For others who might have this problem, I've reduced this to the following:
(defrecord A ...)
(instance? A (A. ...)) ;=> true
(instance? A (map->A {...})) ;=> can be false (e.g., in a tomcat servlet)
The map->A constructor uses the create method, and as Chris said earlier,
this seems
to have to do with some deep class loading magic. But this strikes me as a
bug in
the sense that the above should under all circumstances be a guarantee made
by map->A.
On Thursday, December 13, 2012 11:49:20 PM UTC-5, CGAT wrote:
>
> I've now put :aot [ns1] in the projects.clj, did lein compile, then lein
> ring uberwar.
> There was no difference. The class loaders are distinct and instance?
> still returns false.
>
> 1. Did I do the AOT compile incorrectly? Is it even necessary?
> (The class files are in the war file but *all* of them are, not just
> ns1.)
> 2. If not, is it not being ``transmitted'' correctly into the war file?
> 3. Or, am I misunderstanding the problem?
>
> I'm really not clear how to fix this. A record constructed with
> the map->A constructor does not show up as being an instance of A.
> Why?
>
> Help!?
>
> On Thursday, December 13, 2012 11:13:29 PM UTC-5, CGAT wrote:
>>
>> Without AOT, the classLoader's are not the same
>>
>> (str (.getClassLoader A))
>>
>> WebappClassLoader
>> delegate: false
>> repositories:
>> /WEB-INF/classes/
>> ----------> Parent Classloader:
>> org.apache.catalina.loader.StandardClassLoader@790bc49d
>>
>> (str (.getClassLoader (class an-A-record)))
>> clojure.lang.DynamicClassLoader@3ac803e6
>>
>> I'll try AOT compile next, as that DynamicClassLoader seems to suggest.
>> But I'm not sure what else can be done.
>>
>>
>> On Thursday, December 13, 2012 10:53:53 PM UTC-5, CGAT wrote:
>>>
>>> I did not AOT compile it, though I can try that. I'll check the
>>> ClassLoader as well, thanks. I do know that (class an-A-record) is A,
>>> suitably ns-qualified.
>>>
>>> Thanks for the ideas. What did you finally do when this happened to you?
>>>
>>> On Thursday, December 13, 2012 8:19:48 PM UTC-5, cjeris wrote:
>>>>
>>>> In the Tomcat uberwar setting, is ns1 AOT-compiled? Are
>>>> (.getClassLoader A) and (.getClassLoader (.getClass an-A-record))
>>>> different
>>>> class loaders? That's what the problem was when this one bit me. I never
>>>> did track down exactly why it happens or how to fix it.
>>>>
>>>> peace, Chris Jeris
>>>>
>>>> On Thu, Dec 13, 2012 at 6:13 PM, CGAT <[email protected]> wrote:
>>>>
>>>>> I have a frustrating situation and would appreciate any help.
>>>>> The call
>>>>>
>>>>> (instance? A an-A-record)
>>>>>
>>>>> is returning false when an-A-record is in fact an instance of A.
>>>>>
>>>>> This is in a noir/compojure/ring server. But the problem *only* *occurs
>>>>>
>>>>> *when run under tomcat, *not* when run under Jetty. I'm guessing
>>>>> that I've missed declaring something properly, but I don't see what it
>>>>> is.
>>>>> (I hope this is a suitable forum for this question.)
>>>>>
>>>>> Here's the situation in more detail.
>>>>>
>>>>> In ns1
>>>>>
>>>>> (defrecord A ...)
>>>>> (defrecord B ...)
>>>>> (defn AorB? [x] (or (instance? A x) (instance? B x)))
>>>>> (defn Afactory [] ...) ; makes an A
>>>>>
>>>>> In ns2, which :require's ns 1 refering AorB? and Afactory and
>>>>> :imports A and B properly from ns1.
>>>>>
>>>>> (AorB? (Afactory)) ;=> true in jetty, false in tomcat
>>>>>
>>>>> I'm using Clojure 1.4.0, compojure 1.1.3, noir 1.3.0-beta10, ....
>>>>> I use lein run to start the jetty server (with a suitable -main),
>>>>> and lein ring uberwar to create a warfile which runs under
>>>>> Apache Tomcat 6.0.24 (with JVM 1.6.0_26-b03).
>>>>> All the rest of the code appears to be behaving identically under
>>>>> either platform.
>>>>>
>>>>> I've been banging my head against the wall over this and would
>>>>> appreciate any help or pointers. Thanks!
>>>>>
>>>>>
>>>>> --
>>>>> You received this message because you are subscribed to the Google
>>>>> Groups "Clojure" group.
>>>>> To post to this group, send email to [email protected]
>>>>> Note that posts from new members are moderated - please be patient
>>>>> with your first post.
>>>>> To unsubscribe from this group, send email to
>>>>> [email protected]
>>>>> For more options, visit this group at
>>>>> http://groups.google.com/group/clojure?hl=en
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Chris Jeris
>>>> [email protected] (617) 686-3271
>>>> freenode/twitter/github: ystael
>>>>
>>>>
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to [email protected]
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
