* jlforr...@berkeley.edu <jlforr...@berkeley.edu> [190329 15:28]: > Thanks everyone for responding! > > However, in the case that I illustrated there is no such problem. I > would have thought that the language would allow this construct with > any expression on the right side, as long as it only has one map > access. > > This isn't a big deal since it's so easy to work around, but it > surprised me as a newcomer to Go.
I'm glad you got it figured out! This may help you next time you run into something similar: Go is a strongly typed language, and its syntax consists of statements and expressions. Each expression only looks at its operands, and does not depend on what complex sub-expression was evaluated to obtain that operand. Likewise, a statement only looks at the very outer layer of its expressions. So, for an assignment such as a, ok = foo(x[i]).y[z] the compiler sees a potential assignment statement. An assignment statement must have a list of addressable expressions and/or map index expressions on the left, which this does. The expression on the right must either be multi-valued, with two values, or it must be a map indexing expression. The only way this can be a map indexing expression is if «foo(x[i]).y» evaluates to a map. The Go syntax never looks deeper than this. In fact, the Go specification calls out this particular statement syntax as being a "special form". It only applies if the RHS _is_ a map indexing expression, not if the RHS _contains_ a map indexing expression. Hope this helps! ...Marvin -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
