Merge 'Fix non-determinism in simulator ' from Pedro Muniz

Use IndexSet for order preserving iteration instead of HashSet

Closes #2934

Cargo.lock

@@ -1800,9 +1800,9 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "2.10.0"
+version = "2.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fe4cd85333e22411419a0bcae1297d25e58c9443848b11dc6a86fefe8c78a661"
+checksum = "f2481980430f9f78649238835720ddccc57e52df14ffce1c6f37391d61b563e9"
 dependencies = [
  "equivalent",
  "hashbrown 0.15.2",
@@ -1822,7 +1822,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "232929e1d75fe899576a3d5c7416ad0d88dbfbb3c3d6aa00873a7408a50ddb88"
 dependencies = [
  "ahash",
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "is-terminal",
  "itoa",
  "log",
@@ -2192,7 +2192,7 @@ dependencies = [
  "env_logger 0.10.2",
  "garde",
  "hex",
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "itertools 0.14.0",
  "json5",
  "log",
@@ -2774,7 +2774,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "eac26e981c03a6e53e0aee43c113e3202f5581d5360dae7bd2c70e800dd0451d"
 dependencies = [
  "base64",
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "quick-xml 0.32.0",
  "serde",
  "time",
@@ -3672,6 +3672,7 @@ dependencies = [
  "anyhow",
  "garde",
  "hex",
+ "indexmap 2.11.0",
  "itertools 0.14.0",
  "rand 0.9.2",
  "rand_chacha 0.9.0",
@@ -4064,7 +4065,7 @@ version = "0.8.22"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "05ae329d1f08c4d17a59bed7ff5b5a769d062e64a62d34a3261b219e62cd5aae"
 dependencies = [
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "serde",
  "serde_spanned",
  "toml_datetime",
@@ -4086,7 +4087,7 @@ version = "0.22.26"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "310068873db2c5b3e7659d2cc35d21855dbafa50d1ce336397c666e3cb08137e"
 dependencies = [
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "serde",
  "serde_spanned",
  "toml_datetime",
@@ -4371,7 +4372,7 @@ dependencies = [
  "cc",
  "env_logger 0.11.7",
  "fallible-iterator",
- "indexmap 2.10.0",
+ "indexmap 2.11.0",
  "log",
  "memchr",
  "miette",

simulator/generation/plan.rs

@@ -1,11 +1,11 @@
 use std::{
-    collections::HashSet,
     fmt::{Debug, Display},
     path::Path,
     sync::Arc,
     vec,
 };
 
+use indexmap::IndexSet;
 use serde::{Deserialize, Serialize};
 
 use sql_generation::{
@@ -166,13 +166,13 @@ impl Interactions {
 }
 
 impl Interactions {
-    pub(crate) fn dependencies(&self) -> HashSet<String> {
+    pub(crate) fn dependencies(&self) -> IndexSet<String> {
         match self {
             Interactions::Property(property) => {
                 property
                     .interactions()
                     .iter()
-                    .fold(HashSet::new(), |mut acc, i| match i {
+                    .fold(IndexSet::new(), |mut acc, i| match i {
                         Interaction::Query(q) => {
                             acc.extend(q.dependencies());
                             acc
@@ -181,7 +181,7 @@ impl Interactions {
                     })
             }
             Interactions::Query(query) => query.dependencies(),
-            Interactions::Fault(_) => HashSet::new(),
+            Interactions::Fault(_) => IndexSet::new(),
         }
     }
 

simulator/model/mod.rs

@@ -1,6 +1,7 @@
-use std::{collections::HashSet, fmt::Display};
+use std::fmt::Display;
 
 use anyhow::Context;
+use indexmap::IndexSet;
 use itertools::Itertools;
 use serde::{Deserialize, Serialize};
 use sql_generation::model::{
@@ -32,19 +33,19 @@ pub enum Query {
 }
 
 impl Query {
-    pub fn dependencies(&self) -> HashSet<String> {
+    pub fn dependencies(&self) -> IndexSet<String> {
         match self {
             Query::Select(select) => select.dependencies(),
-            Query::Create(_) => HashSet::new(),
+            Query::Create(_) => IndexSet::new(),
             Query::Insert(Insert::Select { table, .. })
             | Query::Insert(Insert::Values { table, .. })
             | Query::Delete(Delete { table, .. })
             | Query::Update(Update { table, .. })
-            | Query::Drop(Drop { table, .. }) => HashSet::from_iter([table.clone()]),
+            | Query::Drop(Drop { table, .. }) => IndexSet::from_iter([table.clone()]),
             Query::CreateIndex(CreateIndex { table_name, .. }) => {
-                HashSet::from_iter([table_name.clone()])
+                IndexSet::from_iter([table_name.clone()])
             }
-            Query::Begin(_) | Query::Commit(_) | Query::Rollback(_) => HashSet::new(),
+            Query::Begin(_) | Query::Commit(_) | Query::Rollback(_) => IndexSet::new(),
         }
     }
     pub fn uses(&self) -> Vec<String> {

simulator/runner/env.rs

@@ -185,59 +185,6 @@ impl SimulatorEnv {
     ) -> Self {
         let mut rng = ChaCha8Rng::seed_from_u64(seed);
 
