andrew blinn

@cyrus not sure what you mean, that will just break the annotated-with-type-hole case... that's the case that needs the InconsistentBranches(..., Ana) to avoid impossible-syn

my understanding: 1. if we're analyzing against SynPatternVar, we just switch to synthetic mode; everything just works as it did in the classic bidi pre-pattern annotation world. 2. if we're...

@cyrus cursorinfo also seems wonky now:

@cyrus: made the changed we talked, in particular moved the SynPatternVar check to the top of ana. A couple issues: 1. If I put in the check for SynPatternVar in...

@cyrus ah i figured out (2), kind of... so this is the problematic expression: ``` ((ExpLine (OpSeq (Placeholder 0) (S (Case (InconsistentBranches 22 Ana) ((ExpLine (OpSeq (Placeholder 0) (S (EmptyHole...

For (1): syn_fix_holes is being called on the examples to generate an edit state from a zexp; this is what's causing the immediate crash. If I bypass that though there...

(2) okay changed it to only do the check in the Syn case

(1) 80 printlns later: it's not enough to add checks in the root ana function. for example if the pattern is a pair (x,x) then the type won't be synpatternvar...

Added the SynPatternVar check to all ana exp judgements in statics, elaborator, action, and cursor_info. still getting consistent check fails from the Examples, for example from here :

Also: