<< Back to previous view

[CLJ-1591] Symbol not being bound in namespace when name clashes with clojure.core Created: 14/Nov/14  Updated: 25/Nov/14

Status: Reopened
Project: Clojure
Component/s: None
Affects Version/s: Release 1.6, Release 1.7
Fix Version/s: Release 1.7

Type: Defect Priority: Major
Reporter: Mike Anderson Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: None

Approval: Incomplete

 Description   

The following code fails (both in 1.6 and latest 1.7-alpha4):

user=> (ns foo)
nil
foo=>  (def inc inc)
WARNING: inc already refers to: #'clojure.core/inc in namespace: foo, being replaced by: #'foo/inc
#'foo/inc

;; Note inc is unbound at this point, which causes the exception below
foo=> inc
#<Unbound Unbound: #'foo/inc>
foo=> (ns bar)
nil
bar=> (require ['foo :refer ['inc]])
WARNING: inc already refers to: #'clojure.core/inc in namespace: bar, being replaced by: #'foo/inc
nil
bar=> (inc 8)

IllegalStateException Attempting to call unbound fn: #'foo/inc  clojure.lang.Var$Unbound.throwArity (Var.java:43)

Further investigation shows that foo/inc is unbound:

foo/inc
=> #<Unbound Unbound: #'foo/inc>

Further investigation also shows that replacing the (def inc inc) with almost anything else, e.g. (def inc dec), (def inc clojure.core/inc), or (def inc (fn [n] (+ n 1))), causes no exception (but the warnings remain).

I would expect:
a) foo/inc should be bound and have the same value as clojure.core/inc
b) No error when requiring foo/inc
c) bar/inc should be bound to foo/inc



 Comments   
Comment by Nicola Mometto [ 14/Nov/14 10:04 PM ]

The second error should be expected, the right syntax should be (require ['foo :refer ['inc]]) (note the leading quote before inc)

Comment by Mike Anderson [ 14/Nov/14 10:20 PM ]

Thanks for the catch Nicola - I've edited the description. Still get the same error however (just with a slightly different message)

Comment by Alex Miller [ 14/Nov/14 10:22 PM ]

See comment...

Comment by Mike Anderson [ 14/Nov/14 10:24 PM ]

@Alex what comment? Note that the error still occurs even with the right syntax....

Comment by Mike Anderson [ 14/Nov/14 10:26 PM ]

Appears to have been closed prematurely

Comment by Alex Miller [ 14/Nov/14 10:39 PM ]

I can't reproduce with the correct syntax:

Clojure 1.7.0-master-SNAPSHOT
user=> (ns foo)
nil
foo=> (def inc inc)
WARNING: inc already refers to: #'clojure.core/inc in namespace: foo, being replaced by: #'foo/inc
#'foo/inc
foo=> (ns bar)
nil
bar=> (require ['foo :refer ['inc]])
WARNING: inc already refers to: #'clojure.core/inc in namespace: bar, being replaced by: #'foo/inc
nil
Comment by Mike Anderson [ 14/Nov/14 10:55 PM ]

The problem is that the var is still unbound and causes e.g. the following error:

=> (foo/inc 8)
IllegalStateException Attempting to call unbound fn: #'foo/inc clojure.lang.Var$Unbound.throwArity (Var.java:43)

I don't think that should be expected - or am I missing something?

Comment by Alex Miller [ 14/Nov/14 10:57 PM ]

Ah, will take a look. But not right now.

Comment by Andy Fingerhut [ 15/Nov/14 1:09 PM ]

Updated the description with a few more details. The exception goes away if you do (def inc (fn [n] (+ n 1))) instead of (def inc inc), for example. The warnings remain.

