Todd Ditchendorf

@iTod Keep related blocks of code together.

@mjtsai So it is a syntactical difference. Thanks. (AFAICT, this is the only significant difference besides cultural).

@iTod Yes, mainly the syntax is much more convenient and readable. And there is the codegen difference because it’s not based on exceptions.

@mjtsai Hm, you're fist to mention that. Is `try/finally` (with no `catch`) significantly different from `defer` from a codegen perspective?

@iTod Haven't looked at the assembly, but I would assume yes because try/finally has to essentially catch and rethrow the exception.

@mjtsai Hm, that's interesting. So they're only different if an exception is thrown. But in non-exceptional case, it should be same as defer

@iTod Yes, if we're talking about runtime behavior, and we have zero-cost exceptions. But I thought we were talking about codegen.

.@iTod defer doesn't have to be used with try-catch; put your cleanup code right by the initialization. e.g. file open/close.

@owensd I'm pretty sure you misunderstood me. I said `try/finally`, not `try/catch` nor `try/catch/finally`.

.@iTod that's what's better: there is only one cleanup mechanism. Promotion to a throwing function doesn't require moving code around.

@owensd And as far as 'one cleanup mechanism' I'd say `try/finally` is even less conceptual overhead than `defer`, not to mention familiar.

@iTod also, finally happens in order, defer happens on scope exit.

@owensd I don't grok the difference there. Seems like both offer fine-grained ordering control. Maybe `finally` even finer control?

@iTod @joconor it lets you associate cleanup work with setup work. Similar to C89 vs C99 variable declaration.

@Catfish_Man @joconor So the only diff is syntactic one? The cleanup code goes adjacent to code of your choice rather than after curly block

@iTod @joconor there are other minor differences, but really I think @finally is very under-used, probably due to association w/ exceptions.

@Catfish_Man @joconor Ok, so syntactic plus cultural difference.

@iTod @joconor yeah… now, why association with exceptions scares people away is a much more interesting kettle of fish :)

@Catfish_Man @iTod Because Apple has told people for decades that exceptions are only for fatal situations?