-        let total = 100.0;
-
-        let mut create_percent = 0.0;
-        let mut create_index_percent = 0.0;
-        let mut drop_percent = 0.0;
-        let mut delete_percent = 0.0;
-        let mut update_percent = 0.0;
-
-        let read_percent = rng.random_range(0.0..=total);
-        let write_percent = total - read_percent;
-
-        if !cli_opts.disable_create {
-            // Create percent should be 5-15% of the write percent
-            create_percent = rng.random_range(0.05..=0.15) * write_percent;
-        }
-        if !cli_opts.disable_create_index {
-            // Create indexpercent should be 2-5% of the write percent
-            create_index_percent = rng.random_range(0.02..=0.05) * write_percent;
-        }
-        if !cli_opts.disable_drop {
-            // Drop percent should be 2-5% of the write percent
-            drop_percent = rng.random_range(0.02..=0.05) * write_percent;
-        }
-        if !cli_opts.disable_delete {
-            // Delete percent should be 10-20% of the write percent
-            delete_percent = rng.random_range(0.1..=0.2) * write_percent;
-        }
-        if !cli_opts.disable_update {
-            // Update percent should be 10-20% of the write percent
-            // TODO: freestyling the percentage
-            update_percent = rng.random_range(0.1..=0.2) * write_percent;
-        }
-
-        let write_percent = write_percent
-            - create_percent
-            - create_index_percent
-            - delete_percent
-            - drop_percent
-            - update_percent;
-
-        let summed_total: f64 = read_percent
-            + write_percent
-            + create_percent
-            + create_index_percent
-            + drop_percent
-            + update_percent
-            + delete_percent;
-
-        let abs_diff = (summed_total - total).abs();
-        if abs_diff > 0.0001 {
-            panic!("Summed total {summed_total} is not equal to total {total}");
-        }
-
         let opts = SimulatorOpts {
             seed,
             ticks: rng.random_range(cli_opts.minimum_tests..=cli_opts.maximum_tests),

sql_generation/Cargo.toml

@@ -21,6 +21,7 @@ anyhow = { workspace = true }
 tracing = { workspace = true }
 schemars = { workspace = true }
 garde = { workspace = true, features = ["derive", "serde"] }
+indexmap = { version = "2.11.0" }
 
 [dev-dependencies]
 rand_chacha = "0.9.0"

sql_generation/generation/query.rs

@@ -10,6 +10,7 @@ use crate::model::query::select::{
 use crate::model::query::update::Update;
 use crate::model::query::{Create, CreateIndex, Delete, Drop, Insert, Select};
 use crate::model::table::{JoinTable, JoinType, JoinedTable, SimValue, Table, TableContext};
+use indexmap::IndexSet;
 use itertools::Itertools;
 use rand::Rng;
 use turso_parser::ast::{Expr, SortOrder};
@@ -104,7 +105,7 @@ impl Arbitrary for SelectInner {
                 if order_by_col_count == 0 {
                     return None;
                 }
-                let mut col_names = std::collections::HashSet::new();
+                let mut col_names = IndexSet::new();
                 let mut order_by_cols = Vec::new();
                 while order_by_cols.len() < order_by_col_count {
                     let table = pick(&order_by_table_candidates, rng);

sql_generation/generation/table.rs

@@ -1,5 +1,4 @@
-use std::collections::HashSet;
-
+use indexmap::IndexSet;
 use rand::Rng;
 use turso_core::Value;
 
@@ -28,7 +27,7 @@ impl Arbitrary for Table {
         } else {
             rng.random_range(opts.column_range)
         } as usize;
-        let mut column_set = HashSet::with_capacity(column_size);
+        let mut column_set = IndexSet::with_capacity(column_size);
         for col in std::iter::repeat_with(|| Column::arbitrary(rng, context)) {
             column_set.insert(col);
             if column_set.len() == column_size {

sql_generation/model/query/select.rs

@@ -1,6 +1,7 @@
-use std::{collections::HashSet, fmt::Display};
+use std::fmt::Display;
 
 pub use ast::Distinctness;
+use indexmap::IndexSet;
 use itertools::Itertools;
 use serde::{Deserialize, Serialize};
 use turso_parser::ast::{
@@ -105,12 +106,12 @@ impl Select {
         }
     }
 
-    pub fn dependencies(&self) -> HashSet<String> {
+    pub fn dependencies(&self) -> IndexSet<String> {
         if self.body.select.from.is_none() {
-            return HashSet::new();
+            return IndexSet::new();
         }
         let from = self.body.select.from.as_ref().unwrap();
-        let mut tables = HashSet::new();
+        let mut tables = IndexSet::new();
         tables.insert(from.table.clone());
 
         tables.extend(from.dependencies());