This repository has been archived on 2021-03-14. You can view files and clone it, but cannot push or open issues or pull requests.
git-config/src/parser.rs

1446 lines
47 KiB
Rust
Raw Normal View History

2021-02-20 17:41:34 -08:00
//! This module handles parsing a `git-config` file. Generally speaking, you
//! want to use a higher abstraction such as [`GitConfig`] unless you have some
//! explicit reason to work with events instead.
2021-02-19 09:08:42 -08:00
//!
//! The general workflow for interacting with this is to use one of the
//! `parse_from_*` function variants. These will return a [`Parser`] on success,
//! which can be converted into an [`Event`] iterator. The [`Parser`] also has
//! additional methods for accessing leading comments or events by section.
2021-02-20 10:49:01 -08:00
//!
//! [`GitConfig`]: crate::config::GitConfig
2021-02-19 09:08:42 -08:00
2021-02-23 16:47:24 -08:00
use bstr::{BStr, ByteSlice};
2021-02-24 20:19:45 -08:00
use nom::branch::alt;
2021-02-19 16:41:58 -08:00
use nom::bytes::complete::{escaped, tag, take_till, take_while};
2021-02-18 09:49:47 -08:00
use nom::character::complete::{char, none_of, one_of};
use nom::character::{is_newline, is_space};
use nom::combinator::{map, opt};
use nom::error::{Error as NomError, ErrorKind};
2021-02-24 20:19:45 -08:00
use nom::multi::{many0, many1};
2021-02-18 09:49:47 -08:00
use nom::sequence::delimited;
use nom::IResult;
2021-02-21 18:16:08 -08:00
use std::fmt::Display;
use std::iter::FusedIterator;
2021-02-24 20:23:23 -08:00
use std::{borrow::Cow, error::Error};
2021-02-18 09:49:47 -08:00
2021-02-23 08:30:48 -08:00
/// Syntactic events that occurs in the config. Despite all these variants
2021-02-23 17:22:15 -08:00
/// holding a [`Cow`] instead over a simple reference, the parser will only emit
2021-02-23 08:30:48 -08:00
/// borrowed `Cow` variants.
///
/// The `Cow` smart pointer is used here for ease of inserting events in a
/// middle of an Event iterator. This is used, for example, in the [`GitConfig`]
/// struct when adding values.
///
/// [`Cow`]: std::borrow::Cow
/// [`GitConfig`]: crate::config::GitConfig
2021-02-18 21:12:59 -08:00
#[derive(Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug)]
2021-02-18 09:49:47 -08:00
pub enum Event<'a> {
2021-02-19 21:33:17 -08:00
/// A comment with a comment tag and the comment itself. Note that the
/// comment itself may contain additional whitespace and comment markers
/// at the beginning.
2021-02-18 21:12:59 -08:00
Comment(ParsedComment<'a>),
2021-02-19 21:33:17 -08:00
/// A section header containing the section name and a subsection, if it
/// exists.
2021-02-18 21:12:59 -08:00
SectionHeader(ParsedSectionHeader<'a>),
2021-02-19 21:33:17 -08:00
/// A name to a value in a section.
2021-02-23 16:47:24 -08:00
Key(Cow<'a, BStr>),
2021-02-19 21:33:17 -08:00
/// A completed value. This may be any string, including the empty string,
/// if an implicit boolean value is used. Note that these values may contain
/// spaces and any special character. This value is also unprocessed, so it
/// it may contain double quotes that should be replaced.
2021-02-23 16:47:24 -08:00
Value(Cow<'a, BStr>),
2021-02-18 09:49:47 -08:00
/// Represents any token used to signify a new line character. On Unix
/// platforms, this is typically just `\n`, but can be any valid newline
2021-02-21 18:16:08 -08:00
/// sequence. Multiple newlines (such as `\n\n`) will be merged as a single
/// newline event.
2021-02-23 16:47:24 -08:00
Newline(Cow<'a, BStr>),
2021-02-18 09:49:47 -08:00
/// Any value that isn't completed. This occurs when the value is continued
/// onto the next line. A Newline event is guaranteed after, followed by
2021-02-19 21:33:17 -08:00
/// either a ValueDone, a Whitespace, or another ValueNotDone.
2021-02-23 16:47:24 -08:00
ValueNotDone(Cow<'a, BStr>),
2021-02-18 09:49:47 -08:00
/// The last line of a value which was continued onto another line.
