fix: enforce size constraints on window creation on Windows and Linux

[mitchchn]

Description of Change

I noticed that size constraints were not enforced properly on window creation on Linux and Windows:

new BrowserWindow({
  width: 200,
  minWidth: 500,
});

In some environments, this simply creates a window that is too small. In others, the conflict is resolved by the window manager on show, but not before the window gets centered in the wrong position (and initially reports the wrong size).

This PR applies all size constraints on init and moves centering/positioning so it happens after the final window size has been determined.

Checklist

• PR description included
• I have built and tested this PR
• npm test passes
• tests are changed or added
• relevant API documentation, tutorials, and examples are updated and follow the documentation style guide
• PR release notes describe the change in a way relevant to app developers, and are capitalized, punctuated, and past tense.

Release Notes

Notes: BrowserWindow now enforces min/max size constraints on window creation, even if they conflict with the requested width and height.

[ckerr]

These CI failures eg. here seem related to the PR:

Failure in test: "BrowserWindow module sizing BrowserWindow.setMinimum/MaximumSize(width, height) creates window at max size when a larger size is requested"
AssertionError: expected 184 to equal 200

Same test failure on windows-arm64, windows-x86, and windows-x64

[mitchchn]

These CI failures eg. here seem related to the PR:

Failure in test: "BrowserWindow module sizing BrowserWindow.setMinimum/MaximumSize(width, height) creates window at max size when a larger size is requested"
AssertionError: expected 184 to equal 200

Same test failure on windows-arm64, windows-x86, and windows-x64

Fixed!

[ckerr]

@mitchchn the bots want target/* labels. Which branches need this fix?

[mitchchn]

@mitchchn the bots want target/* labels. Which branches need this fix?

Only 41 and up.

[jkleinsc]

Looks like mac arm64 tests are failing. Can you rebase this PR to see if that resolves those failing tests?

[jkleinsc]

Looks like this change is causing mac arm64 mas tests to fail. I'm not sure why its failing but it does look like the removed code would have executed on mac so maybe its needed there.

[mitchchn]

@jkleinsc yup, that test caught a subtle bug and I did need to put back a condition I removed. (Not sure why it only failed on ARM64/MAS...)

[ckerr]

Needs rereview from @jkleinsc

[release-clerk]

Release Notes Persisted

BrowserWindow now enforces min/max size constraints on window creation, even if they conflict with the requested width and height.

[trop]

I have automatically backported this PR to "41-x-y", please check out #50753

[trop]

I have automatically backported this PR to "42-x-y", please check out #50754