Jason, thanks for the quick reply. I understand your point, but sometimes
is better to have an in-situ check without needing to return an if-cljs
form, like in:
(defmacro aaa []
(let [xxx (if-cljs yyy zzz)]
... do something complex with xxx))
(defmacro bbb []
(aaa))
So as you said the alternative options are:
A- refactorize aaa to return `(if-cljs ...) forms... always possible?
B- add an explicit env parameter to aaa, but then all your macro libraries
intended to be used from CLJ&CLJS must follow this style.
C- use the old approach, but is incompatible with having 'cljs.analyzer
loaded for other reasons.
I feel there is an unsatisfied need here, but maybe A is sufficient. Also,
I would like to know if there is a rationale behind &env being in a way
(CLJS) in the first macro level and in another way (CLJ) in the next ones,
if is intended or is an implementation limitation.
Saludos,
Nahuel Greco.
On Thu, Jun 19, 2014 at 2:59 AM, Jason Wolfe <[email protected]> wrote:
> The problem with our old solution is that if you called a macro in cljx
> code, and that fn was itself called from Clojure code involved in
> macroexpansion of ClojureScript, it would get the cljs version and barf.
>
> I think our current solution mostly solves this, in that macros that want
> to emit cross-platform code should emit forms like `(if-cljs cljs-form
> clj-form) -- notice the backtick -- and even if they are called from other
> macros, the ultimate expansion takes place in the correct target context (I
> hope).
>
> If a macro wants to directly switch on the env (rather than emitting an
> if-cljs form) the env must be passed in from the parent, but in general I
> think it seems like a better approach. In any case, it seems to have
> solved our immediate issues that some people were having with cljs
> compilation, without breaking any tests.
>
> In your specific example, there's no backtick before "if-cljs", which I
> believe is causing the unexpected behavior (although I think your log fn
> would have to be changed as well to emit a form, as is typical for macro
> helper fns). But I haven't actually tested this, so if this is incorrect
> please let us know.
>
>
> On Wed, Jun 18, 2014 at 10:30 PM, Nahuel Greco <[email protected]> wrote:
>
>> I found a problem in the Chas approach. When you call a macro from a cljs
>> file, and that macro calls another macro with a(:ns &env) test, it sees
>> a Clojure &env, so it thinks is being compiled in a CLJ environment.
>> This doesn't happens with the Prismatic approach (the old one, it was
>> recently changed for the Chas one). See below:
>>
>> ;;;----- p/macros.clj ------
>> (ns p.macros)
>>
>> (defn log [s]
>> (spit "/tmp/log" (str (pr-str s) "\n") :append true))
>>
>> (defmacro if-cljs ;; Chas approach
>> [then else]
>> (log [:if-cljs-env &env])
>> (if (:ns &env) then else))
>>
>>
>> (defn compiling-cljs? ;; Prismatic (old) approach
>> []
>> (boolean
>> (when-let [n (find-ns 'cljs.analyzer)]
>> (when-let [v (ns-resolve n '*cljs-file*)]
>> @v))))
>>
>> (defmacro other-macro
>> []
>> (log [:other-macro-env &env])
>> (log [:if-cljs (if-cljs :cljs :clj)])
>> (log [:if-compiling-cljs (if (compiling-cljs?) :cljs :clj)]))
>>
>> ;;;----- p/core.cljs ------
>> (ns p.macros
>> (:require-macros [p.macros]))
>>
>> (p.macros/other-macro)
>>
>> When you run this, you get the following contents in the /tmp/log file:
>>
>> [:if-cljs-env {&env #<LocalBinding
>> clojure.lang.Compiler$LocalBinding@18b3f8f5>, &form #<LocalBinding
>> clojure.lang.Compiler$LocalBinding@7a69c3fb>}]
>> [:other-macro-env {:column 1, :line 14, :ns {:defs .... }}]
>> [:if-cljs :clj] ;;; WRONG!
