fix(query): fix stack overflow if excessive use of union all (#17475)

* fix(query): add stateless tests

* fix(query): add recursive for clone

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

* fix(query): try fix stackoverflow

Author: zhang2014

Cargo.lock

@@ -288,7 +288,7 @@ derive-visitor = { version = "0.4.0", features = ["std-types-drive"] }
 derive_more = { version = "1.0.0", features = ["full"] }
 dtparse = { git = "https://github.com/datafuse-extras/dtparse.git", rev = "30e28ca" }
 dyn-clone = "1.0.9"
-educe = "0.4" # FIXME: failed to upgrade to educe 0.6
+educe = { version = "0.6", features = ["default", "full"] }
 either = "1.9"
 enquote = "1.1.0"
 enum-as-inner = "0.6"
@@ -624,6 +624,7 @@ async-recursion = { git = "https://github.com/datafuse-extras/async-recursion.gi
 backtrace = { git = "https://github.com/rust-lang/backtrace-rs.git", rev = "72265be" }
 color-eyre = { git = "https://github.com/eyre-rs/eyre.git", rev = "e5d92c3" }
 deltalake = { git = "https://github.com/delta-io/delta-rs", rev = "3038c145" }
+educe = { git = "https://github.com/zhang2014/educe.git", rev = "ab34a19" }
 ethnum = { git = "https://github.com/datafuse-extras/ethnum-rs", rev = "4cb05f1" }
 openai_api_rust = { git = "https://github.com/datafuse-extras/openai-api", rev = "819a0ed" }
 openraft = { git = "https://github.com/databendlabs/openraft", tag = "v0.10.0-alpha.7" }

Cargo.toml

@@ -11,14 +11,14 @@ doctest = false
 test = true
 
 [dependencies]
-databend-common-ast = { workspace = true }
 
 anyhow = { workspace = true }
 arrow-flight = { workspace = true }
 arrow-schema = { workspace = true }
 backtrace = { workspace = true, features = ["std", "serialize-serde"] }
 bincode = { workspace = true }
 cidr = { workspace = true }
+derive-visitor = { workspace = true }
 geozero = { workspace = true }
 gimli = { workspace = true }
 http = { workspace = true }
@@ -32,6 +32,7 @@ paste = { workspace = true }
 prost = { workspace = true }
 redis = { workspace = true }
 reqwest = { workspace = true }
+rspack-codespan-reporting = { workspace = true }
 rustc-demangle = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }

src/common/exception/Cargo.toml

@@ -19,12 +19,12 @@ use std::fmt::Display;
 use std::fmt::Formatter;
 use std::marker::PhantomData;
 
-use databend_common_ast::span::pretty_print_error;
-use databend_common_ast::Span;
 use thiserror::Error;
 
 use crate::exception_backtrace::capture;
+use crate::span::pretty_print_error;
 use crate::ErrorFrame;
+use crate::Span;
 use crate::StackTrace;
 
 #[derive(Error)]

src/common/exception/src/exception.rs

@@ -17,12 +17,13 @@ use std::fmt::Debug;
 use std::fmt::Display;
 use std::fmt::Formatter;
 
-use databend_common_ast::Span;
 use geozero::error::GeozeroError;
 
 use crate::exception_backtrace::capture;
 use crate::ErrorCode;
 use crate::ErrorFrame;
+use crate::ParseError;
+use crate::Span;
 use crate::StackTrace;
 
 #[derive(thiserror::Error)]
@@ -221,8 +222,8 @@ impl From<std::string::FromUtf8Error> for ErrorCode {
     }
 }
 
-impl From<databend_common_ast::ParseError> for ErrorCode {
-    fn from(error: databend_common_ast::ParseError) -> Self {
+impl From<ParseError> for ErrorCode {
+    fn from(error: ParseError) -> Self {
         ErrorCode::SyntaxException(error.1).set_span(error.0)
     }
 }

src/common/exception/src/exception_into.rs

@@ -24,6 +24,8 @@ mod exception_backtrace;
 mod exception_code;
 mod exception_flight;
 mod exception_into;
+mod parser_error;
+mod span;
 
 pub use context::display_error_stack;
 pub use context::ErrorFrame;
@@ -35,3 +37,9 @@ pub use exception_backtrace::set_backtrace;
 pub use exception_backtrace::StackTrace;
 pub use exception_backtrace::USER_SET_ENABLE_BACKTRACE;
 pub use exception_into::SerializedError;
+pub use parser_error::ParseError;
+pub use span::merge_span;
+pub use span::offset_span;
+pub use span::pretty_print_error;
+pub use span::Range;
+pub use span::Span;

src/common/exception/src/lib.rs

@@ -18,8 +18,6 @@ use std::fmt::Formatter;
 use crate::span::pretty_print_error;
 use crate::Span;
 
-pub type Result<T> = std::result::Result<T, ParseError>;
-
 #[derive(Debug)]
 pub struct ParseError(pub Span, pub String);
 

src/query/ast/src/error.rs src/common/exception/src/parser_error.rssrc/query/ast/src/error.rs renamed to src/common/exception/src/parser_error.rs

@@ -12,7 +12,9 @@ repository = "https://github.com/datafuselabs/databend/tree/main/src/query/ast"
 doctest = false
 
 [dependencies]
+databend-common-exception = { workspace = true }
 derive-visitor = { workspace = true }
+educe = { workspace = true }
 enum-as-inner = { workspace = true }
 ethnum = { workspace = true }
 fast-float2 = { workspace = true }
@@ -27,7 +29,6 @@ percent-encoding = { workspace = true }
 pratt = { workspace = true }
 pretty_assertions = { workspace = true }
 recursive = { workspace = true }
-rspack-codespan-reporting = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }
 strsim = { workspace = true }

src/query/ast/src/span.rs src/common/exception/src/span.rssrc/query/ast/src/span.rs renamed to src/common/exception/src/span.rs

@@ -16,8 +16,10 @@ use std::collections::HashMap;
 use std::fmt::Display;
 use std::fmt::Formatter;
 
+use databend_common_exception::merge_span;
 use derive_visitor::Drive;
 use derive_visitor::DriveMut;
+use educe::Educe;
 use enum_as_inner::EnumAsInner;
 use ethnum::i256;
 use pratt::Affix;
@@ -40,12 +42,16 @@ use crate::ast::write_comma_separated_list;
 use crate::ast::Identifier;
 use crate::ast::Query;
 use crate::ast::SetExpr;
-use crate::span::merge_span;
 use crate::ParseError;
 use crate::Result;
 use crate::Span;
 
-#[derive(Debug, Clone, PartialEq, Drive, DriveMut)]
+#[derive(Educe, Drive, DriveMut)]
+#[educe(
+    PartialEq(bound = false, attrs = "#[recursive::recursive]"),
+    Clone(bound = false, attrs = "#[recursive::recursive]"),
+    Debug(bound = false, attrs = "#[recursive::recursive]")
+)]
 pub enum Expr {
     /// Column reference, with indirection like `table.column`
     ColumnRef {