[Bug] import of available module fails if a long docstring is added to the main module

[BertLindeman]

Describe the bug
There is a large program (> 1000 lines) and I use a docstring at the end of that program to keep my test info.
That log got a bit large and then the program would compile but the run soon crashes due to missing sub module. The docstring > 49000 characters.
Tested on beta.pybricks.com and on code.pybricks.com.
Beta version: v4.0.0b10 (Pybricks Beta v3.0.0-beta.7)
After I reduced the trailing docstring to two empty lines the program ran again.

To reproduce
Steps to reproduce the behavior:

1. Go to beta.pybricks.com
2. Try to run my program (Will try to get a smaller program to test with)
3. Terminal pane:

Traceback (most recent call last):
  File "m258673_h2.py", line 39, in <module>
ImportError: no module named 'm258673_config'

Expected behavior
What did you expect to happen instead?
A normal run as usual.

Screenshots
There is a saying that a picture is worth a 1000 words. Screenshots really help to identify and solve problems.

Console log
On beta I see an error, but on code.pybricks.com I did not see that error.

The console log shows that m258673_config.py is loaded:

5514.f94d1b05.chunk.js:1 creating code completion worker
7346.00d5129d.chunk.js:1 starting Pyodide...
pyodide.asm.js:1 Loading docstring_parser, jedi, parso, pybricks_jedi, pybricks, typing_extensions
main.d018c504.js:2 Web Bluetooth is experimental on this platform. See https://github.com/WebBluetoothCG/web-bluetooth/blob/main/implementation-status.md
pyodide.asm.js:1 Loaded docstring_parser, parso, jedi, pybricks_jedi, pybricks, typing_extensions
7346.00d5129d.chunk.js:1 copied issue_1678_calibrate_gyro.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_config.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_functions.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_h1.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_h2.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_hoist_control.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_main.py to emscripten fs
7346.00d5129d.chunk.js:1 copied m258673_test_grab.py to emscripten fs
7346.00d5129d.chunk.js:1 copied magweg.py to emscripten fs
7346.00d5129d.chunk.js:1 copied release_remote.py to emscripten fs
7346.00d5129d.chunk.js:1 copied reset_hoist.py to emscripten fs
7346.00d5129d.chunk.js:1 copied set_42146_base_xbox_v2.py to emscripten fs
7346.00d5129d.chunk.js:1 copied set_42146_config.py to emscripten fs
7346.00d5129d.chunk.js:1 copied set_42146_crane_v2.py to emscripten fs
7346.00d5129d.chunk.js:1 copied test_hoist.py to emscripten fs
7346.00d5129d.chunk.js:1 copied xbox_example_config.py to emscripten fs
7346.00d5129d.chunk.js:1 copied xbox_example_receiver.py to emscripten fs
7346.00d5129d.chunk.js:1 copied xbox_example_sender.py to emscripten fs
pyodide.asm.js:1 preloading pybricks_jedi...
main.d018c504.js:2 Downloading 10920 bytes
pyodide.asm.js:1 preloading done.
7346.00d5129d.chunk.js:1 Pyodide is ready.
5514.f94d1b05.chunk.js:1 code completion engine is ready
5514.f94d1b05.chunk.js:1 required headers missing for SharedArrayBuffer, cancellation will not work
R @ 5514.f94d1b05.chunk.js:1
[Violation] Forced reflow while executing JavaScript took 41ms

[dlech]

Sounds similar to #1804

[BertLindeman]

closed dup

[BertLindeman]

try again to close dup

[laurensvalk]

It could be similar bit since

On beta I see an error, but on code.pybricks.com I did not see that error.

We should keep it open.