Replace built-in docs codeblock's leading spaces with tabs

[Mickeon]

This PR replaces the spaces used in the codeblocks of the class references with actual tabs.
If your Editor Setting's preference is using spaces, nothing changes.
With one GIF, this means the following (notice the tabs):

Before	After

However, hm... It comes at a cost. For the tab to display in a way that actually looks good, the RichTextLabel's tab_size has to be set to 8, otherwise the tab's indent would be way, way too small, for some reason. Because of this, there's the side effect of increasing the indentation across the whole built-in documentation:

Take a closer look:

Before	After

Some may say this is a happy accident, some may absolutely not. So keep that in mind.

[YuriSizov]

Could you please rebase on the current master? We had an issue with macOS builds which failed your CI.

[Mickeon]

Done just that.

[MewPurPur]

I don't like the increased indent but I think it's worth it. I'm just confused why we'd need an 8 tab size for what's supposed to be monospace. Also, I would expect the mono tabs to be as wide as my code editor theme, which isn't 4.

This is still better than before, I don't think these nitpicks are more important than getting it merged.

[Mickeon]

I'm just confused why we'd need an 8 tab size for what's supposed to be monospace.

Honestly, I am not sure why RichTextLabel acts like this. I tried a value of 4 which you would think would be appropriate, but it resulted in the padding being way too short for some reason.

It originally was also based on the "indent size" Editor Setting, but because it would've affected the entire RichTextLabel I was not sure about it.

[AThousandShips]

LGTM

[Calinou]

Tested locally, there are issues with code samples that contain multiple tabs on the same line.

For instance, the HTTPRequest class description has a code sample where the push_error() line has two leading tabs, but none are displayed and copied here:

If you select the block and paste it in a script, you get this:

[Mickeon]

You will never guess who the idiot that completely f_u^c_k^ed up the formula is.

Should work fine now:

[Mickeon]

Updated the PR to use StringBuilder. I tested and nothing really changed on the outside.

Bugs fixed.

[akien-mga]

Thanks!

[Mickeon]

And from here on out begins the built-in reference revolution