Set transform tokensset#708
Conversation
|
e1d1f2b to
07ef9bb
Compare
|
Hmm, this has had a wider impact that anticipated. Some of the changes just look like re-ordering, but there are also some dropped tokens. Is this a caching issue again? |
Yeah possibly. I think as an approach we DON’T want to call There’s a refMap data structure where we reference which tokens should alias to what. I forget if we always add all tokens to that structure, but it wouldn’t hurt to simply add all tokens as we encounter them there (even if the alias is nothing), and then we’d at least have a complete list of every ID encountered. Anyway doing something more “active” like this would be better performance-wise and all. But also, yes, it does seem like there may be a caching issue, too 🤔. Maybe we can figure that out separately if we can get around calling |
setTransformfails if token not present in default input #707Changes
ensures that
setTransformlooks up tokens on the applied input rather than the default input.