Reimplement attributes view for component debug screen #3475

britmyerss · 2024-03-26T23:32:10Z

This PR brings back the Debug view in the Attribute Panel, with some cleanups along the way:

Created debug modules for the various Node types on the graph: Socket, Attribute Prototype, Attribute Value, Component
Moved logic out of SDF and into the DAL
Moved the debug components into a Debug folder in app/web
Fixed up the interfaces a bit there for this data
Created a AttributeValuePath struct to consolidate path creation for attribute values for props and sockets (including the key/index for child attribute values of maps/arrays)
Includes input/output socket and attribute value debug data!
One realllllllllly basic integration test, but this should be expanded

britmyerss · 2024-03-27T00:23:30Z

bors try

si-bors-ng · 2024-03-27T00:35:35Z

try

Build failed:

buildkite/si-merge-queue

britmyerss · 2024-03-27T01:29:45Z

bors try

britmyerss · 2024-03-27T01:30:32Z

bors cancel

si-bors-ng · 2024-03-27T01:35:28Z

try

Build failed:

buildkite/si-merge-queue

britmyerss · 2024-03-27T05:30:00Z

bors try

si-bors-ng · 2024-03-27T05:39:32Z

try

Build succeeded:

buildkite/si-merge-queue

nickgerace

Alright @britmyerss, you wanted nits... you got nits!

nickgerace · 2024-03-27T18:20:31Z

app/web/src/components/AttributeDebugView.vue

@@ -1,7 +1,7 @@
 <template>
  <div class="overflow-hidden my-xs p-xs border-opacity-10 border-l-2">
    <dl class="flex flex-col gap-xs">
-      <DebugViewItem title="Attribute Value Id" :data="data.valueId" />
+      <DebugViewItem title="Attribute Value Id" :data="data.attributeValueId" />


I like this change

nickgerace · 2024-03-27T18:21:59Z

app/web/src/components/AttributeDebugView.vue

Question about the name of this file. Is it an AttributeValue view, an AttributePrototype view, both or something else?

We're overloading the usage of the word "Attribute" across the system to mean subtly different things and should tighten it up. However, the whole world is unlikely in scope for this PR.

yeah I asked myself the same thing here. I decided to keep it as is both because the whole world is not in scope for this, and because this actually is a debug view of the attributes and we're displaying the AttributePrototype info from the perspective of the AttributeValues - but your questions below, I separated this on the backend because the next step is to bring back input/output sockets, which will also display the same AttributePrototype debug data for them AND I want to surface this on the viz diagram eventually.

Okay that sounds good to me.

nickgerace · 2024-03-27T18:22:09Z

lib/dal/src/attribute/prototype.rs

@@ -33,6 +33,7 @@ use crate::{
 };

 pub mod argument;
+pub mod debug;


I like the module!

nickgerace · 2024-03-27T18:23:48Z

lib/dal/src/attribute/prototype/debug.rs

+use std::collections::HashMap;
+
+use serde::{Deserialize, Serialize};
+
+use thiserror::Error;
+
+use crate::attribute::prototype::{
+    argument::{
+        static_value::StaticArgumentValue,
+        value_source::{ValueSource, ValueSourceError},
+        AttributePrototypeArgument, AttributePrototypeArgumentError,
+    },
+    AttributePrototypeError,
+};
+use crate::attribute::value::{AttributeValueError, ValueIsFor};
+
+use crate::prop::{PropError, PropPath};
+
+use crate::func::execution::{FuncExecution, FuncExecutionError};
+use crate::socket::input::InputSocketError;
+use crate::socket::output::OutputSocketError;
+use crate::workspace_snapshot::node_weight::NodeWeightError;
+use crate::workspace_snapshot::WorkspaceSnapshotError;
+use crate::{
+    AttributePrototype, AttributePrototypeId, AttributeValue, AttributeValueId, Component,
+    ComponentError, DalContext, Func, FuncError, FuncId, InputSocket, Prop,
+};
+use serde_json::Value;
+use telemetry::prelude::*;


We should clean up these imports. Let's put all use <non-crate> statements at the top, then a blank line, then all use crate statements.

Some folks split use std statements in their files too, but we usually group them together (unclear if by habit or intention, but that's a topic for another day).

