You're supposed to handle exceptions yourself in the pod loop and send an exception response

I think it's documented in the README of the pods project. Let me know it isn't. There's also tests for it

I think I got it as advertised, basically copied it over from other pods: https://github.com/TimoKramer/datahike/blob/feat/babashka-pod/src/datahike/pod.clj#L239-267

yeah. is there a question left at this point?

Just thought this is a problem in bb but if you're sure it's not I guess you're right:

#error {
 :cause "clojure.lang.PersistentVector cannot be cast to byte[]"
 :via
 [{:type java.lang.ClassCastException
   :message "clojure.lang.PersistentVector cannot be cast to byte[]"
   :at [babashka.pods.impl$bytes__GT_string invokeStatic "impl.clj" 28]}]
 :trace
 [[babashka.pods.impl$bytes__GT_string invokeStatic "impl.clj" 28]
  [babashka.pods.impl$processor invokeStatic "impl.clj" 190]
  [babashka.pods.sci$load_pod$fn__27430 invoke "sci.clj" 122]
  [sci.impl.vars$binding_conveyor_fn$fn__424 invoke "vars.cljc" 133]
  [clojure.core$binding_conveyor_fn$fn__5823 invoke "core.clj" 2047]
  [clojure.lang.AFn call "AFn.java" 18]
  [java.util.concurrent.FutureTask run "FutureTask.java" 264]
  [java.util.concurrent.ThreadPoolExecutor runWorker "ThreadPoolExecutor.java" 1128]
  [java.util.concurrent.ThreadPoolExecutor$Worker run "ThreadPoolExecutor.java" 628]
  [java.lang.Thread run "Thread.java" 829]
  [com.oracle.svm.core.thread.PlatformThreads threadStartRoutine "PlatformThreads.java" 705]
  [com.oracle.svm.core.posix.thread.PosixPlatformThreads pthreadStartRoutine "PosixPlatformThreads.java" 202]]}

doesn't ring a bell. what I would do is debug with the JVM pods library. perhaps you can do the same

alright, I'll try this

thanks

if you aren't aware, I'd also recommend running your pod in a JVM, so you don't have to graal compile it every time you make a change, this iterates much faster

right, that makes sense

I can take a look later tonight, if you haven't found the issue

hm, I tried a few things, but couldn't find the problem. It think it's the input to the bytes->string -fn... 🤷

would be great if you took a look

have you tried debugging what goes into the bytes->string function?

https://github.com/babashka/pods/blob/8b717eb001811bc5da5d15d1163565de00b4ffa4/src/babashka/pods/impl.clj#L190-L193

apparently there is a "value" in the response, else it couldn't have ended up there

just try printing what goes into that function and report back. check out the pods library using :local/root and make a println change

that's what I am doing 🙂

cool. what does it print?

so the function receives an empty persistenvector presumably from the ex-message part of the processor

why do you think it hits that path? don't the locations you pasted above show that it's this codepath? https://github.com/babashka/pods/blob/8b717eb001811bc5da5d15d1163565de00b4ffa4/src/babashka/pods/impl.clj#L190-L193

that is even before it processes ex-message, etc

that's what I thought but I couldn't find the problem there since it has the some->> macro in there

but of course you're right

BABASHKA_POD=true "clojure" "-M" "-m" "datahike.cli"
Execution error (ClassNotFoundException) at java.net.URLClassLoader/findClass (URLClassLoader.java:445).
datahike.java.IEntity

then I did:

bb jcompile

new error:

$ BABASHKA_POD=true "clojure" "-M" "-m" "datahike.cli"
Execution error (FileNotFoundException) at datahike.pod/eval37638$loading (pod.clj:1).
Could not locate bencode/core__init.class, bencode/core.clj or bencode/core.cljc on classpath.

