the testing. Tweaked the cached invalidation logic a bit. Cleaned up, added the docs, and undrafted the PR. It should be finished by 2023-12-26.

Next Day: Next day I will be working on the #7184 task. Continue working on the task