2021-02-23 16:47:24 -08:00
ValueDone(Cow<'a, BStr>),
2021-02-19 21:33:17 -08:00
/// A continuous section of insignificant whitespace. Values with internal
/// spaces will not be separated by this event.
2021-02-23 16:47:24 -08:00
Whitespace(Cow<'a, BStr>),
2021-02-23 08:30:48 -08:00
/// This event is emitted when the parser counters a valid `=` character
/// separating the key and value. This event is necessary as it eliminates
/// the ambiguity for whitespace events between a key and value event.
KeyValueSeparator,
2021-02-18 09:49:47 -08:00
}
2021-02-21 18:16:08 -08:00
impl Display for Event<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
Self::Comment(e) => e.fmt(f),
Self::SectionHeader(e) => e.fmt(f),
Self::Key(e) => e.fmt(f),
Self::Value(e) => e.fmt(f),
Self::Newline(e) => e.fmt(f),
Self::ValueNotDone(e) => e.fmt(f),
Self::ValueDone(e) => e.fmt(f),
Self::Whitespace(e) => e.fmt(f),
2021-02-23 08:30:48 -08:00
Self::KeyValueSeparator => write!(f, "="),
2021-02-21 18:16:08 -08:00
}
}
}
2021-02-19 23:28:50 -08:00
/// A parsed section containing the header and the section events.
2021-02-18 21:12:59 -08:00
#[derive(Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug, Default)]
pub struct ParsedSection<'a> {
2021-02-19 23:28:50 -08:00
/// The section name and subsection name, if any.
2021-02-19 09:08:42 -08:00
pub section_header: ParsedSectionHeader<'a>,
2021-02-19 23:28:50 -08:00
/// The syntactic events found in this section.
2021-02-19 09:08:42 -08:00
pub events: Vec<Event<'a>>,
2021-02-18 09:49:47 -08:00
}
2021-02-23 08:30:48 -08:00
2021-02-21 18:16:08 -08:00
impl Display for ParsedSection<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(f, "{}", self.section_header)?;
for event in &self.events {
event.fmt(f)?;
}
Ok(())
}
}
2021-02-18 09:49:47 -08:00
2021-02-23 16:47:24 -08:00
impl<'a> Into<Event<'a>> for ParsedSectionHeader<'a> {
fn into(self) -> Event<'a> {
Event::SectionHeader(self)
}
}
2021-02-19 09:08:42 -08:00
/// A parsed section header, containing a name and optionally a subsection name.
2021-02-21 11:43:41 -08:00
#[derive(Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug, Default)]
2021-02-18 21:12:59 -08:00
pub struct ParsedSectionHeader<'a> {
2021-02-19 23:28:50 -08:00
/// The name of the header.
2021-02-23 16:47:24 -08:00
pub name: Cow<'a, BStr>,
2021-02-19 16:18:02 -08:00
/// The separator used to determine if the section contains a subsection.
/// This is either a period `.` or a string of whitespace. Note that
/// reconstruction of subsection format is dependent on this value. If this
/// is all whitespace, then the subsection name needs to be surrounded by
/// quotes to have perfect reconstruction.
2021-02-23 16:47:24 -08:00
pub separator: Option<Cow<'a, BStr>>,
2021-02-19 16:18:02 -08:00
/// The subsection name without quotes if any exist.
2021-02-23 16:47:24 -08:00
pub subsection_name: Option<Cow<'a, BStr>>,
2021-02-18 21:12:59 -08:00
}
2021-02-21 18:16:08 -08:00
impl Display for ParsedSectionHeader<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(f, "[{}", self.name)?;
if let Some(v) = &self.separator {
v.fmt(f)?;
let subsection_name = self.subsection_name.as_ref().unwrap();
2021-02-23 16:47:24 -08:00
if *v == b".".as_bstr() {
subsection_name.fmt(f)?;
} else {
write!(f, "\"{}\"", subsection_name)?;
2021-02-21 18:16:08 -08:00
}
}
write!(f, "]")
}
}
2021-02-19 09:08:42 -08:00
/// A parsed comment event containing the comment marker and comment.
2021-02-21 11:43:41 -08:00
#[derive(Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug, Default)]
2021-02-18 21:12:59 -08:00
pub struct ParsedComment<'a> {
2021-02-19 23:28:50 -08:00
/// The comment marker used. This is either a semicolon or octothorpe.
2021-02-19 09:08:42 -08:00
pub comment_tag: char,
2021-02-19 23:28:50 -08:00
/// The parsed comment.
2021-02-23 16:47:24 -08:00
pub comment: Cow<'a, BStr>,
2021-02-18 09:49:47 -08:00
}
2021-02-21 18:16:08 -08:00
impl Display for ParsedComment<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
self.comment_tag.fmt(f)?;
self.comment.fmt(f)
}
}
2021-02-26 15:35:49 -08:00
/// A parser error reports the one-indexed line number where the parsing error
/// occurred, as well as the last parser node and the remaining data to be
/// parsed.
2021-02-18 21:12:59 -08:00
#[derive(PartialEq, Debug)]
2021-02-24 20:19:45 -08:00
pub struct ParserError<'a> {
line_number: usize,
2021-02-26 15:10:24 -08:00
last_attempted_parser: ParserNode,
2021-02-24 20:19:45 -08:00
parsed_until: &'a [u8],
2021-02-18 21:12:59 -08:00
}
2021-02-26 15:35:49 -08:00
impl ParserError<'_> {
/// The one-indexed line number where the error occurred. This is determined
/// by the number of newlines that were successfully parsed.
pub fn line_number(&self) -> usize {
self.line_number + 1
}
2021-02-25 08:11:26 -08:00
2021-02-26 15:35:49 -08:00
/// The remaining data that was left unparsed.
pub fn remaining_data(&self) -> &[u8] {
self.parsed_until
2021-02-25 08:11:26 -08:00
}
2021-02-24 20:19:45 -08:00
}
impl Display for ParserError<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
let data_size = self.parsed_until.len();
let data = std::str::from_utf8(self.parsed_until);
write!(
f,
2021-02-25 08:11:26 -08:00
"Got an unexpected token on line {} while trying to parse a {}: ",
self.line_number + 1,
2021-02-26 15:10:24 -08:00
self.last_attempted_parser,
2021-02-24 20:19:45 -08:00
)?;
match (data, data_size) {
(Ok(data), _) if data_size > 10 => write!(
f,
"'{}' ... ({} characters omitted)",
&data[..10],
data_size - 10
),
(Ok(data), _) => write!(f, "'{}'", data),
(Err(_), _) if data_size > 10 => write!(
f,
"'{:02x?}' ... ({} characters omitted)",
&self.parsed_until[..10],
data_size - 10
),
(Err(_), _) => write!(f, "'{:02x?}'", self.parsed_until),
}
2021-02-18 21:12:59 -08:00
}
}
2021-02-24 20:23:23 -08:00
impl Error for ParserError<'_> {}
2021-02-26 15:35:49 -08:00
#[derive(PartialEq, Debug, Clone, Copy)]
enum ParserNode {
SectionHeader,
ConfigName,
ConfigValue,
Comment,
}
impl Display for ParserNode {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
Self::SectionHeader => write!(f, "section header"),
Self::ConfigName => write!(f, "config name"),
Self::ConfigValue => write!(f, "config value"),
Self::Comment => write!(f, "comment"),
}
}
}
2021-02-18 21:12:59 -08:00
/// A zero-copy `git-config` file parser.
///
2021-02-19 23:28:50 -08:00
/// This is parser exposes low-level syntactic events from a `git-config` file.
/// Generally speaking, you'll want to use [`GitConfig`] as it wraps
/// around the parser to provide a higher-level abstraction to a `git-config`
/// file, including querying, modifying, and updating values.
///
/// This parser guarantees that the events emitted are sufficient to
/// reconstruct a `git-config` file identical to the source `git-config`.
2021-02-18 21:12:59 -08:00
///
/// # Differences between a `.ini` parser
///
/// While the `git-config` format closely resembles the [`.ini` file format],
/// there are subtle differences that make them incompatible. For one, the file
/// format is not well defined, and there exists no formal specification to
/// adhere to. Thus, attempting to use an `.ini` parser on a `git-config` file
/// may successfully parse invalid configuration files.
///
/// For concrete examples, some notable differences are:
/// - `git-config` sections permit subsections via either a quoted string
/// (`[some-section "subsection"]`) or via the deprecated dot notation
/// (`[some-section.subsection]`). Successful parsing these section names is not
/// well defined in typical `.ini` parsers. This parser will handle these cases
/// perfectly.
/// - Comment markers are not strictly defined either. This parser will always
/// and only handle a semicolon or octothorpe (also known as a hash or number
/// sign).
/// - Global properties may be allowed in `.ini` parsers, but is strictly
/// disallowed by this parser.
/// - Only `\t`, `\n`, `\b` `\\` are valid escape characters.
