Thanks for this.

I'm somewhat concerned that this causes a regression in one of our internal tests. However, the test seems to generate TRAP that is not representative of what the C/C++ extractor will generate (which will always include an appropriate location). So I think this change is ok, but I'd like to hear the opinion of someone else in @github/codeql-c-analysis.

If the test isn't representative of what the extractor actually generates I'm also totally okay with this. In fact, we should probably update the test to better reflect what the extractor does.

In fact, we should probably update the test to better reflect what the extractor does.

Not sure whether that's worth it. The test tests something funny with sizes of things, and locations/urls are not really relevant for that and the size part of the test doesn't break by this change.

@erik-krogh Can you add a change note? I think we can approve/merge this.

@erik-krogh The change note check failed. Also happy 2023 (see date on change note).

I think the date on the change note should be 2023-06-19, not 2023-19-06.

Rebased on main to get around a submodule bump.

And another rebase on main to get around a submodule bump.