[Issue #82] updated backlog; organized t-formula

FORMULAS-BACKLOG-AND-FEATURE-IDEAS.md

@@ -9,26 +9,21 @@
 * [x] figure out how to make syntax validation errors show something more sensible than the 
       error message you'd see for a problem with a fact validation problem 
 
-* [ ] don't run more tests than need be if there is already failure in this formula's batch.
-
-* [ ] make line numbers report correctly
+* [x] don't run more tests than need be if there is already failure in this formula's batch.
 
 * [ ] implement shrinking. Report only the first fully shrunken failure
-         [  ] to do correctly, 'shrink' depends on 'generate'
-
-* [ ] consider implementing with @marick's metaconstant syntax
+         [ ] to do correctly, 'shrink' depends on 'generate'
 
 * [ ] shrink function overrides
    a. [ ] no shrink
-
-* [ ] if we do metaconstant style, implement generator overriding
-
-    --- getting there with the new validate method for "formula"
-      a. [] fix:  however, there's an interesting form of failure if you run (formula [a 1] 1 =>)
-          ... since the formula macro splices in :formula :formula-in-progress
-
-          will the below note help this?
+
+* [ ] fix strange error if you run (formula [a 1] 1 =>)
+      ... since the formula macro splices in :formula :formula-in-progress
+      possibly solution is to not using fact macro inside of formula, 
+      but instead do something like tabular
 
-NOTES:
+
+* [ ] if line numbers shift, then ensure that they always report correctly -- so far I don't know if this even needs to change, since it seems to work fine.  Think about it and decide if tests to prevent regressions are useful here.
 
-* possibly don't use fact macro inside of formula, but instead do something like tabular
+* [ ] consider implementing with @marick's metaconstant syntax
+   a. [ ] if we do metaconstant style, implement generator overriding

test/midje/ideas/t_formulas.clj

@@ -36,7 +36,7 @@
   "Uniform distribution from lo (inclusive) to high (exclusive).
    Defaults to range of Java long."
   (^long [] (.nextLong *rnd*))
-  (^long[lo hi] {:pre [(< lo hi)]}
+  (^long [lo hi] {:pre [(< lo hi)]}
     (clojure.core/long (Math/floor (+ lo (* (.nextDouble *rnd*) (- hi lo)))))))
 
 
@@ -57,25 +57,24 @@
 
 ;;;; Formulas
 
+;; the first formula use ever!
+(defn make-string []
+  (rand-nth ["a" "b" "c" "d" "e" "f" "g" "i"]))
+(formula "can now use simple generative-style formulas"
+  [a (make-string) b (make-string)]
+  (str a b) => (has-prefix a))
+
+
 ;; failed formulas report once per formula regardless how many generations were run
 (after-silently
   (formula "some description" [a "y"]
     a => :foo))
 (fact @reported => (one-of (contains {:type :mock-expected-result-failure
                                       :description "some description"})))
 
-(defn make-string []
-  (rand-nth ["a" "b" "c" "d" "e" "f" "g" "i"]))
-
-(formula
-  "can now use simple generative-style formulas"
-  [a (make-string) b (make-string)]
-  (str a b) => (has-prefix a))
-
 
 ;; passing formulas run the generator many times, and evaluate 
 ;; their body many times - number of generations is rebindable
-
 (defn-verifiable y-maker [] "y")
 (defn-verifiable my-str [s] (str s))
 
@@ -85,6 +84,7 @@
 (fact @y-maker-count => 77)
 (fact @my-str-count => 77)
 
+
 ;; runs only as few times as needed to see a failure
 (defn-verifiable z-maker [] "z")
 (defn-verifiable my-identity [x] (identity x))