/// - Quoted and semi-quoted values will be parsed (but quotes will be included
/// in event outputs). An example of a semi-quoted value is `5"hello world"`,
/// which should be interpreted as `5hello world`.
/// - Line continuations via a `\` character is supported.
/// - Whitespace handling similarly follows the `git-config` specification as
/// closely as possible, where excess whitespace after a non-quoted value are
/// trimmed, and line continuations onto a new line with excess spaces are kept.
/// - Only equal signs (optionally padded by spaces) are valid name/value
/// delimiters.
///
/// Note that that things such as case-sensitivity or duplicate sections are
/// _not_ handled. This parser is a low level _syntactic_ interpreter (as a
/// parser should be), and higher level wrappers around this parser (which may
2021-02-19 09:08:42 -08:00
/// or may not be zero-copy) should handle _semantic_ values. This also means
/// that string-like values are not interpreted. For example, `hello"world"`
/// would be read at a high level as `helloworld` but this parser will return
/// the former instead, with the extra quotes. This is because it is not the
/// responsibility of the parser to interpret these values, and doing so would
/// necessarily require a copy, which this parser avoids.
2021-02-18 21:12:59 -08:00
///
/// # Trait Implementations
///
/// - This struct does _not_ implement [`FromStr`] due to lifetime
2021-02-19 23:28:50 -08:00
/// constraints implied on the required `from_str` method. Instead, it provides
2021-02-18 21:12:59 -08:00
/// [`Parser::from_str`].
///
2021-02-19 23:28:50 -08:00
/// # Idioms
///
/// If you do want to use this parser, there are some idioms that may help you
/// with interpreting sequences of events.
///
/// ## `Value` events do not immediately follow `Key` events
///
/// Consider the following `git-config` example:
///
/// ```text
/// [core]
/// autocrlf = input
/// ```
///
/// Because this parser guarantees perfect reconstruction, there are many
/// non-significant events that occur in addition to the ones you may expect:
///
/// ```
2021-02-24 12:47:35 -08:00
/// # use git_config::parser::{Event, ParsedSectionHeader, parse_from_str};
2021-02-21 11:43:41 -08:00
/// # use std::borrow::Cow;
2021-02-19 23:28:50 -08:00
/// # let section_header = ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
/// # name: Cow::Borrowed("core".into()),
2021-02-19 23:28:50 -08:00
/// # separator: None,
/// # subsection_name: None,
/// # };
/// # let section_data = "[core]\n autocrlf = input";
/// # assert_eq!(parse_from_str(section_data).unwrap().into_vec(), vec![
/// Event::SectionHeader(section_header),
2021-02-23 16:47:24 -08:00
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::Whitespace(Cow::Borrowed(" ".into())),
/// Event::Key(Cow::Borrowed("autocrlf".into())),
/// Event::Whitespace(Cow::Borrowed(" ".into())),
2021-02-23 08:30:48 -08:00
/// Event::KeyValueSeparator,
2021-02-23 16:47:24 -08:00
/// Event::Whitespace(Cow::Borrowed(" ".into())),
/// Event::Value(Cow::Borrowed("input".into())),
2021-02-19 23:28:50 -08:00
/// # ]);
/// ```
///
/// Note the two whitespace events between the key and value pair! Those two
/// events actually refer to the whitespace between the name and value and the
/// equal sign. So if the config instead had `autocrlf=input`, those whitespace
/// events would no longer be present.
