Re: Understanding vars
Opened ticket with fix + test http://dev.clojure.org/jira/browse/CLJ-1187 Mark Engelberg writes: On Tue, Mar 19, 2013 at 12:57 AM, Bronsa brobro...@gmail.com wrote: If I remember correctly, this is a bug due to the fact that constant empty literals are handled in a special way from the compiler. Interesting. I see you are correct that the problem only occurs on metadata attached to an empty literal. So does that mean this is a known bug? -- -- -- 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.
Re: Understanding vars
On Tuesday, March 19, 2013 4:41:33 AM UTC+1, puzzler wrote:
On Mon, Mar 18, 2013 at 8:31 PM, Kemar ke.m...@gmail.com javascript:wrote:
Explicitly derefing the var and calling meta on it works:
(meta @#'c) - {:amazing true}
No idea as to why though...
Quirky. I assume that explicitly derefing the var defeats the speed
benefits of ^:const, yes?
The way speed is achieved for :const is that it is given the same treatment
as Java's *compile-time constants*, so you're not even touching the var
when you refer to it by name. Now, *meta* could be accepted as a special
case, explicitly detected by the compiler, but that mechanism would once
again be easily defeated. So I'd say it's not a bug, but a case of *you
can't have your cake and eat it, too.*
*
*
-marko
--
--
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.
Re: Understanding vars
The way speed is achieved for :const is that it is given the same treatment as Java's *compile-time constants*, so you're not even touching the var when you refer to it by name. Now, *meta* could be accepted as a special case, explicitly detected by the compiler, but that mechanism would once again be easily defeated. So I'd say it's not a bug, but a case of *you can't have your cake and eat it, too.* On second thought, I my be wrong since the meta is not on the var, but on the map object itself. Maybe it really is just a bug due to which the compiled-in value doesn't have metadata copied to it. -marko -- -- 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.
Re: Understanding vars
If I remember correctly, this is a bug due to the fact that constant empty literals are handled in a special way from the compiler. Il giorno 19/mar/2013 08.49, Marko Topolnik marko.topol...@gmail.com ha scritto: The way speed is achieved for :const is that it is given the same treatment as Java's *compile-time constants*, so you're not even touching the var when you refer to it by name. Now, *meta* could be accepted as a special case, explicitly detected by the compiler, but that mechanism would once again be easily defeated. So I'd say it's not a bug, but a case of *you can't have your cake and eat it, too.* On second thought, I my be wrong since the meta is not on the var, but on the map object itself. Maybe it really is just a bug due to which the compiled-in value doesn't have metadata copied to it. -marko -- -- 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. -- -- 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.
Re: Understanding vars
Indeed:
(def ecr (with-meta [1] {:amazing true}))
(def ^:const c ecr)
(meta ecr) - {:amazing true}
(meta c) - {:amazing true}
On Tuesday, March 19, 2013 8:57:28 AM UTC+1, Nicola Mometto wrote:
If I remember correctly, this is a bug due to the fact that constant empty
literals are handled in a special way from the compiler.
Il giorno 19/mar/2013 08.49, Marko Topolnik
marko.t...@gmail.comjavascript:
ha scritto:
The way speed is achieved for :const is that it is given the same
treatment as Java's *compile-time constants*, so you're not even
touching the var when you refer to it by name. Now, *meta* could be
accepted as a special case, explicitly detected by the compiler, but that
mechanism would once again be easily defeated. So I'd say it's not a bug,
but a case of *you can't have your cake and eat it, too.*
On second thought, I my be wrong since the meta is not on the var, but on
the map object itself. Maybe it really is just a bug due to which the
compiled-in value doesn't have metadata copied to it.
-marko
--
--
You received this message because you are subscribed to the Google
Groups Clojure group.
To post to this group, send email to clo...@googlegroups.comjavascript:
Note that posts from new members are moderated - please be patient with
your first post.
To unsubscribe from this group, send email to
clojure+u...@googlegroups.com javascript:
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+u...@googlegroups.com javascript:.
For more options, visit https://groups.google.com/groups/opt_out.
--
--
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.
Re: Understanding vars
On Tue, Mar 19, 2013 at 12:57 AM, Bronsa brobro...@gmail.com wrote: If I remember correctly, this is a bug due to the fact that constant empty literals are handled in a special way from the compiler. Interesting. I see you are correct that the problem only occurs on metadata attached to an empty literal. So does that mean this is a known bug? -- -- 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.
Re: Understanding vars
I remember finding out about it a few months ago. I don't know whether or not there is a jira ticket for it, I'll check out. If nobody can get a patch for this, I'll try and see if I can work out one in the next days Il giorno 19/mar/2013 09.02, Mark Engelberg mark.engelb...@gmail.com ha scritto: On Tue, Mar 19, 2013 at 12:57 AM, Bronsa brobro...@gmail.com wrote: If I remember correctly, this is a bug due to the fact that constant empty literals are handled in a special way from the compiler. Interesting. I see you are correct that the problem only occurs on metadata attached to an empty literal. So does that mean this is a known bug? -- -- 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. -- -- 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.
Re: Understanding vars
On Monday, March 18, 2013 7:16:17 PM UTC+1, puzzler wrote: It used to be that a common way to refactor this for best performance would be: (let [pi (compute-estimate-of-pi)] (defn f [x] ; Do stuff with pi here )) This is just about the same as (def ^:const pi (compute-estimate-of-pi)) (defn f [x] ...) Otherwise, for a regular static var you pay AtomicBoolean.get + volatile read of the actual value. For a dynamic var you pay much more (ThreadLocal.get() + hashmap lookup). -marko -- -- 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.
Re: Understanding vars
On Mon, Mar 18, 2013 at 12:38 PM, Marko Topolnik
marko.topol...@gmail.comwrote:
This is just about the same as
(def ^:const pi (compute-estimate-of-pi))
OK, just tried out ^:const in my code and it seems to strip off any meta
data associated with the object stored in the var. That ruins its utility
for my purposes.
(def ecr (with-meta [] {:amazing true}))
(def ^:const c ecr)
(meta ecr) - {:amazing true}
(meta c) - nil
Bug?
--
--
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.
Re: Understanding vars
Explicitly derefing the var and calling meta on it works:
(meta @#'c) - {:amazing true}
No idea as to why though...
Am Dienstag, 19. März 2013 03:45:57 UTC+1 schrieb puzzler:
On Mon, Mar 18, 2013 at 12:38 PM, Marko Topolnik
marko.t...@gmail.comjavascript:
wrote:
This is just about the same as
(def ^:const pi (compute-estimate-of-pi))
OK, just tried out ^:const in my code and it seems to strip off any meta
data associated with the object stored in the var. That ruins its utility
for my purposes.
(def ecr (with-meta [] {:amazing true}))
(def ^:const c ecr)
(meta ecr) - {:amazing true}
(meta c) - nil
Bug?
--
--
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.
Re: Understanding vars
On Mon, Mar 18, 2013 at 8:31 PM, Kemar ke.mar...@gmail.com wrote:
Explicitly derefing the var and calling meta on it works:
(meta @#'c) - {:amazing true}
No idea as to why though...
Quirky. I assume that explicitly derefing the var defeats the speed
benefits of ^:const, yes?
--
--
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.