Comment by Tom Crayford [ 20/Nov/14 11:07 AM ]

Unsure if this is the same issue (I think it might be?), but I reproduced the exact same error message with AOT compilation involved:

reproduced in this git repository: https://github.com/yeller/compiler_update_not_referenced_bug

clone it, run `lein do clean, uberjar, test`, and that error message will show up every time for me

Comment by Andy Fingerhut [ 20/Nov/14 5:43 PM ]

Mike, I think replacing (def inc inc) in your example with (def inc clojure.core/inc) should be considered as a reasonable workaround for this issue, unless you have some use case where you need to def inc to something that is not in clojure.core (and if so, why?)

The reason (def inc inc) behaves this way is, if not absolutely necessary, at least commonly used in Clojure programs to define recursive functions, e.g. (defn fib [n] (if (<= n 1) 1 (+ (fib (dec n)) (fib (- n 2))))), so that the occurrences of fib in the body are resolved to the fib being defined.

Comment by Alex Miller [ 22/Nov/14 9:05 AM ]

Moving to 1.7 until I can look at this more deeply.

Comment by Mike Anderson [ 23/Nov/14 6:08 PM ]

Andy - yes the workaround is fine for me right now.

I don't think this is an urgent issue but it may be exposing a subtle complexity regarding assumptions about the state of the namespace at different times. Perhaps the semantics should be something like:

  • The def statement itself should be run before the var is interned. e.g. (def inc (inc 5)) should result in (def inc 6)
  • Anything complied / deferred to run after completion of the def statement should use the new var (i.e. the new var should be referenced by fns, lazy sequences etc.)
Comment by Andy Fingerhut [ 23/Nov/14 6:36 PM ]

I'm not sure what your proposal means in a case like this:

(def inc (fn [x] (inc x)))

Is the second inc to be interpreted/resolved before or after the new inc is created? Because it is (fn ...) it should be the after-behavior? What else besides fn should cause the after-behavior, rather than the before-behavior?

Even more fun (not saying that people often write code like this, but the compiler can handle it today):

(def inc (if (> (inc y) 5)
           (fn [x] (inc x))
           (fn [x] (dec x))))

I think the current compiler behavior of 'in the body of a def, the def'd symbol always refers to the new var, not any earlier def'd vars' is fairly straightforward to explain.

Comment by Tom Crayford [ 23/Nov/14 9:15 PM ]

Should I file the AOT issue reproduced in that thing as a new issue?

Comment by Andy Fingerhut [ 24/Nov/14 5:16 PM ]

Tom: Alex Miller or another screener would be best to say whether the AOT issue should be a separate ticket, but my best guess would be "go for it". I tried to look at the link you gave but it seems not to point to anything. Could you double-check that link?

Comment by Tom Crayford [ 24/Nov/14 6:48 PM ]

Andy,

Great. I'll write one up tomorrow sometime. I accidentally left that repo as private, should be visible now.

Comment by Andy Fingerhut [ 24/Nov/14 8:11 PM ]

Tom, looked at your project. Thanks for that. It appears not to have anything like (def inc inc) in it. It throws exception during test step of 'lein do clean, uberjar, test' consistently for me, too, but compiles with only warnings and passes tests with 'lein do clean, test'. I have more test results showing in which Clojure versions these results change. To summarize, the changes to Clojure that appear to make the biggest difference in the results are below (these should be added to the new ticket you create – you are welcome to do so):

Clojure 1.6.0, 1.7.0-alpha1, and later changes up through the commit with description "CLJ-1378: Allows FnExpr to override its reported class with a type hint": No errors or warnings for either lein command above.

Next commit with description "Add clojure.core/update, like update-in but takes a single key" that adds clojure.core/update: 'lein do clean, test' is fine, but 'lein do clean, uberjar' throws exception during compilation, probably due to CLJ-1241.

Next commit with description "fix CLJ-1241": 'lein do clean, test' and 'lein do clean, uberjar' give warnings about clojure.core/update, but no errors or exceptions. 'lein do clean, uberjar, test' throws exception during test step that is same as the one I see with Clojure 1.7.0-alpha4. Debug prints of values of clojure.core/update and int-map/update (in data.int-map and in Tom's namespace compiler-update-not-referenced-bug.core) show things look fine when printed inside data.int-map, and in Tom's namespace when not doing the uberjar, but when doing the uberjar, test, int-map/update is unbound in Tom's namespace.

