On Mon, Jun 24, 2013 at 4:17 PM, Cedric Greevey <cgree...@gmail.com> wrote:
> Ah, thanks. The locking granularity is local to the cons cell (or
> analogue; first/rest pair) being realized, I hope? So one thread actually
> calculates an element and neither call returns until it's calculated, and
> then both promptly return the calculated value, but threads realizing other
> lazy seqs or crawling along earlier parts of the same one don't get blocked?
This is correct, as far as I understand it. One will do the work, the other
will block (awaiting a lock). The first will return, relinquishing the lock
and allow the other to get the computed value too.
The granularity could be anything from a single cons cell to the entire
seq. In the following instance the evaluation of the thunk will yield the
entire vector in one operation, which is synchronized. (The `println` is
there to make clear when/how evaluation takes place.)
(def lazily-initialized-vector (lazy-seq (mapv println (range 10)))
What gets returned by the thunk is entirely dependent on the thunk. LazySeq
just relies on being able to call `seq` on it. Presumably for fully lazy
sequences it will generate one element at a time, and for chunked sequences
it will generate one chunk at a time.
(Just for the record, I don't have any particular knowledge of the Clojure
internals, but I'm fairly confident in my speculation.)
--
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
