NiceGUI renders blank page when behind a main guard on v3

[ClementPinard]

First Check

• I added a very descriptive title here.
• This is not a Q&A. I am sure something is wrong with NiceGUI or its documentation.
• I used the GitHub search to find a similar issue and came up empty.

Example Code

from nicegui import ui

if __name__ in {"__main__", "__mp_main__"}:
    ui.label("Hello world")
    ui.run(reload=True)

Description

This code works with nicegui<3, but with version 3.0.4 I get a blank page.

However, this code works:

if __name__ == "__main__" or current_process().name != "MainProcess":
    ui.label("Hello world")
    ui.run()

I first came upon this problem when trying to run it as an entrypoint script.

Pyproject.toml would look like

[project.scripts]
main = "test_project.main:run"

and the code

def run(reload=False):
    """This is the main function that gets run"""
    ui.label("Hello world")
    ui.run(reload=reload)


if __name__ in {"__main__", "__mp_main__"}:
    run(reload=True)

which is a pattern I saw in this discussion : #3769

And this does not work as well: uv run main gives me blank page.

Not sure these two problems are related though, but they both disappear when I use nicegui==2.24.2 (last version <3)

NiceGUI Version

3.0.4

Python Version

3.13.7

Browser

Chrome

Operating System

macOS

Additional Context

No response

[evnchn]

Hello @ClementPinard thanks for the report.

Can recreate. Using print('__name__:', __name__) we see __name__ is <run_path>.

That is from defaults of

nicegui/nicegui/ui_run.py

Line 129 in b6ebab6

runpy.run_path(sys.argv[0])

Short term fix: You can use main guard if __name__ in {'__main__', '__mp_main__', '<run_path>'}:

Long-term fix: We will have to consider runpy.run_path(sys.argv[0], run_name='__main__')

[evnchn]

@falkoschindler So you are in progress of a PR on this? Because I am quite hesitant.

I fear that we may run into cases of infinite-loop, if we do runpy.run_path(sys.argv[0], run_name='__main__') in our code, as it would basically make our invocation "indistinguishable".

Now to be clear, applying this patch does not lead to infinite-loop by itself, but I am just envisioning if there are user code written with 3.0.4 behaviour in mind that would then infinite-loop or fail catastrophically given this change.