We don't have a linter that will sort those for us?

IDEs typically do. There are a few features needed to do this well via automation and at least one of them is not stable yet: rust-lang/rustfmt#4991.

will do! yeah my IDE does some of this formatting and merging for me but I wasn't sure what 'more right' vs 'less right' looked like :)

nickgerace · 2024-03-27T18:24:01Z

lib/dal/src/attribute/prototype/debug.rs

+
+#[derive(Clone, Debug, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase")]
+pub struct AttributePrototypeDebugView {


I think this answers my other question.

lib/dal/src/attribute/prototype/debug.rs

nickgerace · 2024-03-27T18:37:47Z

lib/dal/src/attribute/prototype/debug.rs

+                        other_source => {
+                            let mut values = vec![];
+
+                            for av_id in other_source
+                                .attribute_values_for_component_id(
+                                    ctx,
+                                    expected_source_component_id,
+                                )
+                                .await?
+                            {
+                                let attribute_value = AttributeValue::get_by_id(ctx, av_id).await?;


A smell for me when match arms get too big is when you have a scope that contains non-trivial iterator or recursive logic in them. I'd look into breaking this up.

Example:

match foo { Some(v) => println!("{v}"), None => { for i in big_ass_vec { // a bunch of code }

lib/dal/src/attribute/prototype/debug.rs

nickgerace · 2024-03-27T18:39:27Z

lib/dal/src/attribute/value.rs

+    #[instrument(level = "info", skip_all)]
+    /// Get the morale equivalent of the prop path for a given attribute value id.
+    /// This includes the key/index in the path, unlike the path for a Prop which doesn't
+    /// include the key/index
+    pub async fn get_path_for_id(


For style, I'd put the comment above the macro.

nickgerace · 2024-03-27T18:41:25Z

lib/dal/src/attribute/value.rs

+            let prop_name = match AttributeValue::prop_id_for_id(ctx, attribute_value_id).await {
+                Ok(prop) => Prop::get_by_id(ctx, prop).await?.name,
+                Err(_) => {
+                    // This means it's an input or output socket, which has no path and should just be empty string
+                    continue;
+                }
+            };


Is the only error possible here the one that you listed? If not, then we should handle the specific error or refactor the prop_id_for_id function to return a different type to help you determine that (e.g. Result<Option<T>, E>> instead of Result<T, E>).

nickgerace · 2024-03-27T18:42:32Z

lib/dal/src/attribute/value.rs

+                        let order_of_thing_node = ctx
+                            .workspace_snapshot()?
+                            .incoming_sources_for_edge_weight_kind(
+                                attribute_value_id,
+                                EdgeWeightKindDiscriminants::Ordinal,
+                            )
+                            .await?
+                            .pop()
+                            .ok_or(AttributeValueError::NoOrderingNodeForAttributeValue(
+                                attribute_value_id,
+                            ))?;


This should be its own function for AttributeValue. We do not want to expose the snapshot.

nickgerace · 2024-03-27T18:43:08Z

lib/dal/src/attribute/value.rs

+                        let order_node_weight = ctx
+                            .workspace_snapshot()?
+                            .get_node_weight(order_of_thing_node)
+                            .await?
+                            .get_ordering_node_weight()?;
+
+                        let order = order_node_weight.order();


Same here as above.... Maybe on WorkspaceSnapshot itself though since "ordering nodes" do not belong to SI-specific entities.

nickgerace · 2024-03-27T18:43:36Z

lib/dal/src/attribute/value.rs

+                        let name = format!("{0}[{1}]", prop_name, &index.to_string());
+                        parts.push_front(name.to_owned());


What is being assembled here? It looks like the path for display purposes in the frontend.

That is correct. This function is supposed to assemble a path to the attribute value, for display in the debugging panel. So if its an array item, it would be something like /root/domain/things/thing[0] with 0 being the index of the item, thing being the prop name for entry and things the prop name for the array.

For a map its the same thing but with key instead of index

Gotcha. Is there a reason why we wouldn't use PropPath instead of strings? Is it because it doesn't work with array items?

Exactly. Prop paths are different from the attribute value paths this creates since we show keys and indices

I see!! String manipulation makes me nervous... I wonder if a theoretical AttributeValuePath entity would be useful.

