Small improvement on refactor group resizable panel test

[tintinthong]

This is 1st step in possibly refactoring resizable panel. Basically, a lot of discussion/confusion fadhlan and I have is which argument is preferred defaultLengthFraction or lengthPx or minLengthPx-- I anticipate that we shud add these arguments that will be tested. And their easily mutable doing the same test setup.

If test pass it should be good to go in. These are only changes to the test

This doesn't change any application code only test code

[github-actions]

Test Results

576 tests  ±0   572 ✔️ ±0   8m 24s ⏱️ -25s
    1 suites ±0       4 💤 ±0 
    1 files   ±0       0 ❌ ±0 

Results for commit 5739ccb. ± Comparison against base commit 789e59d.

[jurgenwerk]

Looks good just wondering why a refactor is needed? Are there bugs in the panel implementation?

[tintinthong]

Looks good just wondering why a refactor is needed? Are there bugs in the panel implementation?

Fair question. Well the bugs which we currently see is

• https://linear.app/cardstack/issue/CS-6642/resizable-panel-id-undefined
• https://linear.app/cardstack/issue/CS-6238/resizable-panel-snaps-out-when-going-into-code-mode-from-empty-stack

The 2nd ticket is merged and fixed. But its a workaround.

Ok as part of the refactor

1. want to solve the bugs above
2. separate the actions of register/unregister and onContainerResize. We seem to rely onContainerResize to often make sure the panel ratios are correct. My question was always, why arent the ratios correct after registering a panel
3. hide ratio computation in an abstraction
4. have greater clarity with usage of arguments

I think if we do 2 . we should fix the bugs. Perhaps that is a good place to stop (tho I suspect itll be a pain to do it without 3)