Variable naming to indicate pointer/value

Author: lukasbindreiter

apis/workflows/v1/core.proto

@@ -206,8 +206,9 @@ message TaskSubmission {
 // To reduce the serialization size, we keep a separate list/lookup table of unique task properties, that can then be
 // referenced by their index.
 message TaskSubmissions {
-  // Concrete instantiations of tasks, grouped by their dependencies and dependants, and referencing entries in the
-  // lookup tables also contained in this message by index.
+  // Concrete instantiations of tasks, grouped by their dependencies and dependants. Each group is uniquely defined by
+  // the set of groups that it depends on (dependencies_on_other_groups) and the set of groups that depend on it,
+  // (which is implicitly given by the inverse of the dependencies on other groups).
   repeated TaskSubmissionGroup task_groups = 1;
   // Unique values of cluster slugs, referenced by index in the task instantiations.
   repeated string cluster_slug_lookup = 2;
@@ -225,9 +226,24 @@ message TaskSubmissions {
 // across groups, since there may be two groups sharing the same dependencies but having different dependants.
 message TaskSubmissionGroup {
   option (buf.validate.message).cel = {
-    id: "task_submission_group.task_fields_size_match"
-    message: "The length of the task related fields must match."
-    expression: "this.inputs.size() == this.max_retries.size() && this.inputs.size() == this.identifier.size() && this.inputs.size() == this.cluster_slug.size() && this.inputs.size() == this.display.size() && this.inputs.size() == this.max_retries.size()"
+    id: "task_submission_group.identifiers_size_match"
+    message: "The number of inputs must match the number of task identifiers."
+    expression: "this.inputs.size() == this.identifier_pointers.size()"
+  };
+  option (buf.validate.message).cel = {
+    id: "task_submission_group.cluster_slugs_size_match"
+    message: "The number of cluster slugs must match the number of inputs."
+    expression: "this.inputs.size() == this.cluster_slug_pointers.size()"
+  };
+  option (buf.validate.message).cel = {
+    id: "task_submission_group.displays_size_match"
+    message: "The number of display pointers must match the number of inputs."
+    expression: "this.inputs.size() == this.display_pointers.size()"
+  };
+  option (buf.validate.message).cel = {
+    id: "task_submission_group.max_retries_size_match"
+    message: "The number of max_retries_values must match the number of inputs."
+    expression: "this.inputs.size() == this.max_retries_values.size()"
   };
 
   // The indices of the groups that this submission group depends on. Indices refer to the groups field of the
@@ -239,15 +255,15 @@ message TaskSubmissionGroup {
   repeated bytes inputs = 2 [(buf.validate.field).repeated.items.bytes.max_len = 2048];
   // Index of the task identifier in the identifier_lookup field of the containing TaskSubmissions message
   // for each task.
-  repeated uint64 identifier = 3;
+  repeated uint64 identifier_pointers = 3;
   // Index of the cluster slug in the cluster_slug_lookup field of the containing TaskSubmissions message for each task,
   // indicating the cluster that the task should be run on.
-  repeated uint64 cluster_slug = 4;
+  repeated uint64 cluster_slug_pointers = 4;
   // Index of the display name in the display_lookup field of the containing TaskSubmissions message for each task,
   // specifying a human-readable description of the task.
-  repeated uint64 display = 5;
+  repeated uint64 display_pointers = 5;
   // The maximum number of retries for each task. Not a pointer to a lookup table, since we just inline the values.
-  repeated int64 max_retries = 6;
+  repeated int64 max_retries_values = 6;
 }
 
 // A lease for a task.