1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
/* Copyright (c) 2020, ZomboDB, LLC Permission to use, copy, modify, and distribute this software and its documentation for any purpose, without fee, and without a written agreement is hereby granted, provided that the above copyright notice and this paragraph and the following two paragraphs appear in all copies. IN NO EVENT SHALL ZomboDB, LLC BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, INCLUDING LOST PROFITS, ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF ZomboDB, LLC HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ZomboDB, LLC SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND ZomboDB, LLC HAS NO OBLIGATIONS TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS. */ //! `postgres-parser` is a safe wrapper around Postgres' SQL query parser. //! //! Its primary purpose is to easily syntax check SQL statements (individually or en masse) in //! addition to providing a parse tree that can be walked, examined, mutated, or transformed. //! //! ## Technical Details //! //! First things first, this crate, as part of its build process, downloads the Postgres source //! code, builds Postgres to LLVM IR (and then bitcode), which is ultimately statically linked into //! the resulting Rust library (rlib), relying on LLVM's "link time optimization" (LTO) features to //! reduce the Postgres LLVM bitcode to just the symbols/code required by this crate. //! //! That's a lot of work, and it requires that building this create, or even crates that use this //! crate as a dependency, have the LLVM toolchain on the system `$PATH`. //! //! The justification for this is that, despite the build complexity, we can always stay current //! with Postgres as it evolves its SQL support and, thus, its parser. //! //! ## What's in the Box? //! //! There's three primary things. The first two are safe interfaces into parsing SQL statements //! and evaluating the resulting parse trees. The third is the set of completely unsafe functions //! and structs upon which the first two are built. //! //! ### `parse_query()` and the `nodes` module //! //! The `parse_query()` function parses a string of SQL statements and returns a Vec of parsed //! nodes, or a parse error //! //! A quick example: //! //! ```rust //! use postgres_parser::{parse_query, PgParserError}; //! let parsetree = parse_query("SELECT * FROM my_table WHERE id = 42; SELECT 2;"); //! match parsetree { //! Ok(nodes) => { //! // one node for each statement parsed from the input string above //! for node in nodes { //! // debug-print the node for this query //! println!("{:#?}", node); //! } //! } //! //! // one possible error, for the first mal-formed SQL statement //! Err(e) => { //! panic!(e); //! } //! } //! ``` //! //! The nodes represented in the parse tree live in the `postgres_parser::nodes` module. The //! top-level node is simply called `Node` and is an enum with a variant for every possible node //! type. //! //! An example of walking a parsetree and examining the expected Node: //! //! ```rust //! use postgres_parser::{parse_query, Node, join_name_list}; //! let parsetree = parse_query("DROP TABLE my_schema.my_table;"); //! match parsetree { //! Ok(mut nodes) => { //! let node = nodes.pop().unwrap(); // we know we only have 1 node here //! match node { //! Node::DropStmt(dropstmt) => { //! // dropstmt.object is a Vec<Node>, where each Node is a Node::List of //! // ultimately, Node::Value, where each value is a String //! for object in dropstmt.objects.unwrap() { //! // join_name_list() will figure out the hard part for us //! // this is a common pattern throughout Postgres' parsetree //! let name = join_name_list(&object).unwrap(); //! assert_eq!(name, "my_schema.my_table"); //! } //! } //! //! _ => panic!("unexpected node: {:#?}", node), //! } //! } //! //! // one possible error, for the first mal-formed SQL statement //! Err(e) => { //! panic!