(Hmm - NoteToSelf - must remember to not use 'nitpick' in a posting referring to a patch
)
re breaking implied line extension on trailing binary ops.
- ideally/realistically - no
- For the example you give - no
(applies to ASSIGN and EQUALS only)
- generically (as written for this prototype if extended for all suitable binary ops) probably
I coded the trailing test for space or tab (not nl) due to some uncertainty as to exactly what is applied wrt patches for implied line extension...
re patch compile itself:
The patch will compile itself (GREATER_THAN is not one of the binary ops checked in this patch)
The patch does not address any of the failing tests due the patch change which includes the patched compiler building itself...
It doesnt cover forcing this on any other binary op than ASSIGN and EQUAL.
The idea is/was for anyone interested and able to apply the patch, rebuild the compiler and try out the effect allowing some experiential data to the opinions given so far.
- This is NOT a patch to push as is into the source tree - = ?experimental/test/try-out/prototype-feature patch?.
Try out effect and tweak as see fit for whatever edge cases make sense or seem most 'comfortable'
(in my case relax the restriction for this check on ops/expressions inside parens).