fix: Add html extension if build.format is file

[fjdelarubia]

Changes

This change adds .html extension to the sitemap urls if the build.format is set to file, this way the urls will match with the canonical url

In the related issue #11575 @ematipico suggest to remove the .html extension from the urls, not sure about it because the extension is forced in core/build/generate (function getUrlForPath), but I can undo this changes and modify that function

Sample project's structure:

Before urls:
https://some-site.com/legal/aviso-legal
https://some-site.com/legal/cookies
https://some-site.com/legal/politica-de-privacidad
https://some-site.com/legal/terminos-y-condiciones
https://some-site.com/legal
https://some-site.com/productos/carteles-personalizados/carteles-para-bodas
https://some-site.com/productos/carteles-personalizados/carteles-para-cumplea%C3%B1os
https://some-site.com/productos/natalicios
https://some-site.com/

After urls:
https://some-site.com/legal/aviso-legal.html
https://some-site.com/legal/cookies.html
https://some-site.com/legal/politica-de-privacidad.html
https://some-site.com/legal/terminos-y-condiciones.html
https://some-site.com/legal.html
https://some-site.com/productos/carteles-personalizados/carteles-para-bodas.html
https://some-site.com/productos/carteles-personalizados/carteles-para-cumplea%C3%B1os.html
https://some-site.com/productos/natalicios.html
https://some-site.com/

Testing

I need some help for the test, would be ok to create new fixture for static build with new build format?

Docs

This PR tries to fix the sitemap behaviour to match the canonical url explained in the docs, see #11575

[changeset-bot]

🦋 Changeset detected

Latest commit: 692d493

The changes in this PR will be included in the next version bump.

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[ematipico]

I am not sure this is the correct fix. You should not add .html to the final URLs.

[matthewp]

Thanks! The build.format option is only for the build, to tell it how to output files. It doesn't signal what URLs are supposed to look like. That's determined by your server. If you want to have .html in your URLs maybe adding an option in the sitemap integration is a better way to go.

[fjdelarubia]

Thanks! The build.format option is only for the build, to tell it how to output files. It doesn't signal what URLs are supposed to look like. That's determined by your server. If you want to have .html in your URLs maybe adding an option in the sitemap integration is a better way to go.

I think not only in the sitemap integration, but on the build process the .html extension is added here https://github.com/withastro/astro/blob/main/packages/astro/src/core/build/generate.ts#L375

This url is used for creating the request which renders the page, and it adds the .html extension to the pathname, which makes the canonical url to contain the .html extension.

The .html extension were added here:
cd154e4

If the .html extension should not be visible, nor in the canonical url nor the sitemap, just removing it from the default state would fix it

switch (format) {
		case 'directory':
		case 'preserve': {
			ending = trailingSlash === 'never' ? '' : '/';
			break;
		}
		default: {
			ending = '';
			break;
		}
	}
``