zprint-filter-1.2.3

[kkinnear]

A fair number of fixes and changes in this release. See the CHANGELOG.md for details, but notable new things are:

• New :range configuration options: :use-previous-!zprint? and :continue-after-!zprint-error?. These allow someone doing :range formatting on a file (string) which has ;!zprint formatting directives to have the lines in the range formatted the same when doing just the range or the whole file. :use-previous-!zprint? if true will scan all previous ;!zprint directives prior to formatting the range. :continue-after-!zprint-error? if true will ignore errors in any ;!zprint directives, and output any errors in the map returned by :output {:range? true}.

• An alternative way to format metadata was created. It is :style :meta-alt. It is presently experiemental. Issue Metadata on deftest (and def) #224.

• letfn didn't indent its fns like fns. Now it does. And if they are multi-arity, the various arities don't hang, but rather flow. Issue letfn doesn't indent its fns like fns. #221.

• The styles :community and :rod didn't compose, because :guided was not considered a "body" function type. Now they do. Issue :community and :rod don't compose well #223.

• zprint-file-str was not thread-safe, for a variety of reasons. Now it is, and configuration is thread-safe as well. Issue Thread safety issues #226.

• Figured out some Figwheel issues. If you use zprint with Figwheel, check out the CHANGELOG.md, or see issue Clojurescript warnings with zprint 1.2.0, 1.2.2 #231.

---

This discussion was created from the release zprint-filter-1.2.3.

[arichiardi]

Hey @kkinnear I added :meta-alt

:style [:community :meta-alt :how-to-ns :binding-nl]

and it definitely addresses #224 - I think the only issue I see is this change:

-(deftest valid-ref?
-  (is (false? (sut/valid-ref? nil)))
+(deftest valid-ref? (is (false? (sut/valid-ref? nil)))
   (is (false? (sut/valid-ref? "")))
   (is (false? (sut/valid-ref? "i/do/not/start/with/a/slash")))
   (is (true? (sut/valid-ref? "/this/is/a/valid/ref")))

The first(is (false? ... looks weird there

I then added this instead only for deftest

{:style [:community :how-to-ns :binding-nl]
 :fn-map
 {"deftest" [:arg1-body {:style :meta-alt}]
 ...
}

That looked better, with only one little change that seems to happen all over the code base, a couple of examples:

 (t/deftest broadcast-to!
   (t/testing "cannot be performed in a transaction"
     (t/is (thrown? IllegalStateException
-                   (dosync (sut/broadcast-to! (mock-socket) (mock-topic) (mock-event)))))))
+                   (dosync (sut/broadcast-to! (mock-socket)
+                                              (mock-topic)
+                                              (mock-event)))))))

-          topic-a (sut/with-event-handlers (sut/->TopicController "topic/a" ::any [] [])
-                                           (sut/->EventHandler "topic.a/x" ::any))
+          topic-a (sut/with-event-handlers
+                   (sut/->TopicController "topic/a" ::any [] [])
+                   (sut/->EventHandler "topic.a/x" ::any))

 (deftest add-template-text-to-note
   (testing "note without value"
     (is (= {::model.note/id note-id ::model.note/value "foo"}
-           (sut/add-note-text-if-not-included {::model.note/id note-id} "foo"))
+           (sut/add-note-text-if-not-included {::model.note/id note-id}
+                                              "foo"))
         "it should add the new text"))

I found this worth mentioning cause there might be a bug lurking, let me know if you want me to try any alpha/beta version or have more questions!

[kkinnear]

Thanks for bringing this up. I totally agree that deftest should be :arg1-body in :style :meta-alt, I'll make that change. I think that the best intermediate approach would be: {:fn-map {"deftest" [:arg1-body {:style :meta-base}]}}. Re-invoking :meta-alt (as you did, above) is probably not the best approach.

That said, I am unable to reproduce the other issue you mention. Well, actually, it reproduces easily for me, since my default width is 80:

zprint.core=> (czprint i232c {:parse-string? true :style [:community :meta-alt :how-to-ns :binding-nl] :fn-map {"deftest" [:arg1-body {:style :meta-base}]}})
(deftest add-template-text-to-note
  (testing "note without value"
           (is (= {::model.note/id note-id, ::model.note/value "foo"}
                  (sut/add-note-text-if-not-included {::model.note/id note-id}
                                                     "foo"))
               "it should add the new text")))
nil

; If I make the width 90, then you get what you want:

zprint.core=> (czprint i232c {:parse-string? true :style [:community :meta-alt :how-to-ns :binding-nl] :fn-map {"deftest" [:arg1-body {:style :meta-base}]} :width 90})
(deftest add-template-text-to-note
  (testing "note without value"
           (is (= {::model.note/id note-id, ::model.note/value "foo"}
                  (sut/add-note-text-if-not-included {::model.note/id note-id} "foo"))
               "it should add the new text")))

So, I'm wondering if the changes in your formatting is somehow related to whatever your current width might be? I've looked, hard, and there isn't anything in the :styles that affects the width.

If you give me your complete configuration (including whatever your current width is) I'll try to reproduce this as well. Thanks!

[arichiardi]

Thank @kkinnear for your answer. I've moved the config to

"deftest" [:arg1-body {:style :meta-base}]

and that did the trick.

The weird thing is that I cannot reproduce the width problem anymore and I am baffled as to why that happened, see #224 on actual test cases