midje-cascalog produces wrong output from basic queries

[timothypratley]

using [cascalog/midje-cascalog "2.1.1"]

(fact
 (<- [?n]
     (num-pair :> ?n ?n))
 => (produces [[0] [1] [4]]]))
;; false, produces [[4] [4] [4]]

(?<- (stdout) [?n]
     (num-pair :> ?n ?n))
;; prints 0 1 4

Also, the tests for midje-cascalog do not run:

✘-1 ~/git/cascalog/midje-cascalog [develop|✔]
17:03 $ lein midje

-> Exception

[sritchie]

Well, doesn't the first clause say that it produces a single 3-tuple?

—
Sent from Mailbox

On Wed, Sep 30, 2015 at 5:47 PM, Timothy Pratley notifications@github.com
wrote:

using [cascalog/midje-cascalog "2.1.1"]

(fact
 (<- [?n]
     (num-pair :> ?n ?n))
 => (produces [[0 1 4]]))
;; false, produces [[4] [4] [4]]
(?<- (stdout) [?n]
     (num-pair :> ?n ?n))
;; prints 0 1 4

---

Reply to this email directly or view it on GitHub:
#292

[timothypratley]

@sritchie I am expecting [[0] [1] [4]] but get [[4] [4] [4]] instead... does that clarify?
num-pair is from cascalog.playground, there are pairs of 0s 1s and 4s in the num-pairs data.

[sritchie]

I think The library wraps the tests in a config that has io.sort.mb equal to 1. Wonder if that's what's going on.

—
Sent from Mailbox

On Wed, Sep 30, 2015 at 7:43 PM, Timothy Pratley notifications@github.com
wrote:

@sritchie I am expecting [[0] [1] [4]] but get [[4] [4] [4]] instead... does that clarify?

num-pair is from cascalog.playground, there are pairs of 0s 1s and 4s in the num-pairs data.

Reply to this email directly or view it on GitHub:
#292 (comment)

[danielstjules]

Just wondering if there's been any update on this? Is there a last known good version we could jump to? I've tried a few with no luck.

[danielstjules]

Turns out it's only an issue with [org.clojure/clojure "1.7.0"]. Things work fine with [org.clojure/clojure "1.6.0"] :)

[sritchie]

Nice, this is good to know. I don't have any time to work on this, but happy to help guide if you guys want take a stab at it and have any questions about the code.

[jiyouyou125]

@sritchie add a test, test success. it seems related iterator-seq. #298 add Implicit equality constraints test,related #292