///
2021-02-23 08:30:48 -08:00
/// ## `KeyValueSeparator` event is not guaranteed to emit
///
/// Consider the following `git-config` example:
///
/// ```text
/// [core]
/// autocrlf
/// ```
///
/// This is a valid config with a `autocrlf` key having an implicit `true`
/// value. This means that there is not a `=` separating the key and value,
/// which means that the corresponding event won't appear either:
///
/// ```
2021-02-24 12:47:35 -08:00
/// # use git_config::parser::{Event, ParsedSectionHeader, parse_from_str};
2021-02-23 08:30:48 -08:00
/// # use std::borrow::Cow;
2021-02-23 16:47:24 -08:00
/// # use bstr::BStr;
2021-02-23 08:30:48 -08:00
/// # let section_header = ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
/// # name: Cow::Borrowed("core".into()),
2021-02-23 08:30:48 -08:00
/// # separator: None,
/// # subsection_name: None,
/// # };
/// # let section_data = "[core]\n autocrlf";
/// # assert_eq!(parse_from_str(section_data).unwrap().into_vec(), vec![
/// Event::SectionHeader(section_header),
2021-02-23 16:47:24 -08:00
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::Whitespace(Cow::Borrowed(" ".into())),
/// Event::Key(Cow::Borrowed("autocrlf".into())),
/// Event::Value(Cow::Borrowed("".into())),
2021-02-23 08:30:48 -08:00
/// # ]);
/// ```
///
2021-02-19 23:28:50 -08:00
/// ## Quoted values are not unquoted
///
/// Consider the following `git-config` example:
///
/// ```text
/// [core]
/// autocrlf=true""
/// filemode=fa"lse"
/// ```
///
/// Both these events, when fully processed, should normally be `true` and
/// `false`. However, because this parser is zero-copy, we cannot process
/// partially quoted values, such as the `false` example. As a result, to
/// maintain consistency, the parser will just take all values as literals. The
/// relevant event stream emitted is thus emitted as:
///
/// ```
2021-02-24 12:47:35 -08:00
/// # use git_config::parser::{Event, ParsedSectionHeader, parse_from_str};
2021-02-21 11:43:41 -08:00
/// # use std::borrow::Cow;
2021-02-19 23:28:50 -08:00
/// # let section_header = ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
/// # name: Cow::Borrowed("core".into()),
2021-02-19 23:28:50 -08:00
/// # separator: None,
/// # subsection_name: None,
/// # };
/// # let section_data = "[core]\nautocrlf=true\"\"\nfilemode=fa\"lse\"";
/// # assert_eq!(parse_from_str(section_data).unwrap().into_vec(), vec![
/// Event::SectionHeader(section_header),
2021-02-23 16:47:24 -08:00
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::Key(Cow::Borrowed("autocrlf".into())),
2021-02-23 08:30:48 -08:00
/// Event::KeyValueSeparator,
2021-02-23 16:47:24 -08:00
/// Event::Value(Cow::Borrowed(r#"true"""#.into())),
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::Key(Cow::Borrowed("filemode".into())),
2021-02-23 08:30:48 -08:00
/// Event::KeyValueSeparator,
2021-02-23 16:47:24 -08:00
/// Event::Value(Cow::Borrowed(r#"fa"lse""#.into())),
2021-02-19 23:28:50 -08:00
/// # ]);
/// ```
///
/// ## Whitespace after line continuations are part of the value
///
/// Consider the following `git-config` example:
///
/// ```text
/// [some-section]
/// file=a\
/// c
/// ```
///
/// Because how `git-config` treats continuations, the whitespace preceding `c`
/// are in fact part of the value of `file`. The fully interpreted key/value
/// pair is actually `file=a c`. As a result, the parser will provide this
/// split value accordingly:
///
/// ```
2021-02-24 12:47:35 -08:00
/// # use git_config::parser::{Event, ParsedSectionHeader, parse_from_str};
2021-02-21 11:43:41 -08:00
/// # use std::borrow::Cow;
2021-02-19 23:28:50 -08:00
/// # let section_header = ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
/// # name: Cow::Borrowed("some-section".into()),
2021-02-19 23:28:50 -08:00
/// # separator: None,
/// # subsection_name: None,
/// # };
/// # let section_data = "[some-section]\nfile=a\\\n c";
/// # assert_eq!(parse_from_str(section_data).unwrap().into_vec(), vec![
/// Event::SectionHeader(section_header),
2021-02-23 16:47:24 -08:00
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::Key(Cow::Borrowed("file".into())),
2021-02-23 08:30:48 -08:00
/// Event::KeyValueSeparator,
2021-02-23 16:47:24 -08:00
/// Event::ValueNotDone(Cow::Borrowed("a".into())),
/// Event::Newline(Cow::Borrowed("\n".into())),
/// Event::ValueDone(Cow::Borrowed(" c".into())),
2021-02-19 23:28:50 -08:00
/// # ]);
/// ```
///
/// [`GitConfig`]: crate::config::GitConfig
2021-02-18 21:12:59 -08:00
/// [`.ini` file format]: https://en.wikipedia.org/wiki/INI_file
/// [`git`'s documentation]: https://git-scm.com/docs/git-config#_configuration_file
/// [`FromStr`]: std::str::FromStr
#[derive(Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug, Default)]
2021-02-18 09:49:47 -08:00
pub struct Parser<'a> {
2021-02-21 18:16:08 -08:00
frontmatter: Vec<Event<'a>>,
2021-02-18 21:12:59 -08:00
sections: Vec<ParsedSection<'a>>,
2021-02-18 09:49:47 -08:00
}
2021-02-18 21:12:59 -08:00
impl<'a> Parser<'a> {
/// Attempt to zero-copy parse the provided `&str`. On success, returns a
2021-02-24 13:30:04 -08:00
/// [`Parser`] that provides methods to accessing leading comments and
/// sections of a `git-config` file and can be converted into an iterator of
/// [`Event`] for higher level processing.
