Peter Ludemann

Also with http://localhost:9999/static/src_browser.html?corpus=CORPUS&root=ROOT&path=tmp%2Fpykythe_test%2FSUBST%2Fhome%2Fpeter%2Fsrc%2Fpykythe%2Ftest_data%2Fpy3_test_grammar.py#L1 ... they all seem to be for `lambda`, which generates a pseudo-token

https://bugs.python.org/issue38551

https://bugs.python.org/issue36541 contains a proposed PR to bring lib2to3 up-to-date.

Consider using BerkeleyDB. More discussion here: https://swi-prolog.discourse.group/t/scaling-to-billions-of-facts/380/9

`black` doesn't do this kind of string reformatting either. `black` does add parens to import statements (`yapf` doesn't based on its design philosophy), although I don't like the format that...

This isn't a test case (yet). I ran the tests, then manually loaded the foreign code with the command line I gave above (which is what the test cases run)...

Also, the bare word "dataclasses" in `from dataclasses import dataclass` doesn't handle mouseover. And the "." in `from . import pod` isn't highlighted. These probably have to do with the...

I've added "delete"d definitions for assignment and copy constructor. I've also gone through the other classes and made explicit whether they are copyable or not. In general, I've disabled the...

We can circumscribe the possibilities to all the (recursive) imports. One possibility would be to add `ref/possible` to the Kythe data model, and then generate `ref/possible` for all entries in...

The move constructors have been added, but leaving this open because `PlRecord` should act like a "smart pointer", taking advantage of the way PL_duplicate_record() and PL_erase() work (with reference counts)....