[ruby-core:109716] [Ruby master Feature#18408] Allow pattern match to set instance variables
From:
"Dan0042 (Daniel DeLorme)" <noreply@...>
Date:
2022-08-26 17:31:36 UTC
List:
ruby-core #109716
Issue #18408 has been updated by Dan0042 (Daniel DeLorme). > it is intentional. Unsuccessful matches remain the assignment result from the internal matches I think everyone here understood the current behavior is intentional, but the more important question would be "is it beneficial?" > Since instance variables (and global variables) can be accessed from outside, pattern matches can break the object status. If we want the object state to stay consistent across threads... well, for one it would require a mutex around every ivar access so IMHO this is not such a relevant concern. But if you don't want ivars to change during match, this *can* be fixed, for example by using temporary local vars like I suggested in #note-11. I don't think this is a fundamental problem with this syntax, although it increases the cost/benefit ratio. > In addition, once we allow instance variables, the request for array reference (a[1]) or attribute acess (obj.attr) would come and things would get more and more complex. In other words it's an inevitable slippery slope? I'm sure there *would* be a request for array reference or attribute access, but that doesn't mean it has to be accepted. There is no reason why one inevitably implies the other, just as `42 => v` doesn't imply the acceptance of `42 => @v`. > So at the moment, we will decide not to support anything but local variables. If the proposal is rejected then that's how it is, but I would honestly appreciate if that rejection came from the suitability of the `42 => @v` syntax itself. It's disappointing to hear a rejection simply based on nitpicks of the implemention. It's the difference between "42=>@v is a bad syntax" and "42=>@v would be acceptable *if* someone is able to address issues A,B,C" ---------------------------------------- Feature #18408: Allow pattern match to set instance variables https://bugs.ruby-lang.org/issues/18408#change-98950 * Author: Dan0042 (Daniel DeLorme) * Status: Rejected * Priority: Normal * Assignee: ktsj (Kazuki Tsujimoto) ---------------------------------------- I expected this to work: ```ruby 42 => @v ``` But instead it raises "syntax error, unexpected instance variable" Is this intentional? -- https://bugs.ruby-lang.org/ Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe> <http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>