Full report at:
/var/folders/j9/xmjlcym958b1fr0npsp9msvh0000gn/T/clojure-14533572372815876965.edn

I'm willing to take another look if you can set up stuff in a way that I can debug with the pods library and have clear instructions how to start

One thing I changed in run-bb-pod-tests was:

(pods/load-pod ["clojure" "-M" "-m" "datahike.cli"])

I created an alias so it's

BABASHKA_POD=true "clojure" "-M:pod" "-m" "datahike.cli"

https://github.com/timokramer/datahike/tree/feat/babashka-pod

thanks for trying

$ ./bb/resources/native-image-tests/run-bb-pod-tests.clj

Testing user

Ran 1 tests containing 20 assertions.
0 failures, 0 errors.

so it works when running the pod in a JVM. Could be a reflection issue perhaps? It might be worth fixing these:

$ BABASHKA_POD=true "clojure" "-M:pod" "-m" "datahike.cli"
:yolo
Reflection warning, superv/async.cljc:834:40 - reference to field getTime on java.lang.Object can't be resolved.
Reflection warning, datahike/array.cljc:45:13 - call to static method compare on java.util.Arrays can't be resolved (argument types: unknown, unknown).
Reflection warning, datahike/db/transaction.cljc:231:32 - reference to field e can't be resolved.
Reflection warning, datahike/query_stats.cljc:6:11 - call to java.math.MathContext ctor can't be resolved.

sorry, played around with it too long yesterday and changed the failing test. here it is again: https://github.com/TimoKramer/datahike/blob/feat/babashka-pod/bb/resources/native-image-tests/run-bb-pod-tests.clj#L63 https://app.circleci.com/pipelines/github/TimoKramer/datahike/462/workflows/ff92a007-bd28-46a0-b44b-405325423017/jobs/2282?invite=true#step-102-288

borkdude@m1 /tmp/datahike (feat/babashka-pod?) $ ./bb/resources/native-image-tests/run-bb-pod-tests.clj
----- Error --------------------------------------------------------------------
Type:     java.lang.NullPointerException
Location: /private/tmp/datahike/./bb/resources/native-image-tests/run-bb-pod-tests.clj:7:1

----- Context ------------------------------------------------------------------
 3: (require '[babashka.pods :as pods]
 4:          '[clojure.test :refer [run-tests deftest testing is]])
 5: (import '[java.util Date])
 6:
 7: (pods/load-pod ["clojure" "-M:pod" "-m" "datahike.cli"])
    ^---
 8:
 9: (require '[datahike.pod :as d])
10:
11: (def config {:keep-history? true,
12:              :search-cache-size 10000,

----- Stack trace --------------------------------------------------------------
babashka.pods.impl/bytes->string      - <built-in>
babashka.pods.impl/describe->metadata - <built-in>
babashka.pods.impl/load-pod           - <built-in>
babashka.pods.sci/load-pod/fn--28088  - <built-in>
babashka.pods.sci/load-pod            - <built-in>
clojure.core/apply                    - <built-in>
babashka.impl.pods/load-pod           - <built-in>
user                                  - /private/tmp/datahike/./bb/resources/native-image-tests/run-bb-pod-tests.clj:7:1

borkdude@m1 /tmp/datahike (feat/babashka-pod?) $ Reflection warning, superv/async.cljc:834:40 - reference to field getTime on java.lang.Object can't be resolved.
Reflection warning, datahike/array.cljc:45:13 - call to static method compare on java.util.Arrays can't be resolved (argument types: unknown, unknown).
Reflection warning, datahike/db/transaction.cljc:231:32 - reference to field e can't be resolved.
Reflection warning, datahike/query_stats.cljc:6:11 - call to java.math.MathContext ctor can't be resolved.
:writing {"format" "transit+json", "namespaces" [{"name" "datahike.pod", "vars" [{"name" "delete-database"} {"name" "pull"} {"name" "entity"} {"name" "metrics"} {"name" "datoms"} {"name" "q"} {"name" "schema"} {"name" "pull-many"} {"name" "since"} {"name" "create-database"} {"name" "db"} {"name" "connect"} {"name" "history"} {"name" "database-exists?"} {"name" "transact"} {"name" "as-of"} {"name" "release-db"} {"name" "db-with"} {"name" "with-db", "code" "(defmacro with-db [bindings & body]\n                (cond\n                  (= (count bindings) 0) `(do ~@body)\n                  (symbol? (bindings 0)) `(let ~(subvec bindings 0 2)\n                                            (try\n                                              (with-db ~(subvec bindings 2) ~@body)\n                                              (finally\n                                                (release-db ~(bindings 0)))))\n                  :else (throw (IllegalArgumentException.\n                                 \"with-db only allows Symbols in bindings\"))))"}]}], "id" "42aaf0eb-4ba1-43b4-b459-0929172b81e5", "ops" {"shutdown" {}}}