In case it makes a difference, my testing was done with Mac OS X 10.9.5, Leiningen 2.5.0 on Java 1.7.0_45 Java HotSpot(TM) 64-Bit Server VM

Comment by Nicola Mometto [ 25/Nov/14 3:44 PM ]

Tom, I've opened a ticket with a patch fixing the AOT issue: http://dev.clojure.org/jira/browse/CLJ-1604





[CLJ-1580] Transient collections should guarantee thread visibility Created: 05/Nov/14  Updated: 25/Nov/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.7
Fix Version/s: Release 1.7

Type: Defect Priority: Major
Reporter: Alex Miller Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: transient

Approval: Incomplete

 Description   

With changes from CLJ-1498, transients are still thread isolated but may move between threads during their lifetime which introduces new concurrency concerns, namely visibility of changes across threads.






[CLJ-1515] Reify the result of range Created: 29/Aug/14  Updated: 14/Nov/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.7
Fix Version/s: Release 1.7

Type: Enhancement Priority: Major
Reporter: Timothy Baldridge Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: None

Attachments: File patch.diff     File range-patch3.diff     File reified-range4.diff    
Patch: Code and Test
Approval: Incomplete

 Description   

Currently range simply returns a lazy seq. If the return value of range were reified into a type (as it is in ClojureScript) we could optimize many functions on that resulting type. Some operations such as count and nth become O(1) in this case, while others such as reduce could receive a performance boost do to the reduced number of allocations.

Approach: this patch revives the unused (but previously existing) clojure.lang.Range class. This class acts as a lazy seq and implements several other appropriate interfaces such as Counted and Indexed. This type is implemented in Java since range is needed fairly on in core.clj before deftype is defined. The attached patch uses Numbers.* methods for all math due to the input types to range being unknown. The class also supplies a .iterator() method which allows for allocation free reducing over range.

Note: this code keeps backwards compatibility with the existing range code. This means some parts of the class (mostly relating to a step size of 0) are a bit more complex than desired, but these bits were needed to get all the tests to pass.

Note: this code does not preserve the chunked-seq nature of the original range. The fact that range used to return chunked seqs was not published in the doc strings and so it was removed to allow for simpler code in Range.java.

Performance:
(timings done at the repl run via java -jar)

(dotimes [x 100] (time (dotimes [x 1] (count (range (* 1024 1024))))))
master => 80-110ms
patch => 0.014ms


(dotimes [x 100] (time (dotimes [x 1] (reduce + (map inc (range (* 1024 1024)))))))
master => 76-87ms
patch => 340-360ms


(dotimes [x 100] (time (dotimes [x 1] (reduce + (map inc (map inc (range (* 1024 1024))))))))
master => 97-123ms
patch=> 490-577ms



(dotimes [x 100] (time (dotimes [x 1] (count (filter odd? (range (* 1024 1024)))))))
master => 87-104ms
patch => 370-330ms


(dotimes [x 100] (time (dotimes [x 1] (transduce (comp (map inc) (map inc)) + (range (* 1024 1024))))))
master=>76-116ms
patch => 44ms-59ms

Patch: reified-range4.diff (some tests fail)



 Comments   
Comment by Alex Miller [ 29/Aug/14 3:19 PM ]

1) Not sure about losing chunked seqs - that would make older usage slower, which seems undesirable.
2) RangeIterator.next() needs to throw NoSuchElementException when walking off the end
3) I think Range should implement IReduce instead of relying on support for CollReduce via Iterable.
4) Should let _hash and _hasheq auto-initialize to 0 not set to -1. As is, I think _hasheq always would be -1?
5) _hash and _hasheq should be transient.
6) count could be cached (like hash and hasheq). Not sure if it's worth doing that but seems like a win any time it's called more than once.
7) Why the change in test/clojure/test_clojure/serialization.clj ?
8) Can you squash into a single commit?

