Paul Zinn-Justin
Paul Zinn-Justin
To clarify, `Symbol`s can be in different dictionaries, but the underlying `Word`s, which carry the parsing information, cannot (and shouldn't precisely because they affect parsing).
Let me try to summarise some of the changes that have been requested during the M2internals meeting: - not let users define *alphanumeric* keywords. though I don't fully agree with...
I think I've dealt with all the issues that were raised: in particular, now alphanumeric keywords are no longer allowed. One pleasant side-effect is the removal of the "math operator...
the checks were successful before, I don't think the ThreadedGB stuff has anything to do with this PR.
Okay. Note that this PR contradicts #3737, and conflicts (not fatally, just some work will be needed) with #3748.
I don't understand any of the comments. seems to be working as intended. you always need to define a keyword before using it. > I think this is better than...
> I was about to merge this, but I realized that all unicode keywords that used to be exported by Core and overloaded by packages are now broken. I suggest...
The only reasonable alternative I can think of is that mathematical symbols are "auto-defined", in the same way as ordinary symbols: the first time they appear, a keyword is created....
I'm still opposed to this change, cf my rejected PR #3599