<< Back to previous view

[CLJS-1062] Incorrect deftype/defrecord definition leads to complex error messages Created: 24/Feb/15  Updated: 24/Feb/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Julien Eluard Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: File deftype-fields.diff    

 Description   

Every once in a while I forget to define fields in deftype/defrecord. In this case compilation fails with a low level error message hard to understand.
Clojure compilation prints a better error message in this case: "No fields vector given."

I propose to backport Clojure error message.






[CLJS-1009] Allow deps.cljs to declare a foreign lib as remote Created: 05/Feb/15  Updated: 21/Feb/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Martin Klepsch Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: None


 Description   

When using things like the Google Maps JS API [1] the Javascript that is required can't be bundled inside a jar as it depends on the used API key.

To be able to provide externs for those kind of libraries there should be a way to declare them as "remote" in deps.cljs.

[1] https://developers.google.com/maps/documentation/javascript/tutorial






[CLJS-994] print a warning when :externs file paths can't be found. Created: 30/Jan/15  Updated: 01/Feb/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Crispin Wellington Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: cljs, enhancement, errormsgs, patch,
Environment:

Linux 64bit

java version "1.7.0_65"
OpenJDK Runtime Environment (IcedTea 2.5.3) (7u71-2.5.3-0ubuntu0.14.04.1)
OpenJDK 64-Bit Server VM (build 24.65-b04, mixed mode)


Attachments: Text File clojurescript-extern-missing-warning.patch    
Patch: Code

 Description   

clojurescript silently ignores missing externs files possibly leading a developer to chase their tail.

Presently it can be very confusing using advanced compilation if you have made a mistake in the path name of one of your :externs files. This patch makes the compiler print a warning on stderr so you can quickly determine the cause of the broken advanced compilation output.

As a side effect, when doing a basic lein-cljsbuild a warning is always printed:

```
WARNING: js resource path closure-js/externs does not exist
```

This is because lein-cljsbuild quietly adds this extra path to your :externs listing without you knowing.



 Comments   
Comment by David Nolen [ 31/Jan/15 1:59 PM ]

You need to bind *out* to *err*, or just print to it directly a la cljs.util/debug-prn.

Comment by Crispin Wellington [ 31/Jan/15 7:30 PM ]

I did bind out to err. Check the patch.

Comment by David Nolen [ 01/Feb/15 12:30 PM ]

Crispin, oops sorry you are correct. Thanks.





[CLJS-970] assert - generate Error string when compiling Created: 10/Jan/15  Updated: 12/Jan/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Thomas Heller Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: File cljs-generate-assert-message.diff    
Patch: Code

 Description   

Currently when using assert in CLJS the assert macro will generate the assert error message through CLJS. We can do that in CLJ which reduces the amount of code generated.

(let [bar 1]
  (assert (string? bar)))

Currently generates

if(typeof bar_11332 === 'string'){
} else {
throw (new Error([cljs.core.str("Assert failed: "),cljs.core.str(cljs.core.pr_str.call(null,cljs.core.list(new cljs.core.Symbol(null,"string?","string?",-1129175764,null),new cljs.core.Symbol(null,"bar","bar",254284943,null))))].join('')));
}

But could generate

if(typeof bar_23183 === 'string'){
} else {
throw (new Error("Assert failed: (string? bar)"));
}

Attached patch does that.



 Comments   
Comment by Thomas Heller [ 12/Jan/15 9:16 AM ]

Forgot to account for the fact that

(assert (string? something) "this might not be a string"))

New patch only generates the string of the form when compiling.





[CLJS-895] Hello-js sample is broken when following README instructions Created: 02/Dec/14  Updated: 02/Dec/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Andrew Rosa Assignee: Andrew Rosa
Resolution: Unresolved Votes: 0
Labels: documentation

Attachments: Text File 0001-Create-sample-to-demonstrate-only-extern-usage.patch    

 Description   

Currently we have on hello-js the usage of externes libraries. But it not only mix different aspects of cljsc as also have some breakage if we follow the README instructions

So, instead of adding more complexity into it, here I suggest us to extract the extern functionality into it's own sample project, which is much more approachable for newcomers that want to grok cljsc usage. This new sample project is included in my first patch, so anyone could see the end result.

If this patch is desirable, I could move around the other samples and split them into separate, focused aspect (simple hello, calling cljs from js, fixing the dom sample). Only the twitter buzz that I'm not sure how to approach it - I guess it has issues with API limitations of the last Twitter updates.






