Store parsed regex::Regex in Value. (#1837)

This commit is contained in:
Finn Bear 2023-04-23 04:55:18 -07:00 committed by GitHub
parent bc7471a6ad
commit f159187dbe
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 98 additions and 53 deletions

View file

@ -66,10 +66,10 @@ pub fn point((arg1, arg2): (Value, Option<Value>)) -> Result<Value, Error> {
} }
pub fn regex((arg,): (Value,)) -> Result<Value, Error> { pub fn regex((arg,): (Value,)) -> Result<Value, Error> {
match arg { Ok(match arg {
Value::Strand(v) => Ok(Value::Regex(v.as_str().into())), Value::Strand(v) => v.parse().map(Value::Regex).unwrap_or(Value::None),
_ => Ok(Value::None), _ => Value::None,
} })
} }
pub fn string((arg,): (Strand,)) -> Result<Value, Error> { pub fn string((arg,): (Strand,)) -> Result<Value, Error> {

View file

@ -4,59 +4,120 @@ use nom::bytes::complete::escaped;
use nom::bytes::complete::is_not; use nom::bytes::complete::is_not;
use nom::character::complete::anychar; use nom::character::complete::anychar;
use nom::character::complete::char; use nom::character::complete::char;
use serde::{Deserialize, Serialize}; use serde::{
use std::fmt; de::{self, Visitor},
use std::ops::Deref; Deserialize, Deserializer, Serialize, Serializer,
};
use std::cmp::Ordering;
use std::fmt::Debug;
use std::fmt::{self, Display, Formatter};
use std::hash::{Hash, Hasher};
use std::str; use std::str;
use std::str::FromStr;
pub(crate) const TOKEN: &str = "$surrealdb::private::sql::Regex"; pub(crate) const TOKEN: &str = "$surrealdb::private::sql::Regex";
#[derive(Clone, Debug, Default, Eq, Ord, PartialEq, PartialOrd, Deserialize, Hash)] #[derive(Clone)]
pub struct Regex(pub(super) String); pub struct Regex(pub(super) regex::Regex);
impl From<&str> for Regex { impl Regex {
fn from(r: &str) -> Self { // Deref would expose `regex::Regex::as_str` which wouldn't have the '/' delimiters.
Self(r.replace("\\/", "/")) pub fn regex(&self) -> &regex::Regex {
}
}
impl Deref for Regex {
type Target = String;
fn deref(&self) -> &Self::Target {
&self.0 &self.0
} }
} }
impl fmt::Display for Regex { impl FromStr for Regex {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { type Err = <regex::Regex as FromStr>::Err;
write!(f, "/{}/", &self.0)
fn from_str(s: &str) -> Result<Self, Self::Err> {
regex::Regex::new(&s.replace("\\/", "/")).map(Self)
} }
} }
impl Regex { impl PartialEq for Regex {
pub fn regex(&self) -> Option<regex::Regex> { fn eq(&self, other: &Self) -> bool {
regex::Regex::new(&self.0).ok() self.0.as_str().eq(other.0.as_str())
}
}
impl Eq for Regex {}
impl Ord for Regex {
fn cmp(&self, other: &Self) -> Ordering {
self.0.as_str().cmp(other.0.as_str())
}
}
impl PartialOrd for Regex {
fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
Some(self.cmp(other))
}
}
impl Hash for Regex {
fn hash<H: Hasher>(&self, state: &mut H) {
self.0.as_str().hash(state);
}
}
impl Debug for Regex {
fn fmt(&self, f: &mut Formatter) -> fmt::Result {
Display::fmt(self, f)
}
}
impl Display for Regex {
fn fmt(&self, f: &mut Formatter) -> fmt::Result {
write!(f, "/{}/", &self.0)
} }
} }
impl Serialize for Regex { impl Serialize for Regex {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error> fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where where
S: serde::Serializer, S: Serializer,
{ {
if is_internal_serialization() { if is_internal_serialization() {
serializer.serialize_newtype_struct(TOKEN, &self.0) serializer.serialize_newtype_struct(TOKEN, self.0.as_str())
} else { } else {
serializer.serialize_none() serializer.serialize_none()
} }
} }
} }
impl<'de> Deserialize<'de> for Regex {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
struct RegexVisitor;
impl<'de> Visitor<'de> for RegexVisitor {
type Value = Regex;
fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
formatter.write_str("a regex str")
}
fn visit_str<E>(self, value: &str) -> Result<Self::Value, E>
where
E: de::Error,
{
Regex::from_str(value).map_err(|_| de::Error::custom("invalid regex"))
}
}
deserializer.deserialize_str(RegexVisitor)
}
}
pub fn regex(i: &str) -> IResult<&str, Regex> { pub fn regex(i: &str) -> IResult<&str, Regex> {
let (i, _) = char('/')(i)?; let (i, _) = char('/')(i)?;
let (i, v) = escaped(is_not("\\/"), '\\', anychar)(i)?; let (i, v) = escaped(is_not("\\/"), '\\', anychar)(i)?;
let (i, _) = char('/')(i)?; let (i, _) = char('/')(i)?;
Ok((i, Regex::from(v))) let regex = v.parse().map_err(|_| nom::Err::Error(crate::sql::Error::Parser(v)))?;
Ok((i, regex))
} }
#[cfg(test)] #[cfg(test)]
@ -71,7 +132,7 @@ mod tests {
assert!(res.is_ok()); assert!(res.is_ok());
let out = res.unwrap().1; let out = res.unwrap().1;
assert_eq!("/test/", format!("{}", out)); assert_eq!("/test/", format!("{}", out));
assert_eq!(out, Regex::from("test")); assert_eq!(out, "test".parse().unwrap());
} }
#[test] #[test]
@ -81,6 +142,6 @@ mod tests {
assert!(res.is_ok()); assert!(res.is_ok());
let out = res.unwrap().1; let out = res.unwrap().1;
assert_eq!(r"/(?i)test/[a-z]+/\s\d\w{1}.*/", format!("{}", out)); assert_eq!(r"/(?i)test/[a-z]+/\s\d\w{1}.*/", format!("{}", out));
assert_eq!(out, Regex::from(r"(?i)test/[a-z]+/\s\d\w{1}.*")); assert_eq!(out, r"(?i)test/[a-z]+/\s\d\w{1}.*".parse().unwrap());
} }
} }