Comment by Timothy Baldridge [ 29/Aug/14 3:40 PM ]

1) I agree, adding chunked seqs to this will dramatically increase complexity, are we sure we want this?
2) exception added
3) I can add IReduce, but it'll pretty much just duplicate the code in protocols.clj. If we're sure we want that I'll add it too.
4) fixed hash init values, defaults to -1 like ASeq
5) hash fields are now transient
6) at the cost of about 4 bytes we can cache the cost of a multiplication and an addition, doesn't seem worth it?
7) the tests in serialization.clj assert that the type of the collection roundtrips. This is no longer the case for range which starts as Range and ends as a list. The change I made converts range into a list so that it properly roundtrips. My assumption is that we shouldn't rely on all implementations of ISeq to properly roundtrip through EDN.
8) squashed.

Comment by Alex Miller [ 29/Aug/14 3:49 PM ]

6) might be useful if you're walking through it with nth, which hits count everytime, but doubt that's common
7) yep, reasonable

Comment by Andy Fingerhut [ 18/Sep/14 6:52 AM ]

I have already pointed out to Edipo in personal email the guidelines on what labels to use for Clojure JIRA tickets here: http://dev.clojure.org/display/community/Creating+Tickets

Comment by Timothy Baldridge [ 19/Sep/14 10:02 AM ]

New patch with IReduce directly on Range instead of relying on iterators

Comment by Alex Miller [ 01/Oct/14 2:00 PM ]

The new patch looks good. Could you do a test to determine the perf difference from walking the old chunked seq vs the new version? If the perf diff is negligible, I think we can leave as is.

Another idea: would it make sense to have a specialized RangeLong for the (very common) case where start, end, and step could all be primitive longs? Seems like this could help noticeably.

Comment by Timothy Baldridge [ 03/Oct/14 10:00 AM ]

Looks like chunked seqs do make lazy seq code about 5x faster in these tests.

Comment by Ghadi Shayban [ 03/Oct/14 10:22 AM ]

I think penalizing existing code possibly 5x is a hard cost to stomach. Is there another approach where a protocolized range can live outside of core? CLJ-993 has a patch that makes it a reducible source in clojure.core.reducers, but it's coll-reduce not IReduce, and doesn't contain an Iterator. Otherwise we might have to take the chunked seq challenge.

Alex: Re long/float. Old reified Ranged.java in clojure.lang blindly assumes ints, it would be nice to have a long vs. float version, though I believe the contract of reduce boxes numbers. (Unboxed math can be implemented very nicely as in Prismatic's Hiphip array manipulation library, which takes the long vs float specialization to the extreme with different namespaces)

Comment by Timothy Baldridge [ 03/Oct/14 10:38 AM ]

I don't think anyone is suggesting we push unboxed math all the way down through transducers. Instead, this patch contains a lot of calls to Numbers.*, if we were to assume that the start end and step params of range are all Longs, then we could remove all of these calls and only box when returning an Object (in .first) or when calling IFn.invoke (inside .reduce)

Comment by Alex Miller [ 03/Oct/14 10:46 AM ]

I agree that 5x slowdown is too much - I don't think we can give up chunked seqs if that's the penalty.

On the long case, I was suggesting what Tim is talking about, in the case of all longs, create a Range that stores long prims and does prim math, but still return boxed objects as necessary. I think the only case worth optimizing is all longs - the permutation of other options gets out of hand quickly.

Comment by Ghadi Shayban [ 03/Oct/14 11:00 AM ]

Tim, I'm not suggesting unboxed math, but the singular fast-path of all-Longs that you and Alex describe. I mistakenly lower-cased Long/Float.

Comment by Timothy Baldridge [ 31/Oct/14 11:30 AM ]

Here's the latest work on this, a few tests fail. If someone wants to take a look at this patch feel free, otherwise I'll continue to work on it as I have time/energy.

Comment by Nicola Mometto [ 14/Nov/14 12:51 PM ]

As discussed with Tim in #clojure, the current patch should not change ArrayChunk's reduce impl, that's an error.





