Section 3.1 of the irregex documentation specifies the meaning of the index parameter of the kons lambda taken by irregex-fold:
http://synthcode.com/scheme/irregex/#SECTION_3.1
(irregex-fold <irx> <kons> <knil> <str> [<finish> <start> <end>])
This performs a fold operation over every non-overlapping place <irx> occurs in the string str.
The <kons> procedure takes the following signature:
(<kons> <from-index> <match> <seed>)
where <from-index> is the index from where we started searching (initially <start> and thereafter the end index of the last match), <match> is the resulting match-data object, and <seed> is the accumulated fold result starting with <knil>.
Equating the search start index with the last match end index is false, as the documentation notes a bit further on:
Note if an empty match is found <kons> will be called on that empty string, and to avoid an infinite loop matching will resume at the next char. It is up to the programmer to do something sensible with the skipped char in this case.
This is a poor design choice, because it effectively mandates that all nontrivial kons lambdas keep track of their own last match end index, as the fixed irregex-split >>154 does. With author powers to change the public contract the proper design choice would be to switch the meaning of the index parameter of the kons to the last match end index. Irregex-fold/fast would keep track internally of both the search start index and the last match end index, and would pass the former to searches and the latter to 'kons' and 'finish'. This would simplify empty match handling, as well as irregex-replace/all and irregex-split. But this is outside the scope of a mere bugfix as it requires a change to the public contract, and is therefore something only ashinn can do.
With the group of fixes for the fold family >>140 >>141 >>154 and the bonus aborting attack >>158, I think this should be enough material for the second irregex bughunt. The first irregex bughunt, with the leftmost longest violations from no backtracking and the bonus DoS attacks, can be found in the group of posts starting at https://textboard.org/prog/39#t39p100 as well as a few preceding posts in that thread.
You don't have a job do you?
I work in academia and have for several decades. In all those decades, however, I've never had such a devoted follower. What's your day job, Anon?
I can't imagine how you'd interact with other people.
You can easily glean the answer by comparing the two sides of the exchange: without any need to resort to insults, and using elements of reality instead of "a stream of trivially refutable falsehoods". >>152