Yeah I debated making an AttributeValuePath entity, because I was honestly surprised we didn't have this yet. But it felt strange to formalize it for just this debug view and I figured if we end up needing it later we can always formalize it then and replace this implementation with that one.

I won't be a blocker here, but I would personally formalize it because there may be a day 2 bug or feature extension here that relies on using/changing/extending the string manipulation.

It's a nit though.

nickgerace · 2024-03-27T18:43:48Z

lib/dal/src/attribute/value.rs

+                            let edge_weight = ctx
+                                .workspace_snapshot()?
+                                .edges_directed_for_edge_weight_kind(
+                                    attribute_value_id,
+                                    Direction::Incoming,
+                                    EdgeWeightKindDiscriminants::Contain,
+                                )
+                                .await?
+                                .pop()
+                                .ok_or(AttributeValueError::ParentAttributeValueMissing(
+                                    attribute_value_id,
+                                ))?;


Same comment as above with not exposing the graph

lib/dal/src/attribute/value.rs

nickgerace · 2024-03-27T18:45:25Z

lib/dal/src/attribute/value/debug.rs

+use std::collections::HashMap;
+
+use serde::{Deserialize, Serialize};
+
+use thiserror::Error;
+
+use crate::attribute::prototype::debug::{
+    AttributePrototypeDebugView, AttributePrototypeDebugViewError,
+};
+use crate::attribute::prototype::{
+    argument::{value_source::ValueSourceError, AttributePrototypeArgumentError},
+    AttributePrototypeError,
+};
+use crate::attribute::value::AttributeValueError;
+use crate::func::execution::FuncExecutionError;
+
+use crate::prop::PropError;
+
+use crate::socket::input::InputSocketError;
+use crate::socket::output::OutputSocketError;
+
+use crate::workspace_snapshot::node_weight::NodeWeightError;
+use crate::workspace_snapshot::WorkspaceSnapshotError;
+
+use crate::ComponentError;
+use telemetry::prelude::*;
+
+use crate::{func::execution::FuncExecution, AttributeValue, AttributeValueId, DalContext, Prop};
+use crate::{AttributePrototypeId, FuncError, FuncId, PropKind};
+
+use super::ValueIsFor;


Same comment about imports as before

nickgerace · 2024-03-27T18:45:37Z

lib/dal/src/attribute/value/debug.rs

+    #[error(transparent)]
+    AttributePrototypeArgumentError(#[from] AttributePrototypeArgumentError),
+    #[error(transparent)]
+    AttributePrototypeDebugViewError(#[from] AttributePrototypeDebugViewError),
+    #[error(transparent)]
+    AttributePrototypeError(#[from] AttributePrototypeError),
+    #[error(transparent)]
+    AttributeValue(#[from] AttributeValueError),
+    #[error(transparent)]
+    ComponentError(#[from] ComponentError),
+    #[error(transparent)]
+    Func(#[from] FuncError),
+    #[error(transparent)]
+    FuncExecution(#[from] FuncExecutionError),
+    #[error(transparent)]
+    InputSocketError(#[from] InputSocketError),
+    #[error(transparent)]
+    NodeWeightError(#[from] NodeWeightError),
+    #[error(transparent)]
+    OutputSocketError(#[from] OutputSocketError),
+    #[error(transparent)]
+    PropError(#[from] PropError),
+    #[error(transparent)]
+    ValueSourceError(#[from] ValueSourceError),
+    #[error(transparent)]
+    WorkspaceSnapshotError(#[from] WorkspaceSnapshotError),


