Context handling improvements and auth fixes

[daniel-de-vera]

This PR includes the following changes:

• Improves context handling across various CLI commands by properly linking contexts to signals, making undo operations more reliable.
• Fixes the auth token refresh issue reported in [CLI] Override exiting with 401 mid session #228.
• Normalizes and aligns code between the override and traffic record commands.

Brief demo (has audio):

screen-capture.43.webm

[gitar-bot]

Gitar analyzes/fixes CI failures, runs prompts as workflows and addresses comments starting with Gitar.

⚙️ Options:

• Automatically accept changes

---

📝 Submit Feedback

[scott-cotton]

LGTM

But maybe some of those context.Background calls on cleanup should somehow take the overall command invocation context instead.

Unless the extra plumbing complexity doesn't seem worth it...

[daniel-de-vera]

But maybe some of those context.Background calls on cleanup should somehow take the overall command invocation context instead.

Unless the extra plumbing complexity doesn't seem worth it...

The problem is that in some cases (e.g. on ctr-c), the context may be already cancelled.

[scott-cotton]

But maybe some of those context.Background calls on cleanup should somehow take the overall command invocation context instead.
Unless the extra plumbing complexity doesn't seem worth it...

The problem is that in some cases (e.g. on ctr-c), the context may be already cancelled.

I think the problem is also that there is a containing context whose cancelation is not respected. One could have a context that does not cancel on control-C as the root, for example.

[daniel-de-vera]

But maybe some of those context.Background calls on cleanup should somehow take the overall command invocation context instead.
Unless the extra plumbing complexity doesn't seem worth it...

The problem is that in some cases (e.g. on ctr-c), the context may be already cancelled.

I think the problem is also that there is a containing context whose cancelation is not respected. One could have a context that does not cancel on control-C as the root, for example.

Ok, let me see what I can do.

[daniel-de-vera]

@scott-cotton, PTAL:

In bc1c763 I implemented:

I think the problem is also that there is a containing context whose cancelation is not respected. One could have a context that does not cancel on control-C as the root, for example.

In 9d50091 a minor file reorg in override package (to make it uniform with the rest).

[scott-cotton]

@scott-cotton, PTAL:

In bc1c763 I implemented:

I think the problem is also that there is a containing context whose cancelation is not respected. One could have a context that does not cancel on control-C as the root, for example.

In 9d50091 a minor file reorg in override package (to make it uniform with the rest).

LGTM