Currently in the UsfmVersificationErrorDetector we only register warnings for missing verses and chapters at the ends of chapters and books respectively. For example, if you run it over 2JN 1 and it only has 3 verses, you will only get one error saying that it expected 2JN 1:13 (the last verse) but saw 2JN 1:3. Similar story for chapters. This means two things:
- A single warning may actually inform you that many verses are missing (which may be misleading to those looking at the warnings). E.g., "I only saw 2 warnings, so it's probably not a big deal".
- We are not capturing mid-chapter missing verses or mid-book missing chapters.
Capturing these would add a good bit of complexity to the detector, so we should consider how important this is (i.e., is this a 'bug' or an 'enhancement'?). If we were to capture these mid-book/verse missing items, I believe that the strategy that we would have to employ would also allow us to count affected verses. That may be a more reliable at-a-glance metric for those looking over the warnings.
Currently in the
UsfmVersificationErrorDetectorwe only register warnings for missing verses and chapters at the ends of chapters and books respectively. For example, if you run it over 2JN 1 and it only has 3 verses, you will only get one error saying that it expected 2JN 1:13 (the last verse) but saw 2JN 1:3. Similar story for chapters. This means two things:Capturing these would add a good bit of complexity to the detector, so we should consider how important this is (i.e., is this a 'bug' or an 'enhancement'?). If we were to capture these mid-book/verse missing items, I believe that the strategy that we would have to employ would also allow us to count affected verses. That may be a more reliable at-a-glance metric for those looking over the warnings.