Fix nested inline fn expansion #34
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
camsaul
commented
Aug 19, 2020
Comment on lines
+1
to
+3
| ((nil . ((indent-tabs-mode . nil) | ||
| (require-final-newline . t))) | ||
| (clojure-mode . ((cljr-favor-prefix-notation . nil)))) |
There was a problem hiding this comment.
This is just a bit of editor config settings for people who use Emacs
camsaul
force-pushed
the
fix-nested-inline-fn-expansion
branch
2 times, most recently
from
005d071
to
93c2875
Compare
camsaul
commented
Aug 19, 2020
project.clj
Outdated
Comment on lines
12
to
16
| :profiles {:provided {:dependencies [[org.clojure/clojure "1.8.0"]]}} | ||
| :profiles {:dev {:dependencies [[org.clojure/clojure "1.10.1"] | ||
| [pjstadig/humane-test-output "0.10.0"]] | ||
| :injections [(require 'pjstadig.humane-test-output) | ||
| (pjstadig.humane-test-output/activate!)]}} |
There was a problem hiding this comment.
Changed the :provided profile to a :dev one and added humane-test-output to make test failures easier to read
camsaul
force-pushed
the
fix-nested-inline-fn-expansion
branch
from
93c2875
to
1b81669
Compare
camsaul
commented
Aug 19, 2020
Comment on lines
+13
to
+16
| :dev {:dependencies [[org.clojure/clojure "1.10.1"] | ||
| [pjstadig/humane-test-output "0.10.0"]] | ||
| :injections [(require 'pjstadig.humane-test-output) | ||
| (pjstadig.humane-test-output/activate!)]}} |
There was a problem hiding this comment.
Added a :dev profile that uses Clojure 1.10.1 by default and adds humane-test-output to make the tests easier to read
camsaul
commented
Aug 19, 2020
Comment on lines
-14
to
+18
| :javac-options ["-target" "1.6" "-source" "1.6"]) | ||
| :javac-options ["-target" "1.7" "-source" "1.7"]) |
There was a problem hiding this comment.
I'm using Java 14 locally and it doesn't support compiling to Java 1.6 bytecode. I think dropping 1.6 support makes sense -- Clojure 1.10+ requires Java 8+ and Clojure 1.5+ requires Java 7 for all features to work out-of-the-box
camsaul
commented
Aug 19, 2020
Comment on lines
-209
to
+316
| 'def def-handler | ||
| 'fn* fn-handler | ||
| 'let* let-handler | ||
| 'loop* let-handler | ||
| 'letfn* letfn-handler | ||
| 'case* case-handler | ||
| 'try try-handler | ||
| 'catch catch-handler | ||
| 'reify* reify-handler | ||
| 'deftype* deftype-handler | ||
| '. dot-handler | ||
| #(doall (map %1 %2))) | ||
| walk-exprs' (special-meta x))) | ||
|
|
||
| (instance? java.util.Map$Entry x) | ||
| (clojure.lang.MapEntry. | ||
| (walk-exprs' (key x)) | ||
| (walk-exprs' (val x))) | ||
|
|
||
| (or | ||
| (set? x) | ||
| (vector? x)) | ||
| (into (empty x) (map walk-exprs' x)) | ||
|
|
||
| (instance? clojure.lang.IRecord x) | ||
| x | ||
|
|
||
| (map? x) | ||
| (into (empty x) (map walk-exprs' x)) | ||
|
|
||
| ;; special case to handle clojure.test | ||
| (and (symbol? x) (-> x meta :test)) | ||
| (vary-meta x update-in [:test] walk-exprs') | ||
|
|
||
| :else | ||
| x)] | ||
| (if (instance? clojure.lang.IObj x') | ||
| (with-meta x' (merge (meta x) (meta x'))) | ||
| x'))))) | ||
| (cmp/with-base-env | ||
| (let [x (try | ||
| (macroexpand x special-form?) | ||
| (catch ClassNotFoundException _ | ||
| x)) | ||
| walk-exprs' (partial walk-exprs predicate handler special-form?) | ||
| x' (cond | ||
|
|
||
| (and (head= x 'var) (predicate x)) | ||
| (handler (eval x)) | ||
|
|
||
| (and (head= x 'quote) (not (predicate x))) | ||
| x | ||
|
|
||
| (predicate x) | ||
| (handler x) | ||
|
|
||
| (seq? x) | ||
| (if (or (and (not try-clause?) | ||
| (#{'catch 'finally} (first x))) | ||
| (not (contains? special-forms (first x)))) | ||
| (doall (map walk-exprs' x)) | ||
| ((condp = (first x) | ||
| 'do do-handler | ||
| 'def def-handler | ||
| 'fn* fn-handler | ||
| 'let* let-handler | ||
| 'loop* let-handler | ||
| 'letfn* letfn-handler | ||
| 'case* case-handler | ||
| 'try try-handler | ||
| 'catch catch-handler | ||
| 'reify* reify-handler | ||
| 'deftype* deftype-handler | ||
| '. dot-handler | ||
| #(doall (map %1 %2))) | ||
| walk-exprs' (special-meta x))) | ||
|
|
||
| (instance? java.util.Map$Entry x) | ||
| (clojure.lang.MapEntry. | ||
| (walk-exprs' (key x)) | ||
| (walk-exprs' (val x))) | ||
|
|
||
| (or | ||
| (set? x) | ||
| (vector? x)) | ||
| (into (empty x) (map walk-exprs' x)) | ||
|
|
||
| (instance? clojure.lang.IRecord x) | ||
| x | ||
|
|
||
| (map? x) | ||
| (into (empty x) (map walk-exprs' x)) | ||
|
|
||
| ;; special case to handle clojure.test | ||
| (and (symbol? x) (-> x meta :test)) | ||
| (vary-meta x update-in [:test] walk-exprs') | ||
|
|
||
| :else | ||
| x)] | ||
| (merge-meta x' (meta x)))))) |
There was a problem hiding this comment.
The only change here was to fix the indentation and use the merge-meta helper function I wrote to simplify things a bit
|
@ztellman bump |
|
Hi, I'm sorry for my lack of response so far, this isn't a repo I've thought about in quite some time. If it were to get moved over to the |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #33
The root cause of the issue was that the last form in expanded inline function calls would be marked with
^::transformedmetadata to prevent infinite expansions in cases where the resultant form looked like it could be an inline function form. e.g.(inc 1)expands to(. clojure.lang.Numbers (inc 1)); the intention was that by marking the(inc 1)as^::transformedwe would know not to try to expand it as an inline function a second time. This however caused problems when using inlineable functions inside other inlineable functions -- see #33.After poking around a bit I determined the
^::transformedmetadata isn't needed because the code that handles java interop forms (dothandler) does not apply the expression-walking function to the method form (e.g.(inc 1)) or the parent.Java interop form. I added tests to verify that this indeed works as expected even without marking things as^::transformed.As part of figuring out how to fix this issue I did a bit of refactoring and broke
macroexpandout into a series of smaller functions to make it easier to follow. I added a bunch of extra tests around this refactored code. Hope that's ok!