Attached is a patch that enhances doctest discovery so that it can correctly report the line number for doctest errors in tests that are string values of the __test__ dictionary.  It only works if the strings are formatted as triple quoted strings, but since that is the most common case I think it is worth doing.

I don't have a test for this...I think it will require a unit test rather than adding to the doctests, since it needs to check the reported line number and other changes to test_doctest.py would change the line numbers.

I am not as good in making nice patches, but that code can be improved upon a bit as follows:
<insert patch intro blurb here, same as original>

+        if isinstance(obj, str) and source_lines is not None:
+            # This will find __test__ string doctests if and only if the string
+            # contains any unique line.
+            for offset,line in enumerate(obj.splitlines(keepends=True):
+                if source_lines.count(line)==1:
+                    lineno = source_lines.index(line)-offset
+                    break

         # We couldn't find the line number.
         return None


I thing this will improve legibility and probably also speed under most circumstances.

Oops. Lousy editing, I should have used return instead of break.

These 6 lines can be inserted in lib/doctest at line 1100:
+        if isinstance(obj, str) and source_lines is not None:
+            # This will find __test__ string doctests if and only if the string
+            # contains any unique line.
+            for offset,line in enumerate(obj.splitlines(keepends=True):
+                if source_lines.count(line)==1:
+                    return source_lines.index(line)-offset

And it works fine for me; the code is quite general and has predictable behaviour: any test string with a unique line in it will work.

I always use the following piece of code that can be trivially used to patch the source.
It basically looks for a using line in the test string and finds it in the source file. If there is a match, we know where we are. Otherwise it falls back to the "normal" way.

Oh wait, I already posted that so long ago I forgot about it.
Oops.
Sorry about the repetition, folks.
- Jurjen

Hey Jurjen,

Do you mind converting your patch in a Pull request please?

Hey  R. David Murray,

Where do you think the unit test for `doctest` should be placed? 

Should it be in `test_doctest.py`?

Without looking at doctest.py, yes.  I believe the doctests in that file should be already plugged in to the unittest framework, so adding new testcase containing a non-doctest unit test should work fine.

Hi, Should I convert patch into PR?
Thanks

Yes. That would make me happy. In the meantime I learned how to use git, so maybe I'll do the pull request myself next time. Thanks for the work.

I tried to make a pull request, but it fails on the format of news file name.
At least the tests all pass.

All is fixed. The PR apparently is correct.
My first dent in the Python universe, however small :-)