2021-02-18 21:12:59 -08:00
///
2021-02-19 09:08:42 -08:00
/// This function is identical to [`parse_from_str`].
2021-02-18 21:12:59 -08:00
///
/// # Errors
///
2021-02-19 23:28:50 -08:00
/// Returns an error if the string provided is not a valid `git-config`.
/// This generally is due to either invalid names or if there's extraneous
/// data succeeding valid `git-config` data.
2021-02-18 21:12:59 -08:00
pub fn from_str(s: &'a str) -> Result<Self, ParserError> {
parse_from_str(s)
}
2021-02-24 13:30:04 -08:00
/// Attempt to zero-copy parse the provided bytes. On success, returns a
/// [`Parser`] that provides methods to accessing leading comments and
/// sections of a `git-config` file and can be converted into an iterator of
/// [`Event`] for higher level processing.
///
/// # Errors
///
/// Returns an error if the string provided is not a valid `git-config`.
/// This generally is due to either invalid names or if there's extraneous
/// data succeeding valid `git-config` data.
pub fn from_bytes(s: impl Into<&'a BStr>) -> Result<Self, ParserError<'a>> {
parse_from_bytes(s.into())
}
2021-02-21 18:16:08 -08:00
/// Returns the leading events (any comments, whitespace, or newlines before
/// a section) from the parser. Consider [`Parser::take_frontmatter`] if
/// you need an owned copy only once. If that function was called, then this
/// will always return an empty slice.
pub fn frontmatter(&self) -> &[Event<'a>] {
&self.frontmatter
2021-02-18 21:12:59 -08:00
}
2021-02-21 18:16:08 -08:00
/// Takes the leading events (any comments, whitespace, or newlines before
/// a section) from the parser. Subsequent calls will return an empty vec.
/// Consider [`Parser::frontmatter`] if you only need a reference to the
/// frontmatter
pub fn take_frontmatter(&mut self) -> Vec<Event<'a>> {
2021-02-18 21:12:59 -08:00
let mut to_return = vec![];
2021-02-21 18:16:08 -08:00
std::mem::swap(&mut self.frontmatter, &mut to_return);
2021-02-18 21:12:59 -08:00
to_return
}
2021-02-19 23:28:50 -08:00
/// Returns the parsed sections from the parser. Consider
/// [`Parser::take_sections`] if you need an owned copy only once. If that
/// function was called, then this will always return an empty slice.
2021-02-18 21:12:59 -08:00
pub fn sections(&self) -> &[ParsedSection<'a>] {
&self.sections
}
2021-02-19 23:28:50 -08:00
/// Takes the parsed sections from the parser. Subsequent calls will return
/// an empty vec. Consider [`Parser::sections`] if you only need a reference
/// to the comments.
2021-02-18 21:12:59 -08:00
pub fn take_sections(&mut self) -> Vec<ParsedSection<'a>> {
let mut to_return = vec![];
std::mem::swap(&mut self.sections, &mut to_return);
to_return
}
2021-02-19 23:28:50 -08:00
/// Consumes the parser to produce a Vec of Events.
2021-02-18 21:12:59 -08:00
pub fn into_vec(self) -> Vec<Event<'a>> {
self.into_iter().collect()
}
2021-02-19 23:28:50 -08:00
/// Consumes the parser to produce an iterator of Events.
2021-02-18 21:12:59 -08:00
pub fn into_iter(self) -> impl Iterator<Item = Event<'a>> + FusedIterator {
let section_iter = self
.sections
.into_iter()
.map(|section| {
vec![Event::SectionHeader(section.section_header)]
.into_iter()
2021-02-19 09:08:42 -08:00
.chain(section.events)
2021-02-18 21:12:59 -08:00
})
.flatten();
2021-02-21 18:16:08 -08:00
self.frontmatter.into_iter().chain(section_iter)
2021-02-18 21:12:59 -08:00
}
}
/// Attempt to zero-copy parse the provided `&str`. On success, returns a
/// [`Parser`] that provides methods to accessing leading comments and sections
/// of a `git-config` file and can be converted into an iterator of [`Event`]
/// for higher level processing.