View file

@ -17,7 +17,6 @@ use crate::sql::Future;
use crate::sql::Ident; use crate::sql::Ident;
use crate::sql::Idiom; use crate::sql::Idiom;
use crate::sql::Param; use crate::sql::Param;
use crate::sql::Regex;
use crate::sql::Strand; use crate::sql::Strand;
use crate::sql::Table; use crate::sql::Table;
use crate::sql::Uuid; use crate::sql::Uuid;
@ -211,7 +210,7 @@ impl ser::Serializer for Serializer {
value.serialize(ser::block::entry::vec::Serializer.wrap())?, value.serialize(ser::block::entry::vec::Serializer.wrap())?,
))))), ))))),
sql::regex::TOKEN => { sql::regex::TOKEN => {
Ok(Value::Regex(Regex(value.serialize(ser::string::Serializer.wrap())?))) Ok(Value::Regex(value.serialize(ser::string::Serializer.wrap())?.parse().unwrap()))
} }
sql::table::TOKEN => { sql::table::TOKEN => {
Ok(Value::Table(Table(value.serialize(ser::string::Serializer.wrap())?))) Ok(Value::Table(Table(value.serialize(ser::string::Serializer.wrap())?)))
@ -743,7 +742,7 @@ mod tests {
#[test] #[test]
fn regex() { fn regex() {
let regex = Regex::default(); let regex = "abc".parse().unwrap();
let value = to_value(&regex).unwrap(); let value = to_value(&regex).unwrap();
let expected = Value::Regex(regex); let expected = Value::Regex(regex);
assert_eq!(value, expected); assert_eq!(value, expected);

View file

@ -1179,22 +1179,13 @@ impl Value {
Value::False => other.is_false(), Value::False => other.is_false(),
Value::Thing(v) => match other { Value::Thing(v) => match other {
Value::Thing(w) => v == w, Value::Thing(w) => v == w,
Value::Regex(w) => match w.regex() { Value::Regex(w) => w.regex().is_match(v.to_string().as_str()),
Some(ref r) => r.is_match(v.to_string().as_str()),
None => false,
},
_ => false, _ => false,
}, },
Value::Regex(v) => match other { Value::Regex(v) => match other {
Value::Regex(w) => v == w, Value::Regex(w) => v == w,
Value::Number(w) => match v.regex() { Value::Number(w) => v.regex().is_match(w.to_string().as_str()),
Some(ref r) => r.is_match(w.to_string().as_str()), Value::Strand(w) => v.regex().is_match(w.as_str()),
None => false,
},
Value::Strand(w) => match v.regex() {
Some(ref r) => r.is_match(w.as_str()),
None => false,
},
_ => false, _ => false,
}, },
Value::Uuid(v) => match other { Value::Uuid(v) => match other {
@ -1211,19 +1202,13 @@ impl Value {
}, },
Value::Strand(v) => match other { Value::Strand(v) => match other {
Value::Strand(w) => v == w, Value::Strand(w) => v == w,
Value::Regex(w) => match w.regex() { Value::Regex(w) => w.regex().is_match(v.as_str()),
Some(ref r) => r.is_match(v.as_str()),
None => false,
},
_ => v == &other.to_strand(), _ => v == &other.to_strand(),
}, },
Value::Number(v) => match other { Value::Number(v) => match other {
Value::Number(w) => v == w, Value::Number(w) => v == w,
Value::Strand(_) => v == &other.to_number(), Value::Strand(_) => v == &other.to_number(),
Value::Regex(w) => match w.regex() { Value::Regex(w) => w.regex().is_match(v.to_string().as_str()),
Some(ref r) => r.is_match(v.to_string().as_str()),
None => false,
},
_ => false, _ => false,
}, },
Value::Geometry(v) => match other { Value::Geometry(v) => match other {
@ -1886,7 +1871,7 @@ mod tests {
assert_eq!(24, std::mem::size_of::<crate::sql::table::Table>()); assert_eq!(24, std::mem::size_of::<crate::sql::table::Table>());
assert_eq!(56, std::mem::size_of::<crate::sql::thing::Thing>()); assert_eq!(56, std::mem::size_of::<crate::sql::thing::Thing>());
assert_eq!(48, std::mem::size_of::<crate::sql::model::Model>()); assert_eq!(48, std::mem::size_of::<crate::sql::model::Model>());
assert_eq!(24, std::mem::size_of::<crate::sql::regex::Regex>()); assert_eq!(16, std::mem::size_of::<crate::sql::regex::Regex>());
assert_eq!(8, std::mem::size_of::<Box<crate::sql::range::Range>>()); assert_eq!(8, std::mem::size_of::<Box<crate::sql::range::Range>>());
assert_eq!(8, std::mem::size_of::<Box<crate::sql::edges::Edges>>()); assert_eq!(8, std::mem::size_of::<Box<crate::sql::edges::Edges>>());
assert_eq!(8, std::mem::size_of::<Box<crate::sql::function::Function>>()); assert_eq!(8, std::mem::size_of::<Box<crate::sql::function::Function>>());