Support standard library (Attempt 3) #109
Conversation
…x with both gast and ast without additional parameters.
…ing gast This should make it possible to derive from beniget to implement support for ast module in addition to gast module.
|
Idk why the tests failed, it works on my local computer and it looks like I can’t access the log of the GitHub CI. Can you help me @serge-sans-paille ? Thanks |
|
I really like where it's going! Some comments inline though |
| return self.node.name | ||
| elif _PY310PLUS and isinstance(self.node, ast.MatchMapping) \ | ||
| and self.node.rest: | ||
| return self.node.rest |
There was a problem hiding this comment.
This is the place I like the less, because we clearly loose the generic aspect of gast
There was a problem hiding this comment.
We could use __class__.__name__ == str for these case to keep it generic.
There was a problem hiding this comment.
Hello @serge-sans-paille, I’ve applied the suggested changes :)
| @@ -559,17 +601,76 @@ def process_annotations(self): | |||
| cb(visitor(annnode)) if cb else visitor(annnode) | |||
| self._scopes = currenthead | |||
| self.defs = compute_defs | |||
|
|
|||
| def _support_stdlib(self): | |||
There was a problem hiding this comment.
Question: should / could that be put in a subclass?
There was a problem hiding this comment.
This way simplifies the usage as well as the testing.
There was a problem hiding this comment.
But to answer your question: yes it could be a subclass. But there is no value for it to be put in a subclass instead of this.
|
Grand merci pour ton feed-back @serge-sans-paille ! |
…eveloment computers (tox).
|
I'm good with this! can you rebase & squash into a nice stack and we're done :-) |
|
Could you just use the « squash and merge » option of GitHub ? Thanks |
|
On Sat, Oct 12, 2024 at 02:37:58PM -0700, tristanlatr wrote:
Could you just use the « squash and merge » option of GitHub ? Thanks
No :-) There are conflicts.
If you can't handle it, I'll copy your branch and do it myself...
|
|
It looks like there are no conflicts… I don’t get it
|
|
Manually rebased, squashed and merged as of de126b0 |
|
Fantastique, merci |
This attempt uses the
pkg()function proposed in #101.This changes should probably be squashed since a lot of the commit messages are irrelevant.
For the record:
if node.__class__.__name__ == <str>.astmodule only at strategics place and then storing it in aself.astvariable.But in any case some methods needs to be added to the
DefUseChainsclass and a couple of places needs to be adjusted to cope with structural differences in between ast and gast, for instance the fact that expect handlers name are stored asNameinstead ofstrandargdoesn't exist ingast.