///
/// # Errors
///
2021-02-19 23:28:50 -08:00
/// Returns an error if the string provided is not a valid `git-config`.
/// This generally is due to either invalid names or if there's extraneous
/// data succeeding valid `git-config` data.
2021-02-18 21:12:59 -08:00
pub fn parse_from_str(input: &str) -> Result<Parser<'_>, ParserError> {
2021-02-24 13:30:04 -08:00
parse_from_bytes(input.as_bytes())
}
/// Attempt to zero-copy parse the provided bytes. On success, returns a
/// [`Parser`] that provides methods to accessing leading comments and sections
/// of a `git-config` file and can be converted into an iterator of [`Event`]
/// for higher level processing.
///
/// # Errors
///
/// Returns an error if the string provided is not a valid `git-config`.
/// This generally is due to either invalid names or if there's extraneous
/// data succeeding valid `git-config` data.
pub fn parse_from_bytes(input: &[u8]) -> Result<Parser<'_>, ParserError> {
2021-02-24 20:19:45 -08:00
let mut newlines = 0;
2021-02-21 18:16:08 -08:00
let (i, frontmatter) = many0(alt((
2021-02-24 09:11:18 -08:00
map(comment, Event::Comment),
2021-02-21 18:16:08 -08:00
map(take_spaces, |whitespace| {
2021-02-23 16:47:24 -08:00
Event::Whitespace(Cow::Borrowed(whitespace.into()))
2021-02-21 18:16:08 -08:00
}),
2021-02-24 20:19:45 -08:00
map(take_newline, |(newline, counter)| {
newlines += counter;
2021-02-23 16:47:24 -08:00
Event::Newline(Cow::Borrowed(newline.into()))
}),
2021-02-24 20:19:45 -08:00
)))(input.as_bytes())
// I don't think this can panic. many0 errors if the child parser returns
// a success where the input was not consumed, but alt will only return Ok
// if one of its children succeed. However, all of it's children are
// guaranteed to consume something if they succeed, so the Ok(i) == i case
// can never occur.
.expect("many0(alt(...)) panicked. Likely a bug in one of the children parser.");
if i.is_empty() {
return Ok(Parser {
frontmatter,
sections: vec![],
});
}
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-24 20:19:45 -08:00
2021-02-25 08:11:26 -08:00
let maybe_sections = many1(|i| section(i, &mut node))(i);
let (i, sections) = maybe_sections.map_err(|_| ParserError {
2021-02-24 20:19:45 -08:00
line_number: newlines,
2021-02-26 15:35:49 -08:00
last_attempted_parser: node,
2021-02-24 20:19:45 -08:00
parsed_until: i,
})?;
let sections = sections
.into_iter()
2021-02-26 15:35:49 -08:00
.map(|(section, additional_newlines)| {
newlines += additional_newlines;
section
2021-02-24 20:19:45 -08:00
})
.collect();
2021-02-18 21:12:59 -08:00
2021-02-26 15:35:49 -08:00
// This needs to happen after we collect sections, otherwise the line number
// will be off.
2021-02-18 21:12:59 -08:00
if !i.is_empty() {
2021-02-24 20:19:45 -08:00
return Err(ParserError {
line_number: newlines,
last_attempted_parser: node,
parsed_until: i,
});
2021-02-18 21:12:59 -08:00
}
2021-02-18 09:49:47 -08:00
Ok(Parser {
2021-02-21 18:16:08 -08:00
frontmatter,
2021-02-18 09:49:47 -08:00
sections,
})
}
2021-02-26 15:14:40 -08:00
fn comment(i: &[u8]) -> IResult<&[u8], ParsedComment> {
2021-02-18 09:49:47 -08:00
let (i, comment_tag) = one_of(";#")(i)?;
2021-02-23 16:47:24 -08:00
let (i, comment) = take_till(|c| c == b'\n')(i)?;
2021-02-18 09:49:47 -08:00
Ok((
i,
2021-02-18 21:12:59 -08:00
ParsedComment {
2021-02-18 09:49:47 -08:00
comment_tag,
2021-02-23 16:47:24 -08:00
comment: Cow::Borrowed(comment.into()),
2021-02-18 09:49:47 -08:00
},
))
}
2021-02-24 20:19:45 -08:00
fn section<'a, 'b>(
i: &'a [u8],
2021-02-26 15:10:24 -08:00
