Respect --file-lines when making various whitespace-related changes.

[martinboehme]

Various whitespace-related changes did not so far respect --file-lines, see also #5136. This PR fixes this in all of the cases reported in #5136:

• Newlines at the beginning of the file
• Spaces at the beginning of the file
• Missing newline at the end of the file
• Space after a doc comment
• Space before a comment in the last line of the file (without trailing newline)

The PR includes tests that fail without the fixes.

[martinboehme]

Edit: Disregard. PR is ready for review now.

Please hold off on reviewing for now. While tests previously passed for me, I see that they are now failing, for reasons that aren't clear to me.

[ytmimi]

Checking the logs shows that the system_tests and idempotence_tests failed:

test test::system_tests ... FAILED
test test::idempotence_tests ... FAILED

Specifically I see:

Mismatch at tests/source/issue-5136-3.rs:4:
 // Important: When editing this file, make sure not to add a newline at the end
 // of the last line.
 use std;
+

Mismatch at tests/target/issue-5136-3.rs:4:
 // Important: When editing this file, make sure not to add a newline at the end
 // of the last line.
 use std;
+

[martinboehme]

Checking the logs shows that the system_tests and idempotence_tests failed:
...

Yes, thanks. See also my previous comment:

Please hold off on reviewing for now. While tests previously passed for me, I see that they are now failing, for reasons that aren't clear to me.

I'm investigating and will let you know when the PR is ready to review. Sorry for the noise.

[martinboehme]

The issue is now fixed.

I had removed part of the code that I had thought I had convinced myself was not needed -- obviously, this was not the case. I have now reinstated the code, and the tests pass.

[martinboehme]

Since the PR wasn't in review yet, I've modified it to add tests and fixes for the two remaining issues from #5136. The PR is now a complete fix for #5136.

The PR currently consists of multiple commits -- should I squash these into a single commit?

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[martinboehme]

I've resolved the merge conflicts and, in the process, squashed the previously multiple commits into one.