[CLJ-1424] Feature Expressions Created: 15/May/14  Updated: 07/Nov/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: Release 1.7

Type: Enhancement Priority: Major
Reporter: Ghadi Shayban Assignee: Alex Miller
Resolution: Unresolved Votes: 0
Labels: reader

Attachments: File CLJ-1424-2.diff     File clj-1424-3.diff     File clojure-feature-expressions.diff    
Patch: Code
Approval: Incomplete

 Description   

Feature expressions based directly on Common Lisp. See Clojure design docs, which includes discussion and links to Common Lisp documentation for feature expressions here: http://dev.clojure.org/display/design/Feature+Expressions

#+ #- and or not
are supported. Unreadable tagged literals are suppressed through the *suppress-read* dynamic var. For example, with *features* being #{:clj}, which is the default, the following should read :foo

#+cljs #js {:one :two} :foo

The initial *features* set can be augmented (clj will always be included) with the clojure.features System property:

-Dclojure.features=production,embedded

Patch: clj-1424-3.diff

Questions: Should *suppress-read* override *read-eval*?

Related: CLJS-27, TRDR-14



 Comments   
Comment by Jozef Wagner [ 16/May/14 2:19 AM ]

Has there been a decision that CL syntax is going to be used? Related discussion can be found at design page, google groups discussion and another discussion.

Comment by Alex Miller [ 16/May/14 8:34 AM ]

No, no decisions on anything yet.

Comment by Ghadi Shayban [ 19/May/14 7:25 PM ]

Just to echo a comment from TRDR-14:

This is WIP and just one approach for feature expressions. There seem to be at least two couple diverging approaches emerging from the various discussion (Brandon Bloom's idea of read-time splicing being the other.)

In any case having all Clojure platforms be ready for the change is probably essential. Also backwards compatibility of feature expr code to Clojure 1.6 and below is also not trivial.

Comment by Kevin Downey [ 04/Aug/14 1:39 PM ]

if you have ever tried to do tooling for a language where the "parser" tossed out information or did some partial evaluation, it is a pain. this is basically what the #+cljs style feature expressions and bbloom's read time splicing both do with clojure's reader. I think resolving this at read time instead of having the compiler do it before macro expansion is a huge mistake and makes the reader much less useful for reading code.

Comment by Ghadi Shayban [ 04/Aug/14 2:00 PM ]

Kevin, what kind of tooling use case are you alluding to?

Comment by Kevin Downey [ 04/Aug/14 3:24 PM ]

any use case that involves reading code and not immediately handing it off to the compiler. if I wanted to write a little snippet to read in a function, add an unused argument to every arity then pprint it back, reader resolved feature expressions would not round trip.

if I want to write snippet of code to generate all the methods for a deftype (not a macro, just at the repl write a `for` expression) I can generate a clojure data structure, call pprint on it, then paste it in as code, reader feature expressions don't have a representation as data so I cannot do that, I would have to generate strings directly.

Comment by Alex Miller [ 22/Aug/14 9:10 AM ]

Changing Patch setting so this is not in Screenable yet (as it's still a wip).

Comment by Alex Miller [ 07/Nov/14 4:39 PM ]

Latest patch brings up to par with related patches in CLJS-27 and TRDR-14 and importantly adds support for loading .cljc files as Clojure files.

Comment by Andy Fingerhut [ 07/Nov/14 5:55 PM ]

Maybe undesirable behavior demonstrated below with latest Clojure master plus patch clj-1424-3.diff, due to the #+cljs skipping the comment, but not the (dec a). I thought it could be fixed simply by moving RT.suppressRead() check after (ret == r) check in read(), but that isn't correct.

user=> (read-string "(defn foo [a] #+clj (inc a) #+cljs ; foo\n (dec a))")
(defn foo [a] (inc a) (dec a))




Generated at Wed Nov 26 23:54:15 CST 2014 using JIRA 4.4#649-r158309.