[CLJS-889] 2 characters not supported in re-pattern: \u2028 \u2029 Created: 18/Nov/14  Updated: 26/Dec/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Dave Sann Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None
Environment:

[org.clojure/clojurescript "0.0-2371"]
ubuntu


Attachments: Text File 0001-re-pattern-works-on-strings-containing-u2028-or-u202.patch     Text File 889.patch    

 Description   

The following 2 patterns

(re-pattern "[\u2028]")
(re-pattern "[\u2029]")

Cause:

SyntaxError: Invalid regular expression: missing terminating ] for character class

They are probably somewhat rare characters in practice.

http://www.fileformat.info/info/unicode/char/2028/index.htm
http://www.fileformat.info/info/unicode/char/2029/index.htm



 Comments   
Comment by Alex Dowad [ 18/Dec/14 1:51 PM ]

The problem is that re-pattern uses a (.*) regexp to pick out the "pattern" portion of a regexp string (separate from the "flags" portion), and "." doesn't match \u2028 and \u2029.

Comment by Alex Dowad [ 18/Dec/14 2:15 PM ]

Sorry, JIRA's helpful formatting mangled that patch. Here it is again.

Comment by Alex Dowad [ 26/Dec/14 1:31 AM ]

Here is a fixed version of the patch, tested on V8, Nashorn, and Spidermonkey. It includes a regression test.





[CLJS-884] with-redefs don't restores previous values when used inside go block Created: 09/Nov/14  Updated: 09/Nov/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Vladimir Iakovlev Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None
Environment:

clojure 1.6, clojurescript 0.0-2371, core.async 0.1.346.0-17112a-alpha


Attachments: Text File cljs_884.patch    

 Description   

For example I have code and test:

(defn test-fn [] :original)

(deftest ^:async test-with-redefs-async
  (go (is (= (test-fn) :original))
      (with-redefs [test-fn (fn [] :mocked)]
        (is (= (test-fn) :mocked)))
      (is (= (test-fn) :original))
      (done)))

And test failed in last `is`:

FAIL in (test-with-redefs-async) (:)
expected: (= (test-fn) :original)
  actual: (not (= :mocked :original))


 Comments   
Comment by Vladimir Iakovlev [ 09/Nov/14 1:26 PM ]

I attached little patch with fix. With it `with-redefs` forced to use previous value for restoring bindings.

But I don't know how to add test for this issue, because it requires core.async. So I tested with:

(defn test-fn [] :original)

(deftest test-with-redefs-sync
  (is (= (test-fn) :original))
  (with-redefs [test-fn (fn [] :mocked)]
    (is (= (test-fn) :mocked)))
  (is (= (test-fn) :original)))

(deftest ^:async test-with-redefs-async
  (go (is (= (test-fn) :original))
      (with-redefs [test-fn (fn [] :mocked)]
        (is (= (test-fn) :mocked)))
      (is (= (test-fn) :original))
      (done)))




[CLJS-872] Ordered collections printed as if they were unordered at the REPL Created: 13/Oct/14  Updated: 14/Oct/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Michał Marczyk Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None


 Description   

This is due to cljs.repl's read-then-print processing of string representations of return values that come back from the JS env. As of release 2371, the relevant code fragment lives here:

https://github.com/clojure/clojurescript/blob/r2371/src/clj/cljs/repl.clj#L156

A larger snippet to demonstrate this:

ClojureScript:cljs.user> (array-map 1 2 3 4 5 6 7 8 9 10 11 12 13 14)
{1 2, 3 4, 5 6, 7 8, 9 10, 11 12, 13 14}
ClojureScript:cljs.user> (array-map 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18)
{7 8, 1 2, 15 16, 13 14, 17 18, 3 4, 11 12, 9 10, 5 6}
ClojureScript:cljs.user> (seq (array-map 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18))
([1 2] [3 4] [5 6] [7 8] [9 10] [11 12] [13 14] [15 16] [17 18])
ClojureScript:cljs.user> (hash-map 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18)
{7 8, 1 2, 15 16, 13 14, 17 18, 3 4, 11 12, 9 10, 5 6}

This issue seems to be the most likely cause of the problem described in this StackOverflow question. It would be nice to print ordered collections in the "expected" way to prevent user confusion.



 Comments   
Comment by David Nolen [ 14/Oct/14 5:23 AM ]

How is this handled in Clojure?

Comment by Michał Marczyk [ 14/Oct/14 7:40 AM ]

The built-in REPL simply prints string representations of values to *out* without attempting to read them back in first.

Comment by David Nolen [ 14/Oct/14 7:54 AM ]

Well the result of REPL evaluation is a string unlike Clojure, in order to get a proper print at the ClojureScript REPL it seems necessary to read back the result. I will say it's not clear to me array-map promises anything about order anyway.

