Releases: PyCQA/baron
Releases · PyCQA/baron
0.9 (2019-02-01)
First version of full python 3.7 grammar support.
- BREAKING CHANGE: annotations are now member of {def,list,dict}_argument to flatten the data structure
- add support for ... in from import by bram
- add support for return annotation by bram
- add support for exec function by bram
- add support for variable annotation #145 by scottbelden and additional work by bram
- add support for *var expressions in tuple assignment by bram
- add support for raise from #120 by odcinek with additional work by bram
- add support for arglist usage in class definition inheritence by bram
- bug fix by 91e839a kyleatmakrs
0.8 (2018-10-29)
- add typed parameters support #140 by Scott Belden and and additional work by bram
0.7 (2018-08-21)
- fix line continuation #92 by ibizaman
- handle corrupt cache file situation #76 by ryu2
- fix special crashing edge case in indentation marker https://github.com/PyCQA/bar by Ahuge
- fixed incorrect tokenization case "d*e-1". Fixes #85 #107 by boxed
- fix endl handling inside groupings by kyleatmakrs (extracted from #126)
Python 3:
- python 3 parsing extracted from #126
- support ellipsis #121 by odcinek
- support matrix operator #117 by odcinek
- support f-strings #110 by odcinek
- support numeric literals #111 by odcinek
- support nonlocal statement #112 by odcinek
- support keyword only markers #108 by boxed
- support yield from statement #113 by odcinek and additional work by bram
- support async/await statements #114 by odcinek and additional work by bram
0.6.6
0.6.5
- fix previous regression fix was broken
0.6.4
- fix regression in case a comment follow the ":" of a if/def/other
0.6.3
- group formatting at start of file or preceded by space with comment
0.6.2
- fix race condition when generating parser cache file
- make all user-facing errors inherit from the same BaronError class
- fix: dotted_name and float_exponant_complex were missing from
nodes_rendering_order
0.6.1
- fix: the string was having a greedy behavior on grouping the string tokens
surrounding it (for string chains), this ends up creating an inconsistancy in
the way string was grouped in general - fix: better number parsing handling, everything isn't fixed yet
- make all (expected) errors inherit from the same BaronError class
- fix: parsing fails correctly if a quoted string is not closed
0.6
- FST structure modification: def_argument_tuple is no more and all arguments
now have a coherent structure:- def_argument node name attribute has been renamed to target, like in assign
- target attribute now points to a dict, not to a string
- old name -> string are now target -> name_node
- def_argument_tuple is now a def_argument where target points to a tuple
- this specific tuple will only has name and comma and tuple members (no more
def_argument for name)
- new node: long, before int and long where merged but that was causing problems