It seems I'm getting an NPE during the "describe" payload

You don't get this locally?

no, I'm sorry

you're just running the tests?

what jvm version are you using?

ah wait, now it works, I had some reflection warning stuff in place in user.clj

I recommend fixing those as graal doesn't work under reflection warnings in clojure

Yes, we have to do that

This is what I'm seeing right now: https://gist.github.com/borkdude/c44eba75eed1e766810324fb420b3251

yeah sorry, disable the debug in datahike.pod-ns. The NPE is what I expect in the test. What I am getting back with java-17-graalvm is

timo@Zockolette> ./bb/resources/native-image-tests/run-bb-pod-tests.clj                                                                                                                                                                                                                      ~/projects/datahike

Testing user

ERROR in (pod-workflow) (/home/timo/projects/datahike/./bb/resources/native-image-tests/run-bb-pod-tests.clj:22)
transact with bad arg
expected: (= [:tempids :db-before :db-after :tx-meta :tx-data] (keys (d/transact "foo" [{:name "Alice", :age 20} {:name "Bob", :age 30} {:name "Charlie", :age 40} {:age 15}])))
  actual: clojure.lang.ExceptionInfo: Cannot invoke "java.util.concurrent.Future.get()" because "fut" is null
{:argument-type "", :type "class java.lang.NullPointerException"}
 at babashka.pods.impl$processor.invokeStatic (impl.clj:219)
    babashka.pods.sci$load_pod$fn__28022.invoke (sci.clj:122)
    sci.impl.vars$binding_conveyor_fn$fn__440.invoke (vars.cljc:133)
    clojure.core$binding_conveyor_fn$fn__5823.invoke (core.clj:2047)
    clojure.lang.AFn.call (AFn.java:18)
    java.util.concurrent.FutureTask.run (FutureTask.java:317)
    java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1144)
    java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:642)
    java.lang.Thread.run (Thread.java:1589)
    com.oracle.svm.core.thread.PlatformThreads.threadStartRoutine (PlatformThreads.java:775)
    com.oracle.svm.core.posix.thread.PosixPlatformThreads.pthreadStartRoutine (PosixPlatformThreads.java:203)

Ran 1 tests containing 21 assertions.
0 failures, 1 errors.

can you maybe make a separate branch if this is easier? and then I'll check that one out and run

sure

https://github.com/timokramer/datahike/tree/borkdude

debugging is now disabled

ok, this is what I'm seeing:

ERROR in (pod-workflow) (/private/tmp/datahike/./bb/resources/native-image-tests/run-bb-pod-tests.clj:22)
transact with bad arg
expected: (= [:tempids :db-before :db-after :tx-meta :tx-data] (keys (d/transact "foo" [{:name "Alice", :age 20} {:name "Bob", :age 30} {:name "Charlie", :age 40} {:age 15}])))
  actual: clojure.lang.ExceptionInfo: Cannot invoke "java.util.concurrent.Future.get()" because "fut" is null
{:argument-type "", :type "class java.lang.NullPointerException"}
 at babashka.pods.impl$processor.invokeStatic (impl.clj:219)
    babashka.pods.sci$load_pod$fn__28124.invoke (sci.clj:122)
    sci.impl.vars$binding_conveyor_fn$fn__440.invoke (vars.cljc:133)
    clojure.core$binding_conveyor_fn$fn__5823.invoke (core.clj:2047)
    clojure.lang.AFn.call (AFn.java:18)
    java.util.concurrent.FutureTask.run (FutureTask.java:317)
    java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1144)
    java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:642)
    java.lang.Thread.runWith (Thread.java:1636)
    java.lang.Thread.run (Thread.java:1623)
    com.oracle.svm.core.thread.PlatformThreads.threadStartRoutine (PlatformThreads.java:807)
    com.oracle.svm.core.posix.thread.PosixPlatformThreads.pthreadStartRoutine (PosixPlatformThreads.java:210)

