I tried to run make test with your patch, and it is failing:
On my side it passes on main branch at commit
6c862699a6db3f6b76391c05380d92d9f1b3838f with patch.
passed 444/1207 test-org-capture/entry (0.099791 sec)
...
passed 447/1207 test-org-capture/org-capture-expand-olp (0.001104 sec)
Apparently, you get
`"* A\n** B\n*** 1970\n**** 1970-01 January\n***** 1970-01-01 Thursday\n****** H1
Capture text\n** C\n"'
Whereas I get the expected result of
`"* A\n** B\n*** 1969\n**** 1969-12 December\n***** 1969-12-31 Wednesday\n****** H1
Capture text\n** C\n"'
What do you get if you evaluate `(calendar-gregorian-from-absolute
(time-to-days 0))'? I get `(12 31 1969)'.
I set `org-overriding-default-time' to `0' for testing with datetree.
I don't have to additionally test for whether the lambda (as a target)
is actually called while visiting the file right?
May you elaborate what you mean?
So I've written a test for lambda-based target.
(should
(equal
"* A\n* B\n** H1 Capture text\n* C\n"
(org-test-with-temp-text-in-file "* A\n* B\n* C\n"
(let* ((file (buffer-file-name))
(org-capture-templates
`(("t" "Todo" entry (file+headline ,file ,(lambda () "B")) "** H1
%?"))))
(org-capture nil "t")
(insert "Capture text")
(org-capture-finalize))
(buffer-string))))
But it does not verify that the lambda `(lambda () "B")' is actually being
called while
visiting the file which would allow reading the file to compute or generate a
headline.
So I'm wondering if I should write a test for that too. Maybe for each new
lambda target?