Comment by Michał Marczyk [ 14/Oct/14 8:02 AM ]

It does – see http://clojure.org/data_structures, where it says

ArrayMaps
When doing code form manipulation it is often desirable to have a map which maintains key order. An array map is such a map - it is simply implemented as an array of key val key val... As such, it has linear lookup performance, and is only suitable for very small maps. It implements the full map interface. New ArrayMaps can be created with the array-map function. Note that an array map will only maintain sort order when un-'modified'. Subsequent assoc-ing will eventually cause it to 'become' a hash-map.

This snippet has been there for as long as I can remember; and this particular feature comes up in various online discussions from time to time. The docstrings for array-map (the core function) are unambiguous in their promise to construct array maps in both Clojure and ClojureScript.

As for the issue at hand, I do think it's a very minor one (hence the "Trivial" priority), but it is real – the representation printed at the REPL is out of line with the string returned from pr-str on the same value. In fact, one way to fix this would be to use pr-str representations computed on the CLJS side. (There may be some complications here, though, which we'd have to work out.)

Comment by David Nolen [ 14/Oct/14 8:07 AM ]

Thanks for pointing out the API promise. OK, yeah I don't really have a good idea for how do this but a patch that isn't too ugly is welcome





[CLJS-844] Optimize js->clj by switching to transducers Created: 22/Aug/14  Updated: 26/Aug/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Darrick Wiebe Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: Text File use-transducers-in-js-to-clj.patch     Text File use-transducers-in-js-to-clj.patch    

 Description   

This simple patch just switches a couple of lines to use transducers. I made this change after finding that using transducers for these types of mappings is significantly faster in (at least in Chrome).

I've checked that this code is being actually tested and both pass under V8 and JavaScriptCore.



 Comments   
Comment by David Nolen [ 23/Aug/14 1:19 PM ]

Did you any benchmarks on other JS engines?

Comment by Darrick Wiebe [ 23/Aug/14 2:14 PM ]

No. My main concern was to check if there was any startup overhead that might offset the basic efficiency improvements for processing small collections. Using into with a transducer turned out to be faster in every scenario, significantly in most.

Comment by David Nolen [ 23/Aug/14 2:19 PM ]

It would be nice to see a jsperf of before and after on this ticket. Thanks.

Comment by Darrick Wiebe [ 23/Aug/14 2:23 PM ]

Is there a existing one that I can work from?

Comment by David Nolen [ 23/Aug/14 2:35 PM ]

There is not. I generally just make a simple project, create an advanced compiled artifact and copy and paste it into jsperf.

Comment by Darrick Wiebe [ 24/Aug/14 7:24 PM ]

Turns out reducing into a transient is considerably better than using a transducer (which was itself a little faster) for this.

http://jsperf.com/js-clj-transducer-test

The code is at:

https://gist.github.com/pangloss/591d77231fda460c2fbe

Let me know if you want me to prepare an updated patch.

Comment by David Nolen [ 25/Aug/14 7:34 AM ]

Thanks for putting this together. Yes please provide an updated patch.

Comment by Darrick Wiebe [ 26/Aug/14 11:19 AM ]

Not sure whether the convention is to comment that I've uploaded a new patch. Regardless, I uploaded it yesterday.





[CLJS-742] Compilation with :output-file option set fails Created: 03/Jan/14  Updated: 08/Oct/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Timothy Pratley Assignee: Timothy Pratley
Resolution: Unresolved Votes: 0
Labels: None

Attachments: File CLJS-742.diff    
Patch: Code

 Description   

./bin/cljsc hello.cljs '{:output-file "foo.js"}'

Expected:
a file called foo.js

Actual:
Compiler throws an internal error

Notes:
The correct option is :output-to
:output-file might be removable (it is broken, nobody is using it?)
Further analysis required to determine if it is useful to keep.



 Comments   
Comment by Timothy Pratley [ 03/Jan/14 12:14 PM ]

Fixes handling of compiling one IJavaScript or many

Comment by Timothy Pratley [ 08/Oct/14 11:03 PM ]

Patch is attached, updating ticket to reflect that





[CLJS-741] seq error message readability is not optimal Created: 02/Jan/14  Updated: 02/Jan/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Julien Eluard Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: Text File cljs_741.patch    

 Description   

When calling seq on an invalid type an errr with following message is thrown: :keywordis not ISeqable.

Adding a space between the argument and the message improves the readability.



 Comments   
Comment by David Nolen [ 02/Jan/14 12:58 PM ]

Excellent thanks!





[CLJS-633] Missing Java type hints in closure.clj Created: 22/Oct/13  Updated: 22/Oct/13

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Julien Eluard Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: Text File cljs_ticket_633.patch    

 Description   

A number of reflection warnings are generated from closure.clj. Fixing those probably would not improved performance but it would be cleaner anyway.



 Comments   
Comment by Julien Eluard [ 22/Oct/13 8:03 AM ]

Attached patch.

