Open bbriscoe opened 1 year ago
This issue is referred to in this thread on tools-discuss: https://mailarchive.ietf.org/arch/msg/tools-discuss/Ti0TecUKdBC42YpUNz5LAsN31Vo/
Here's a screen shot illustrating the issue. At least the last four flagged changes are bogus (different, Also the, in terms of, packet traveled). It looks as if some of the newlines are not being treated as neutral whitespace. I don't have the old rfcdiff installed, so I can't double check, but I never saw this issue there.
(The files compared were intentionally ancient, draft-carpenter-midtax-01.txt and draft-carpenter-midtax-03.txt )
Compare the output of iddiff with the output from rfcdiff for almost any pair of files, and it will be seen that iddiff does not ignore changes in line wrap, which is a serious regression relative to rfcdiff.
For example, if a few words are added or removed in the middle of a para, for the rest of the para there will be blocks of text highlighted around the starts and ends of every subsequent line. Also, sometimes this causes iddiff to give up and just highlight everything from the first line-break after the actual change until the end of the para.
Example: compare the diffs between the most recently published RFC (9329) and the last draft before it was approved (note, however, that almost any pair of files will demonstrate the problem): idiff: https://author-tools.ietf.org/iddiff?url1=draft-ietf-ipsecme-rfc8229bis-09&url2=rfc9329&difftype=--html rfcdiff: https://bobbriscoe.net/tmp/rfc932-from-draft-ietf-ipsecme-rfc8229bis-0.diff.html
I suggest you skip over all the boilerplate to Section 1.
Bob