Good news re leoAst.py

[Edward K. Ream]

In a previous thread I said that work on this project is complete. However, I have some news to report.

Guido and I have been discussing whether it makes sense to add the capabilities of leoAst.py to python's stdlib. We both now agree that doing so would be a bad idea. To summarize: adding even the simplest new function to python's ast module would imply exposing the api's of the TokenOrderGenerator and TokenOrderTraverser classes. Alas, doing so would cast those api's in stone forever. It's not something for which I want to be responsible.

There is some disappointment in the conclusion. Happily, there is a big consolation prize.

Guido is open to adding links to leoAst.py at the bottom of the docs for the ast module. Such a link would be a big deal. The link would amount to be a strong endorsement of leoAst.py. To mute possible pushback against adding these links, I'll be coordinating discussions with the devs for asttokens, parso and LibCST, with the aim of adding links to these tools as well.

All in all, this seems like the best possible result for leoAst.py.

Edward