Same comment about transparent as before

nickgerace · 2024-03-27T18:46:10Z

lib/dal/src/attribute/value/debug.rs

+        let attribute_value = AttributeValue::get_by_id(ctx, attribute_value_id).await?;
+        let prototype_id = AttributeValue::prototype_id(ctx, attribute_value_id).await?;
+        let value_is_for = AttributeValue::is_for(ctx, attribute_value_id).await?;
+
+        let prop_id = AttributeValue::prop_id_for_id(ctx, attribute_value_id).await?;
+
+        let prop = Prop::get_by_id(ctx, prop_id).await?;
+        let path = match AttributeValue::get_path_for_id(ctx, attribute_value_id).await? {
+            Some(path) => path,
+            None => String::new(),
+        };
+        let prop_opt: Option<Prop> = Some(prop);
+        let attribute_prototype_debug_view =
+            AttributePrototypeDebugView::new_from_attribute_value(ctx, attribute_value_id).await?;
+        let materialized_view = attribute_value.materialized_view(ctx).await?;
+        let prop_kind = prop_opt.clone().map(|prop| prop.kind);
+        let value = match attribute_value.unprocessed_value(ctx).await? {
+            Some(value) => Some(value),
+            None => attribute_value.value(ctx).await?,
+        };
+        let view = AttributeDebugView {
+            path,
+            parent_id,
+            attribute_value_id,
+            func_id: attribute_prototype_debug_view.func_id,
+            key,
+            func_execution: attribute_prototype_debug_view.func_execution,
+            prop: prop_opt,
+            prototype_id: Some(prototype_id),
+            value_is_for,
+            func_name: attribute_prototype_debug_view.func_name,
+            func_args: attribute_prototype_debug_view.func_args,
+            arg_sources: attribute_prototype_debug_view.arg_sources,
+            value,
+            prop_kind,
+            materialized_view,
+        };
+        Ok(view)
+    }


Now I'm a bit confused. What's the difference between this and the other view?

nickgerace · 2024-03-27T18:46:27Z

lib/dal/src/component/debug.rs

+    pub attribute_tree: HashMap<AttributeValueId, Vec<AttributeValueId>>,
+    // pub input_sockets: Vec<AttributeDebugView>,
+    // pub output_sockets: Vec<AttributeDebugView>,


We should delete the commented out stuff unless accompanied by a NOTE, TODO or FIXME.

nickgerace · 2024-03-27T18:46:52Z

lib/dal/src/component/debug.rs

+#[remain::sorted]
+#[derive(Error, Debug)]
+pub enum ComponentDebugViewError {
+    #[error(transparent)]


Same comment about transparent

nickgerace · 2024-03-27T18:47:40Z

lib/dal/src/component/debug.rs

+    }
+    #[instrument(level = "info", skip_all)]
+    pub async fn get_attribute_value_tree_for_component(


I'd leave a space here.

lib/dal/src/component/debug.rs

nickgerace · 2024-03-27T18:48:16Z

lib/dal/src/component/debug.rs

+impl ComponentDebugView {
+    #[instrument(level = "info", skip_all)]
+    pub async fn new(ctx: &DalContext, component: &Component) -> ComponentDebugViewResult<Self> {
+        let schema_variant = Component::schema_variant(component, ctx).await?;


nit: You don't need the whole schema variant. We may be able to just fetch the id from the node.

nickgerace · 2024-03-27T18:48:42Z

lib/dal/src/component/debug.rs

+                let maybe_ordering =
+                    AttributeValue::get_child_av_ids_for_ordered_parent(ctx, attribute_value_id)
+                        .await
+                        .ok();


Are we eating a potential error here?

nickgerace · 2024-03-27T18:49:37Z

lib/dal/src/component/debug.rs

+        let mut child_values = HashMap::new();
+        // Get the root attribute value and load it into the work queue.
+        let root_attribute_value_id = Component::root_attribute_value_id(ctx, component_id).await?;
+        let workspace_snapshot = ctx.workspace_snapshot()?;


I'm a bit conflicted. I see other AttributeValue methods called in here, so that's a sign to create more methods to sandbox the snapshot. However, it makes more sense in this function than before because this is more of a raw usage scenario. I'd still investigate breaking them up.

nickgerace · 2024-03-27T18:50:42Z

lib/dal/tests/integration_test/component/debug.rs

+use pretty_assertions_sorted::assert_eq;
+
+#[test]
+async fn get_debug_view(ctx: &mut DalContext) {


Let's leave some spaces and comments. This is difficult to read.