(e); //! } //! } //! ``` //! //! ### The `sys` module //! //! The `sys` module is a 100% "bindgen"-generated module from Postgres' header files. In general, //! it's not expected that users of this crate will interact with this module. //! //! It is upon the items in this module that the rest of `postgres-parser` is built. The module is //! public for completeness only. //! //! //! ### `SqlStatementScanner` //! //! The `SqlStatementScanner` is a simple type intended to work as an iterator over scanning and //! parsing a single string of multiple SQL statements, one at a time. //! //! This is particullary useful to report statement-level parse errors, as opposed to the `parse_query()` //! function that simply reports one error for the entire string. //! //! A quick example: //! //! ```rust //! use postgres_parser::SqlStatementScanner; //! let mut scanner = SqlStatementScanner::new("SELECT 1;\nSELECT 2;").into_iter(); //! //! let first = scanner.next().expect("no first query"); //! assert_eq!(first.sql, "SELECT 1;\n"); // note trailing \n -- trailing whitespace after ';' is included //! assert!(first.payload.is_none()); //! assert!(first.parsetree.is_ok()); //! //! let second = scanner.next().expect("no second query"); //! assert_eq!(second.sql, "SELECT 2;"); //! assert!(second.payload.is_none()); //! assert!(second.parsetree.is_ok()); //! //! assert!(scanner.next().is_none()); //! ``` //! //! ## Serde Support //! //! All the parse tree Node structures supported are `Deserialize, Serialize`, and as such, can be //! directly used by any of the serde serializers, including serde_json. //! //! ```rust //! use postgres_parser::parse_query; //! let as_json = serde_json::to_string_pretty(&parse_query("SELECT 1;")).expect("failed to convert to json"); //! println!("{}", as_json); //! ``` //! //! The above would output: //! //! ```json //! {"SelectStmt":{"targetList":[{"ResTarget":{"val":{"A_Const":{"val":{"int":1},"location":7}},"location":7}}],"op":"SETOP_NONE","all":false}} //! ``` //! //! //! ## Notes on Thread Safety //! //! Postgres is, by design, not thread safe. Rust, on the other hand, is. As we're literally //! statically linking against the compiled Postgres code, this presents an interesting problem. //! //! The solution `postgres-parser` has taken is that the `parse_query()` function (which is also //! used by `SqlStatementScanner`) is guarded under a Rust Mutex. As such, only one query can //! be parsed at a time. //! use serde::{Deserialize, Serialize}; mod convert; mod parser; mod scanner; pub mod nodes; pub mod sys; pub use nodes::Node; pub use parser::*; pub use scanner::*; /// Represents various errors that can occur while parsing a SQL statement. #[derive(Debug, Serialize, Deserialize, Eq, PartialEq)] pub enum PgParserError { /// The SQL statement could not be parsed ParseError { message: String, cursor_pos: i32 }, /// Internal SqlStatementScannerError -- more than one statement was scanned MultipleStatements(Vec<crate::Node>), /// We couldn't determine the error Postgres tried to tell us UnknownParseError, /// One of the returned parsed queries didn't parse to a `postgres-parser::RawStmt` (this is unlikely to ever happen) NotARawStmt, /// The input SQL statement `&str` contained an internal zero-byte InternalNull, /// The result from Postgres' parser was not a `Node::List` (this is unlikely to ever happen) NotAList, /// The evaluated Node is not a String NotAString, } /// A common pattern in Postgres' parse trees, when it needs to represent the name of a thing (a /// table, a field, a view, etc), especially when that name can be qualified, is to represent /// that name as a List of string Values. /// /// This is a helper function to join these qualified name Lists together into a single Rust string. /// /// It's expected that the `list_of_names` argument be a `postgres_parser::Node::List` variant, /// where each of its contained members is a `postgres_parser::Node::Value` where its `string` /// member `is_some()`. /// /// If all these conditions are true, the result will be a dot-delimited string of each part of /// the List of names (`Ok("the.qualified.name")`). /// /// If not, the result will either be `Err(PgParserError::NotAList)` or /// `Err(PgParserError::NotAString)`. pub fn join_name_list(list_of_names: &crate::Node) -> Result<String, PgParserError> { match list_of_names { crate::Node::List(names) => { let mut result = String::new(); for name in names { if !result.is_empty() { result.push('.'); } match name { crate::Node::Value(value) if value.string.is_some() => { result.push_str(value.string.as_ref().unwrap()) } _ => return Err(PgParserError::NotAString), } } Ok(result) } _ => Err(PgParserError::NotAList), } }