News for hackers. Stuff that matters.
Re:Structurally-Typed Condition Handling by gumby on Monday February 14, 2022 @07:54PM (#30341042)
Note to the reader: the article talks about modeling the Common Lisp condition system which is more general than an exception system.

In an exception system, when you signal a condition (“throw”) the stack is unwound until the exception is caught or the stack exhausted.

In a general condition system the stack is searched for a handler; that handler is free to return (at which point the stack is unwound, and then the handler exits, same as catching an exception) or continue (e.g. free some disk space and then have the write operation retry, or return infinity from a divide by zero error).

Re:Structurally-Typed Condition Handling by steveklabnik on Tuesday February 15, 2022 @02:24AM (#30343709)
Fun fact: at one point in Rust's history, it too had conditions. They ended up being removed well before 1.0.
Re:Structurally-Typed Condition Handling by agumonkey on Tuesday February 15, 2022 @02:58AM (#30343875)
was is too costly (memory/instr) or too difficult to model/type properly ?
Re:Structurally-Typed Condition Handling by steveklabnik on Tuesday February 15, 2022 @04:02AM (#30344224)
https://github.com/rust-lang/rust/issues/9795 and the various things linked from it (prs, commits, etc) are the best way to get the context.