Usually, the convention we have to group things by commit when in doubt. Additionally, grouping by "preparing the environment" vs. performing mutations and assertions can be helpful.

nickgerace · 2024-03-27T18:51:42Z

lib/dal/tests/integration_test/component/debug.rs

Yay! Testing!

nickgerace · 2024-03-27T18:52:14Z

lib/sdf-server/src/server/service/component/debug.rs

+use std::time::Instant;
+
 use axum::extract::Query;
 use axum::Json;

 use serde::{Deserialize, Serialize};

 use super::ComponentResult;
 use crate::server::extract::{AccessBuilder, HandlerContext};
-use dal::{Component, ComponentId, SchemaVariantId, Visibility};
+use dal::attribute::value::debug::AttributeDebugView;
+use dal::component::debug::ComponentDebugView;
+use dal::{AttributeValueId, Component, ComponentId, SchemaVariantId, Visibility};
+use telemetry::prelude::*;


Same comment as before

nickgerace · 2024-03-27T18:52:44Z

lib/sdf-server/src/server/service/component/debug.rs

+    //grab input socket info next
+    //grab output socket info next
+
+    dbg!(transform_start.elapsed());


Let's delete these or put this behind an info or debug statement with context for what it's benchmarking.

nickgerace · 2024-03-27T18:53:53Z

lib/sdf-server/src/server/service/component/debug.rs

+    let debug_view = ComponentDebugView::new(&ctx, &component).await?;

-    let attributes = vec![];
+    let mut attributes = vec![];
    let input_sockets = vec![];
    let output_sockets = vec![];

-    // let transform_start = Instant::now();
-    //
-    // for attribute_debug in debug_view.attributes {
-    //     attributes.push(FrontendAttributeDebugView {
-    //         name: attribute_debug
-    //             .prop
-    //             .as_ref()
-    //             .map(|p| p.name())
-    //             .unwrap_or("")
-    //             .into(),
-    //         path: attribute_debug.path.to_owned(),
-    //         debug_data: get_attribute_metadata(&ctx, attribute_debug).await?,
-    //     });
-    // }
-    //
-    // for attribute_debug in debug_view.input_sockets {
-    //     input_sockets.push(FrontendAttributeDebugView {
-    //         name: attribute_debug.path.to_owned(),
-    //         path: "Input Socket".into(),
-    //         debug_data: get_attribute_metadata(&ctx, attribute_debug).await?,
-    //     });
-    // }
-    //
-    // for attribute_debug in debug_view.output_sockets {
-    //     output_sockets.push(FrontendAttributeDebugView {
-    //         name: attribute_debug.path.to_owned(),
-    //         path: "Output Socket".into(),
-    //         debug_data: get_attribute_metadata(&ctx, attribute_debug).await?,
-    //     });
-    // }
-    //
-    // dbg!(transform_start.elapsed());
+    let transform_start = Instant::now();
+    let mut cache: Vec<AttributeValueId> = vec![];
+    //grab attribute value debug views
+    for (av, children) in debug_view.attribute_tree {
+        if !cache.contains(&av) {
+            let avd = AttributeDebugView::new(&ctx, av, None, None).await?;
+            if let Some(name) = avd.prop.as_ref().map(|p| p.name.to_owned()) {
+                attributes.push(FrontendAttributeDebugView {
+                    name,
+                    path: avd.path.to_owned(),
+                    debug_data: avd,
+                });
+                cache.push(av);
+            }
+        }
+
+        for child_av in children {
+            if !cache.contains(&child_av) {
+                let child_avd = AttributeDebugView::new(&ctx, child_av, None, Some(av)).await?;
+                if let Some(name) = child_avd.prop.as_ref().map(|p| p.name.to_owned()) {
+                    attributes.push(FrontendAttributeDebugView {
+                        name,
+                        path: child_avd.path.to_owned(),
+                        debug_data: child_avd,
+                    });
+                    cache.push(child_av);
+                }
+            }
+        }
+    }
+    attributes.sort_by_key(|view| view.path.to_lowercase());


My instinct is to put this behind a dal struct and then test that in integration tests. We want sdf routes thin in lieu of not having API-driven testing infrastructure. Up to you though if you think this is covered in tests.

britmyerss · 2024-04-02T15:00:12Z

bors try

si-bors-ng · 2024-04-02T15:00:15Z

try

Merge conflict.

nickgerace · 2024-04-02T15:03:34Z

lib/si-pkg/src/spec/attribute_value.rs

These changes are excellent! Love the use of the constant for the separator.

