kp-forks / grobid / 28126898290

39%

push

github

Guard against malformed callout interval in toTEITextPiece  (#1472)

When two footnotes share the same superscript marker, the note-matching loop could produce an overlapping/reversed offset interval (start > end). The subsequent List.subList(start, end) then threw IllegalArgumentException: fromIndex > toIndex, aborting the whole document (processFulltextDocument returning HTTP 500). Skip such malformed intervals gracefully so the callout falls back to plain text. This would solve #1024.

As a bonus, we fixed also a regression in the Lexicon refactor #1440

8668 of 24958 branches covered (34.73%)

Branch coverage included in aggregate %.

18454 of 45263 relevant lines covered (40.77%)

1.65 hits per line