>> [:if-compiling-cljs :cljs] ;;; CORRECT
>>
>> Note, [:if-cljs-env ...] is printed before [:if-cljs ...] as the if-cljs
>> macro expands inside the other-macro definition before other-macro is
>> executed from cljs.
>>
>> My question is, why the &env in the if-cljs macro call here is not a
>> ClojureScript one? Bug or feature? :)
>>
>>
>> Saludos,
>> Nahuel Greco.
>>
>>
>> On Tue, Jun 17, 2014 at 11:01 PM, Jason Wolfe <[email protected]> wrote:
>>
>>> Not anymore (thanks Chas):
>>>
>>> https://github.com/Prismatic/schema/pull/113
>>>
>>> On Tuesday, June 3, 2014 7:21:35 AM UTC-7, Nahuel Greco wrote:
>>> > check what Prismatic schema uses:
>>> https://github.com/Prismatic/schema/blob/master/src/clj/schema/macros.clj#L13-20
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > Saludos,
>>> > Nahuel Greco.
>>> >
>>> >
>>> >
>>> > On Tue, Jun 3, 2014 at 9:40 AM, Ambrose Bonnaire-Sergeant <
>>> [email protected]> wrote:
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > I would write two different macros.
>>> >
>>> >
>>> > Thanks,
>>> > Ambrose
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > On Tue, Jun 3, 2014 at 8:07 PM, Максим Карандашов <[email protected]>
>>> wrote:
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > I want to find (in the macro call) what happens: Clojure or
>>> ClojureScript compilation. It needed for using one name of macro inside
>>> Clojure and ClojureScript.
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > Currently I use next approach:
>>> >
>>> >
>>> >
>>> > (if (nil? cljs.env/*compiler*)
>>> >
>>> > (do-something-for-clojure)
>>> >
>>> > (do-something-for-clojurescript))
>>> >
>>> >
>>> >
>>> > And it's working well. But maybe this is not correct? Or is there
>>> another way to do this?
>>> >
>>> >
>>> >
>>> > --
>>> >
>>> > Note that posts from new members are moderated - please be patient
>>> with your first post.
>>> >
>>> > ---
>>> >
>>> > You received this message because you are subscribed to the Google
>>> Groups "ClojureScript" group.
>>> >
>>> > To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>> >
>>> > To post to this group, send email to [email protected].
>>> >
>>> > Visit this group at http://groups.google.com/group/clojurescript.
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > --
>>> >
>>> > Note that posts from new members are moderated - please be patient
>>> with your first post.
>>> >
>>> > ---
>>> >
>>> > You received this message because you are subscribed to the Google
>>> Groups "ClojureScript" group.
>>> >
>>> > To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>> >
>>> > To post to this group, send email to [email protected].
>>> >
>>> > Visit this group at http://groups.google.com/group/clojurescript.
>>>
>>> --
>>> Note that posts from new members are moderated - please be patient with
>>> your first post.
>>> ---
>>> You received this message because you are subscribed to the Google
>>> Groups "ClojureScript" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected].
>>>
>>> To post to this group, send email to [email protected].
>>> Visit this group at http://groups.google.com/group/clojurescript.
>>>
>>
>> --
>> Note that posts from new members are moderated - please be patient with
>> your first post.
>> ---
>> You received this message because you are subscribed to a topic in the
>> Google Groups "ClojureScript" group.
>> To unsubscribe from this topic, visit
>> https://groups.google.com/d/topic/clojurescript/iBY5HaQda4A/unsubscribe.
>> To unsubscribe from this group and all its topics, send an email to
>> [email protected].
>>
>> To post to this group, send email to [email protected].
>> Visit this group at http://groups.google.com/group/clojurescript.
>>
>
> --
> Note that posts from new members are moderated - please be patient with
> your first post.
> ---
> You received this message because you are subscribed to the Google Groups
> "ClojureScript" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/clojurescript.
>
--
Note that posts from new members are moderated - please be patient with your
first post.
---
You received this message because you are subscribed to the Google Groups
"ClojureScript" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/clojurescript.
