On 14.01.2024 20:18, gerard.vermeulen@posteo.net wrote:
> On 14.01.2024 13:16, Ihor Radchenko wrote:
>> gerard.vermeulen@posteo.net writes:
> [...]
>> I tried make test and the tests are still failing with this new patch:
>> 
>> 1 unexpected results:
>>    FAILED  test-ob/demarcate-block  ((should (string= region-text
>> (org-trim (nth 1 info)))) :form (string= "mark this line as region"
>> "") :value nil :explanation (arrays-of-different-length 24 0 "mark
>> this line as region" "" first-mismatch-at 0))
> 
> This is a tough lesson: the tests pass always on my system.
> 
> I think the failure you see is related to a problem marking a region
> in my test code (wish: support in `org-test-with-temp-text' for
> "<mark>" besides "<point>", but maybe that depends on ERT), else
> the problem would have shown up while testing the patch interactively.
> 
> I think that I have improved my region marking code by using "<point>"
> in the temp-text as a start. Then, I only have to find where to set 
> mark,
> and eventually exchange point and mark.
> 
> The test code now checks (mark) in the 3 places where a region is 
> marked.
> This looks superfluous if the code is really robust, but at least it 
> checks
> whether the region marking is (or was) the problem.
> 

To converge faster, I have split the test into 6 tests:

test-ob/demarcate-block-split-duplication
test-ob/demarcate-block-split-prefix-point
test-ob/demarcate-block-split-prefix-region
test-ob/demarcate-block-split-user-errors
test-ob/demarcate-block-wrap-point
test-ob/demarcate-block-wrap-region

The test failure on your system was due to the sub-test that is now
test-ob/demarcate-block-wrap-region

Regards -- Gerard