Ran 1 tests containing 21 assertions.
0 failures, 1 errors.

I would expect to see the NPE from the debug output from your gist

and in the catch-block from my pod-impl I am not sending a value back as you are not doing either in your samples. so I don't get the exception message since there is a some->> and it should not get into the bytes->string -fn

this is not what I'm seeing here either. could there be laziness at play or so?

I wouldn't expect laziness in transact but I can check with @UB95JRKM3

no need

This doesn't ring a bell, I'd have to dig into your implementation I guess

ok, I guess it's not worth it then.

thanks a lot for taking the time though

not worth what?

you investigating my implementation 🙂

when it's a problem with my impl

well, I don't know off hand, would need more time

which I don't have in the coming days probably :)

alright, I'll be happy if you do

you could try to throw a manual exception and see if that is handled correctly

instead of relying on the datahike stuff

to exclude any external factors

When you feed "foo" into this function:

(defn transact [conn arg-map]
  (let [c (get @conns conn)
        {:keys [db-before db-after tx-meta tx-data tempids]} (d/transact c arg-map)]
    (-> {:tempids tempids}
        (assoc :db-before (select-keys db-before [:max-tx :max-eid]))
        (assoc :db-after (select-keys db-after [:max-tx :max-eid]))
        (assoc :tx-meta tx-meta)
        (assoc :tx-data (map seq tx-data)))))

then c is null. Could it be that the Cannot invoke "java.util.concurrent.Future.get()" because "fut" is null is coming from that?

Yes, that should be the error.

I can try some exceptions manually

so what is then the problem?

if you transact with a bad argument, why don't you expect an exception?

(testing "transact with bad arg"
      (is (= [:tempids :db-before :db-after :tx-meta :tx-data]
             (keys (d/transact
                     "foo"
                     [{:name  "Alice", :age   20}
                      {:name  "Bob", :age   30}
                      {:name  "Charlie", :age   40}
                      {:age 15}])))))

I do expect an exception but I don't understand why I am getting the stacktrace from bb/sci

the test seems to suggest you just get some transaction data back, rather than an exception?

I don't get a stacktrace from bb/sci, I just get a failing test, like expected, I would say

hmm, so this is what you want to see when a test is failing?

at babashka.pods.impl$processor.invokeStatic (impl.clj:219)
    babashka.pods.sci$load_pod$fn__28022.invoke (sci.clj:122)
    sci.impl.vars$binding_conveyor_fn$fn__440.invoke (vars.cljc:133)
    clojure.core$binding_conveyor_fn$fn__5823.invoke (core.clj:2047)
    clojure.lang.AFn.call (AFn.java:18)
    java.util.concurrent.FutureTask.run (FutureTask.java:317)
    java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1144)
    java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:642)
    java.lang.Thread.run (Thread.java:1589)
    com.oracle.svm.core.thread.PlatformThreads.threadStartRoutine (PlatformThreads.java:775)
    com.oracle.svm.core.posix.thread.PosixPlatformThreads.pthreadStartRoutine 

ok, then it's my bad

yes, this is the exception thrown from the pod processor as this is is the exception you wanted to raise by sending exception data

so you write a test with a thrown? or thrown-with-msg?

that's what I would have done yes

I guess it's sorted then :)

oh man, sorry for keeping you busy

the internal pod failure we saw before isn't something I can reproduce here based on the borkdude branch

no worries!

thanks

if you have more questions, feel free to bug me. the better the repro, the faster I can provide you with more info