I thought I had fixed this, but apparently did not.

1 patch for repository http://darcs.net/screened:

patch a6ebb72f5926f8ef2af860fcb46816e0f9ce5945
Author: Ben Franksen <ben.franksen@online.de>
Date:   Sat Aug  1 13:54:52 CEST 2020
  * resolve issue2570: coalesceHunk should take arguments in the natural order

Attachments

• patch-preview.txt
• resolve-issue2570_-coalescehunk-should-take-arguments-in-the-natural-order.dpatch
• unnamed

FYI the tests pass on Windows with this patch

Fine.

(not sure if it merits a test?)

> (not sure if it merits a test?)

What kind of test were you thinking of?

I thought you mentioned in issue2570 that some cases you expected to coalesce
didn't, so that might turn into a test using amend-record.

Not worth much effort, if any, though.

> I thought you mentioned in issue2570 that some cases you expected to coalesce
> didn't, so that might turn into a test using amend-record.

Ah, no. I was confused by the code itself and did not understand how it
could possibly work, so I constructed a few simple examples and tested
them in ghci with the raw function. All of the "problems" I noticed were
due to me not realising that the argument order was "wrong" i.e. the
arguments named xxx1 were actually from the second patch and vice versa.
This counter-intuitive naming and placement of the arguments is what the
patch fixes. There is no change in behavior at all.