Fix a few entry_point parameters

[nical]

Description

A few must have been missed in #5148 and somehow CI let it pass but caught it in #6030.

[ErichDonGubler]

I can see how that happened/that PR became necessary. Those tests were added after the original base of the optional entry point PR. $OP-and-merging would not have caught that for either of them unless somebody had rebased and then checked CI after the other had merged. 😅

This is a point in favor of forcing PRs to be up-to-date before bringing into mainline history; CC @jimblandy, @cwfitzgerald.

[cwfitzgerald]

Forcing PR's to be up to date is really annoying, but this is a good argument due merge trains, which we can enable.

[sagudev]

$OP-and-merging would not have caught that for either of them unless somebody had rebased and then checked CI after the other had merged. 😅

This is a point in favor of forcing PRs to be up-to-date before bringing into mainline history

We use GitHub's merge queue in servo for this.

[ErichDonGubler]

Oddly, MacOS CI timed out. Re-running.

[ErichDonGubler]

Even if CI times out again, this PR is likely worth pushing to mainline anyway, since it fixes compilation failures. 😅

[teoxoy]

Oddly, MacOS CI timed out. Re-running.

Yeah, it started with #6021.