[node-core-library] Fix Text.truncateWithEllipsis() exceeding maximumLength

[Alexandre Kohler (kwy404)]

Summary

Text.truncateWithEllipsis() checks s.length <= 3 where it should check maximumLength < 3 before deciding to skip the ellipsis. So for any input longer than 3 characters with maximumLength of 0, 1 or 2, it returns '...', which is longer than maximumLength. For example Text.truncateWithEllipsis('12345', 2) and Text.truncateWithEllipsis('12345', 0) both return '...'.

This PR changes the check to maximumLength < 3.

Details

With the old check, those inputs fall through to s.substring(0, maximumLength - 3) + '...'. The negative end index is treated as 0, so the result is just '...'.

Results do not change for any other input: an input of 3 characters or fewer only reaches this check when maximumLength is already below 3, and maximumLength of 3 or more still gets the ellipsis as before.

rushx can hit this when it truncates script bodies to the console width on a narrow terminal.

How it was tested

Added a unit test in src/test/Text.test.ts that truncates '12345' to lengths 0 and 2. Without the fix it fails (Expected: "", Received: "..."). With the fix all 16 Text tests pass. Also ran the node-core-library build with lint and API Extractor, rush prettier and rush change --verify.