nickgerace · 2024-04-02T15:04:00Z

lib/dal/tests/integration_test/component/debug.rs

+    let attribute_path = AttributeValue::get_path_for_id(ctx, rigid_designator_value_id)
+        .await
+        .expect("can't get the path");
+    assert_eq!(
+        attribute_path,
+        Some(rigid_prop_path.with_replaced_sep("/").to_string())
+    );


nickgerace · 2024-04-02T15:04:16Z

Really awesome work @britmyerss 🥳

britmyerss · 2024-04-02T15:19:42Z

bors try

si-bors-ng · 2024-04-02T15:36:01Z

try

Build succeeded:

buildkite/si-merge-queue

britmyerss · 2024-04-02T15:44:37Z

bors merge

si-bors-ng · 2024-04-02T15:58:17Z

Build succeeded:

buildkite/si-merge-queue

britmyerss requested a review from vbustamante March 26, 2024 23:32

github-actions bot added A-sdf Area: Primary backend API service [Rust] A-dal A-web labels Mar 26, 2024

britmyerss force-pushed the feat/component-debug branch from caff78d to a4989ff Compare March 27, 2024 00:02

si-bors-ng bot added a commit that referenced this pull request Mar 27, 2024

Try #3475:

3ce80a8

britmyerss force-pushed the feat/component-debug branch from a4989ff to 8271b39 Compare March 27, 2024 01:27

si-bors-ng bot added a commit that referenced this pull request Mar 27, 2024

Try #3475:

a973c3b

britmyerss force-pushed the feat/component-debug branch from 8271b39 to 0e135b5 Compare March 27, 2024 01:31

si-bors-ng bot added a commit that referenced this pull request Mar 27, 2024

Try #3475:

3ce281b

nickgerace requested changes Mar 27, 2024

View reviewed changes

nickgerace reviewed Mar 27, 2024

View reviewed changes

nickgerace mentioned this pull request Mar 27, 2024

feat(dal,sdf,app): Reimplement variant_defs endpoints #3474

Merged

britmyerss force-pushed the feat/component-debug branch from 0e135b5 to 0cb9e87 Compare April 2, 2024 02:56

github-actions bot added the A-si-pkg label Apr 2, 2024

britmyerss requested a review from nickgerace April 2, 2024 03:10

britmyerss force-pushed the feat/component-debug branch from 0cb9e87 to ec9fe2c Compare April 2, 2024 03:34

nickgerace approved these changes Apr 2, 2024

View reviewed changes

Reimplement attributes view for component debug screen

475c755

britmyerss force-pushed the feat/component-debug branch from ec9fe2c to 475c755 Compare April 2, 2024 15:12

si-bors-ng bot added a commit that referenced this pull request Apr 2, 2024

Try #3475:

0a62c0e

si-bors-ng bot merged commit ee467b6 into main Apr 2, 2024
9 checks passed

si-bors-ng bot deleted the feat/component-debug branch April 2, 2024 15:58

		let name = format!("{0}[{1}]", prop_name, &index.to_string());
		parts.push_front(name.to_owned());

Reimplement attributes view for component debug screen #3475

Reimplement attributes view for component debug screen #3475

Conversation

britmyerss commented Mar 26, 2024 • edited Loading

britmyerss commented Mar 27, 2024

si-bors-ng bot commented Mar 27, 2024

try

britmyerss commented Mar 27, 2024

britmyerss commented Mar 27, 2024

si-bors-ng bot commented Mar 27, 2024

try

britmyerss commented Mar 27, 2024

si-bors-ng bot commented Mar 27, 2024

try

nickgerace left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nickgerace Mar 27, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nickgerace Mar 27, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

britmyerss commented Apr 2, 2024

si-bors-ng bot commented Apr 2, 2024

try

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nickgerace commented Apr 2, 2024

britmyerss commented Apr 2, 2024

si-bors-ng bot commented Apr 2, 2024

try

britmyerss commented Apr 2, 2024

si-bors-ng bot commented Apr 2, 2024

britmyerss commented Mar 26, 2024 •

edited

Loading

nickgerace Mar 27, 2024 •

edited

Loading

nickgerace Mar 27, 2024 •

edited

Loading