From 3d8feb1c8af3aa8b4b4a1a85c3d2497966e64f13 Mon Sep 17 00:00:00 2001
From: Lining Pan <lining.pan@yale.edu>
Date: Sat, 2 Jul 2022 16:43:38 -0400
Subject: [PATCH 1/2] add instruction, register, and cursor state memorization

---
 sqlx-core/src/sqlite/connection/explain.rs | 77 ++++++++++++++++++++--
 1 file changed, 70 insertions(+), 7 deletions(-)

diff --git a/sqlx-core/src/sqlite/connection/explain.rs b/sqlx-core/src/sqlite/connection/explain.rs
index 101ca56a2b..0dca29ae2c 100644
--- a/sqlx-core/src/sqlite/connection/explain.rs
+++ b/sqlx-core/src/sqlite/connection/explain.rs
@@ -4,6 +4,7 @@ use crate::sqlite::connection::{execute, ConnectionState};
 use crate::sqlite::type_info::DataType;
 use crate::sqlite::SqliteTypeInfo;
 use crate::HashMap;
+use std::collections::HashSet;
 use std::str::from_utf8;
 
 // affinity
@@ -117,7 +118,7 @@ const OP_CONCAT: &str = "Concat";
 const OP_RESULT_ROW: &str = "ResultRow";
 const OP_HALT: &str = "Halt";
 
-#[derive(Debug, Copy, Clone, Eq, PartialEq)]
+#[derive(Debug, Copy, Clone, Eq, PartialEq, Hash)]
 struct ColumnType {
     pub datatype: DataType,
     pub nullable: Option<bool>,
@@ -141,7 +142,7 @@ impl ColumnType {
     }
 }
 
-#[derive(Debug, Clone, Eq, PartialEq)]
+#[derive(Debug, Clone, Eq, PartialEq, Hash)]
 enum RegDataType {
     Single(ColumnType),
     Record(Vec<ColumnType>),
@@ -178,7 +179,7 @@ impl RegDataType {
     }
 }
 
-#[derive(Debug, Clone, Eq, PartialEq)]
+#[derive(Debug, Clone, Eq, PartialEq, )]
 enum CursorDataType {
     Normal(HashMap<i64, ColumnType>),
     Pseudo(i64),
@@ -323,6 +324,49 @@ struct QueryState {
     pub result: Option<Vec<(Option<SqliteTypeInfo>, Option<bool>)>>,
 }
 
+#[derive(Debug, Hash, PartialEq, Eq)]
+struct BranchStateHash {
+    instruction: usize,
+    registers: Vec<(i64, RegDataType)>,
+    cursors: Vec<(i64, i64, Option<ColumnType>)>,
+}
+
+impl BranchStateHash {
+    pub fn from_query_state(st: &QueryState) -> Self {
+        let mut reg = vec![];
+        for (k, v) in &st.r {
+            reg.push((*k, v.clone()));
+        }
+        reg.sort_by_key(|v| v.0);
+
+        let mut cur = vec![];
+        for (k, v) in &st.p {
+            match v {
+                CursorDataType::Normal(hm) => {
+                    for (i, col) in hm {
+                        cur.push((*k, *i, Some(col.clone())));
+                    }
+                }
+                CursorDataType::Pseudo(i) => {
+                    cur.push((*k, *i, None));
+                }
+            }
+        }
+        cur.sort_by(|a, b| {
+            if a.0 == b.0 {
+                a.1.cmp(&b.1)
+            } else {
+                a.0.cmp(&b.0)
+            }
+        });
+        Self {
+            instruction: st.program_i,
+            registers: reg,
+            cursors: cur,
+        }
+    }
+}
+
 // Opcode Reference: https://sqlite.org/opcode.html
 pub(super) fn explain(
     conn: &mut ConnectionState,
@@ -348,6 +392,8 @@ pub(super) fn explain(
         result: None,
     }];
 
+    let mut visited_branch_state: HashSet<BranchStateHash> = HashSet::new();
+
     let mut result_states = Vec::new();
 
     while let Some(mut state) = states.pop() {
@@ -389,7 +435,12 @@ pub(super) fn explain(
 
                     let mut branch_state = state.clone();
                     branch_state.program_i = p2 as usize;
-                    states.push(branch_state);
+
+                    let bs_hash = BranchStateHash::from_query_state(&branch_state);
+                    if !visited_branch_state.contains(&bs_hash) {
+                        visited_branch_state.insert(bs_hash);
+                        states.push(branch_state);
+                    }
 
                     state.program_i += 1;
                     continue;
@@ -473,15 +524,27 @@ pub(super) fn explain(
 
                     let mut branch_state = state.clone();
                     branch_state.program_i = p1 as usize;
-                    states.push(branch_state);
+                    let bs_hash = BranchStateHash::from_query_state(&branch_state);
+                    if !visited_branch_state.contains(&bs_hash) {
+                        visited_branch_state.insert(bs_hash);
+                        states.push(branch_state);
+                    }
 
                     let mut branch_state = state.clone();
                     branch_state.program_i = p2 as usize;
-                    states.push(branch_state);
+                    let bs_hash = BranchStateHash::from_query_state(&branch_state);
+                    if !visited_branch_state.contains(&bs_hash) {
+                        visited_branch_state.insert(bs_hash);
+                        states.push(branch_state);
+                    }
 
                     let mut branch_state = state.clone();
                     branch_state.program_i = p3 as usize;
-                    states.push(branch_state);
+                    let bs_hash = BranchStateHash::from_query_state(&branch_state);
+                    if !visited_branch_state.contains(&bs_hash) {
+                        visited_branch_state.insert(bs_hash);
+                        states.push(branch_state);
+                    }
                 }
 
                 OP_COLUMN => {

From 92289c087c36d343517209ccfbea44ed3e765998 Mon Sep 17 00:00:00 2001
From: Lining Pan <lining.pan@yale.edu>
Date: Mon, 1 Aug 2022 13:22:02 -0400
Subject: [PATCH 2/2] fix: fixed formating

---
 sqlx-core/src/sqlite/connection/explain.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sqlx-core/src/sqlite/connection/explain.rs b/sqlx-core/src/sqlite/connection/explain.rs
index 0dca29ae2c..097dcccf9f 100644
--- a/sqlx-core/src/sqlite/connection/explain.rs
+++ b/sqlx-core/src/sqlite/connection/explain.rs
@@ -179,7 +179,7 @@ impl RegDataType {
     }
 }
 
-#[derive(Debug, Clone, Eq, PartialEq, )]
+#[derive(Debug, Clone, Eq, PartialEq)]
 enum CursorDataType {
     Normal(HashMap<i64, ColumnType>),
     Pseudo(i64),