I've been using ClojureScript rather successfully for a year now, but
ran into an interesting issue recently.
I've discovered that js->clj isn't quite working as expected. After
quite a bit of tearing things down and apart, I discovered it was one
of the keys in some JSON data coming back from the server that looks
like:
"Bi": 8
And the resultant error would be:
No protocol method IEncodeClojure.-js->clj defined for type
object: [object Object]
I don't know why this particular sequence caused a problem, but I
ended trying to vary things a bit just to see what happens. If I use:
"Bi": 0
I don't see the issue. Any non-zero value though does seem to bring
the problem about for this sequence of letters though. Similarly, I
tried varying the two letter sequence, and discovered two others with
similar problems: Ai, and ba. The latter brought about a different
error:
[object Object] is not ISeqable
And the pattern with the value needing to be non-zero follows. If the
value associated with the key is 0, then I don't see a problem.
As I wasn't seeing this issue before, I backpedaled my version of the
ClojureScript compiler and eventually found that 1.9.56 didn't seem to
sufffer from the problem. After checking out ClojureScript, and
bisecting more I found that commit
bcf60ce194e5292fbc5c4b2d89dfc5a7b886b94c is when the behavior changed
from behaving correctly to behaving badly, but I think it's a false
positive.
I then created a much smaller test case, with no other libraries
involved. This time it's a different set of values that are broken:
* "Ta" with a non-zero value gives: [object Object] is not ISeqable
* "Ub" with a non-zero value gives: No protocol method
IEmptyableCollection.-empty defined for type object: [object Object]
* "Wb" with a non-zero value gives: No protocol method
IEncodeClojure.-js->clj defined for type object: [object Object]
If I back up to before the previously mentioned commit, I can still
cause the problem, though the letter sequences change a bit.
Also, this only appears to happen with advanced optimizations. With
no optimizations or with whitespace optimizations, there's no issues.
So it smells like a strange interaction between the ClojureScript
runtime and the compiler.
Has anyone seen an issue like this before? Does anyone have any ideas
what is going on? We have these short keys in our data, and it'd be
pretty painful to make them something else (especially since the names
are meaningful to those in the know, despite them being cryptic--they
refer to bit settings on various motors).
I'd be happy to file a ticket, if that's what needs to get done too.
Thanks!
-John
PS Here's the code I used to hunt down problematic keys:
(def alphabet [\a \b \c \d \e \f \g \h \i \j \k \l \m \n \o \p \q \r \s \t \u
\v \w \x \y \z \A \B \C \D \E \F \G \H \I \J \K \L \M \N \O \P
\Q \R \S \T \U \V \W \X \Y \Z])
(enable-console-print!)
(defn ^:export main-page []
(println"Starting...")
(doseq [c1 alphabet
c2 alphabet
i (range 256)]
(let [s (str "{\"" c1 c2 "\": " i "}")]
(try
;; (prn :parse-s (js/JSON.parse s))
;; (prn :js-clj-parse (js->clj (js/JSON.parse s)))
(js->clj (js/JSON.parse s))
(catch js/Error e
(println "didn't work" s (str e))))
))
(println "Done!"))
Then you can observe the results from the JS console.
--
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/d/optout.
