feat: refactor entirely json processing in favor or rawjson to handle larger payloads (#2446)

* raw json

* raw json

* raw json

* raw json

* the big rawjson rewrite

* resolve merge conflicts

* all

* progress

* progress

* all compile

* remove mut

* remove mut

* remove mut

* arc queuedjob

* finish

* finish

* finish

* fixes

* progress

* fix result

* small fixes

* small fixes

* fix tests

* fix tests

* fix tests

* optimize eval timeout

* optimize eval timeout

* optimize eval timeout

* handle big args and result in get jobs

* improve args

* fix

* fix

* fix tests

* fix tests

* done

* add bigquery to native picker

* add bigquery

---------

Co-authored-by: gbouv <guillaume@windmill.dev>

backend/windmill-api/src/scripts.rs

@@ -46,7 +46,9 @@ use windmill_common::{
         not_found_if_none, paginate, query_elems_from_hub, require_admin, Pagination, StripPath,
     },
 };
-use windmill_queue::{self, schedule::push_scheduled_job, PushIsolationLevel, QueueTransaction};
+use windmill_queue::{
+    self, schedule::push_scheduled_job, PushArgs, PushIsolationLevel, QueueTransaction,
+};
 
 const MAX_HASH_HISTORY_LENGTH_STORED: usize = 20;
 
@@ -592,7 +594,7 @@ async fn create_script(
             tx,
             &w_id,
             JobPayload::Dependencies { hash, dependencies, language: ns.language, path: ns.path },
-            serde_json::Map::new(),
+            PushArgs::empty(),
             &authed.username,
             &authed.email,
             username_to_permissioned_as(&authed.username),