There are a couple warnings left, similar to:
Reflection warning, cljs/closure.clj:232:1 - reference to field url can't be resolved.
They appear to be due to some unexpected protocol method name? It looks harmless.





[CLJS-625] apply constructor generates invalid JS Created: 18/Oct/13  Updated: 18/Oct/13

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: George Fraser Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None
Environment:

0.0-1913



 Description   

(apply js/String. "foo")

generates the javscript:

cljs.core.apply.call(null,String.,"foo")

which does not parse. This seems like a bad failure mode.






[CLJS-555] CLONE - Implement ratios Created: 27/Jul/13  Updated: 02/Dec/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Trivial
Reporter: Stuart Halloway Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: None


 Description   

Clojure.java contains support for Ratio types. It would be nice to have them in Clojurescript as well, but as far as I can tell this would be new development (please comment if I'm wrong). That is, there is no implementation of Ratio types in GClosure so this feature would need to be implemented from the ground up. In additional to the Ratio type, the following corresponding functions would also need implementations:

  • `ratio?`
  • `numerator`
  • `denominator`
  • `rationalize`

Plus the ratio type would need to be rolled into the existing mathematical frameworkings.

Imported from github issue #66



 Comments   
Comment by David Nolen [ 02/Dec/14 6:14 AM ]

Lower priority, unsolvable without numerics overhaul.





[CLJS-507] Persistent Data Structure Benchmark Created: 20/May/13  Updated: 02/Dec/14

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Trivial
Reporter: David Nolen Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None


 Description   

Put together an easy to run series of persistent data structure benchmarks that can be easily run. For example we would like to submit to this Mozilla ticket http://bugzilla.mozilla.org/show_bug.cgi?id=874174






[CLJS-450] (ns) within (do) inconsistent with Clojure behaviour Created: 27/Dec/12  Updated: 21/Feb/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Stuart Campbell Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None


 Description   

An (ns) within a (do) doesn't quite work as expected at the REPL:

ClojureScript:cljs.user> (do (ns foo) (def x 42))   
nil
ClojureScript:foo> x
nil
ClojureScript:cljs.user> cljs.user/x
42

The Clojure equivalent:

user=> (do (ns foo) (def x 42))
#'foo/x
foo=> x
42


 Comments   
Comment by David Nolen [ 05/Jan/13 2:05 PM ]

Looks like we need to do something similar to what is done in Clojure with top level do - http://github.com/frenchy64/typed-clojure/pull/4





[CLJS-353] Use different primitives for array access and property/object access Created: 09/Aug/12  Updated: 21/Feb/15

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Raphaël AMIARD Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: Text File 0001-CLJS-353-Use-different-primitives-for-array-access-a.patch     Text File 0001-CLJS-353-Use-different-primitives-for-array-access-a.patch    

 Description   

In javascript, array access (array[index]) and property access (object["property"]) have the same syntax, hence aget/aset are used for both in the core library.

But this isn't true for every target language. For example in Lua, there are no arrays, so if you want to have an array like container, array access will need to go through a function.

This patch proposes to add new builtins : oget and oset (for obj-get obj-set) and use them where appropriate. The generated code will be the same for javascript, but will enable alternate backend implementers to treat both differently



 Comments   
Comment by Raphaël AMIARD [ 09/Aug/12 7:24 AM ]

All test pass

Comment by David Nolen [ 09/Aug/12 5:26 PM ]

Not sure I like the names oget/set. How about obj-get obj-set?

Comment by Raphaël AMIARD [ 10/Aug/12 4:08 AM ]

fine by me, added a new patch !

Comment by David Nolen [ 10/Aug/12 11:07 AM ]

The patches are identical.





[CLJS-288] Compilation of unordered collections Created: 31/May/12  Updated: 23/Sep/12

Status: Open
Project: ClojureScript
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Trivial
Reporter: Brandon Bloom Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None


 Description   

Given:

(defn f [x] (println x) x)

{(f 5) (f 10), (f :x) (f :y)}

Clojure produces:

5
10
:x
:y

{5 10, :x :y}

ClojureScript produces:

5
:x
10
:y

{5 10, :x :y}

 Comments   
Comment by Brandon Bloom [ 16/Aug/12 9:28 PM ]

See also: CLJ-1043

I realized that this problem is actually only partially solvable as is. We could assign the interleaved keys and values to locals before constructing the map. Unfortunately, that doesn't solve a bigger underlying problem: The reader returns unordered sets and maps.

Comment by David Nolen [ 31/Aug/12 9:23 AM ]

Is this actually a problem?

Comment by Brandon Bloom [ 23/Sep/12 7:40 PM ]

See discussion at http://dev.clojure.org/jira/browse/CLJ-1043?focusedCommentId=29526#comment-29526





Generated at Thu Feb 26 22:03:37 CST 2015 using JIRA 4.4#649-r158309.