node: &'b mut ParserNode,
2021-02-24 20:19:45 -08:00
) -> IResult<&'a [u8], (ParsedSection<'a>, usize)> {
2021-02-26 16:30:04 -08:00
let (mut i, section_header) = section_header(i)?;
2021-02-25 08:11:26 -08:00
2021-02-24 20:19:45 -08:00
let mut newlines = 0;
2021-02-26 16:30:04 -08:00
let mut items = vec![];
// This would usually be a many0(alt(...)), the manual loop allows us to
// optimize vec insertions
loop {
let old_i = i;
if let Ok((new_i, v)) = take_spaces(i) {
if old_i != new_i {
i = new_i;
items.push(Event::Whitespace(Cow::Borrowed(v.into())));
}
}
if let Ok((new_i, (v, new_newlines))) = take_newline(i) {
if old_i != new_i {
i = new_i;
newlines += new_newlines;
items.push(Event::Newline(Cow::Borrowed(v.into())));
}
}
if let Ok((new_i, _)) = section_body(i, node, &mut items) {
if old_i != new_i {
i = new_i;
// items.push(Event::Key(Cow::Borrowed(key.into())));
}
}
if let Ok((new_i, comment)) = comment(i) {
if old_i != new_i {
i = new_i;
items.push(Event::Comment(comment));
}
}
if old_i == i {
break;
}
}
2021-02-25 08:11:26 -08:00
2021-02-18 09:49:47 -08:00
Ok((
i,
2021-02-24 20:19:45 -08:00
(
ParsedSection {
section_header,
2021-02-26 16:30:04 -08:00
events: items,
2021-02-24 20:19:45 -08:00
},
newlines,
),
2021-02-18 09:49:47 -08:00
))
}
2021-02-26 15:14:40 -08:00
fn section_header(i: &[u8]) -> IResult<&[u8], ParsedSectionHeader> {
2021-02-18 09:49:47 -08:00
let (i, _) = char('[')(i)?;
// No spaces must be between section name and section start
2021-02-23 16:47:24 -08:00
let (i, name) = take_while(|c: u8| c.is_ascii_alphanumeric() || c == b'-' || c == b'.')(i)?;
2021-02-18 09:49:47 -08:00
2021-02-23 16:47:24 -08:00
if let Ok((i, _)) = char::<_, NomError<&[u8]>>(']')(i) {
2021-02-18 09:49:47 -08:00
// Either section does not have a subsection or using deprecated
// subsection syntax at this point.
2021-02-23 16:47:24 -08:00
let header = match name.rfind(&[b'.']) {
2021-02-18 21:12:59 -08:00
Some(index) => ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
name: Cow::Borrowed(name[..index].into()),
separator: name
.get(index..index + 1)
.map(|slice| Cow::Borrowed(slice.into())),
subsection_name: name
.get(index + 1..)
.map(|slice| Cow::Borrowed(slice.into())),
2021-02-18 09:49:47 -08:00
},
2021-02-18 21:12:59 -08:00
None => ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
name: Cow::Borrowed(name.into()),
2021-02-19 16:18:02 -08:00
separator: None,
2021-02-18 09:49:47 -08:00
subsection_name: None,
},
};
return Ok((i, header));
}
// Section header must be using modern subsection syntax at this point.
2021-02-19 16:18:02 -08:00
let (i, whitespace) = take_spaces(i)?;
2021-02-18 09:49:47 -08:00
let (i, subsection_name) = delimited(
char('"'),
opt(escaped(none_of("\"\\\n\0"), '\\', one_of(r#""\"#))),
tag("\"]"),
)(i)?;
Ok((
i,
2021-02-18 21:12:59 -08:00
ParsedSectionHeader {
2021-02-23 16:47:24 -08:00
name: Cow::Borrowed(name.into()),
separator: Some(Cow::Borrowed(whitespace.into())),
2021-02-18 09:49:47 -08:00
// We know that there's some section name here, so if we get an
// empty vec here then we actually parsed an empty section name.
2021-02-23 16:47:24 -08:00
subsection_name: subsection_name
.or(Some(b""))
.map(|slice| Cow::Borrowed(slice.into())),
2021-02-18 09:49:47 -08:00
},
))
}
2021-02-26 16:30:04 -08:00
fn section_body<'a, 'b, 'c>(
2021-02-25 08:11:26 -08:00
i: &'a [u8],
2021-02-26 15:10:24 -08:00
node: &'b mut ParserNode,
2021-02-26 16:30:04 -08:00
items: &'c mut Vec<Event<'a>>,
) -> IResult<&'a [u8], ()> {
2021-02-18 09:49:47 -08:00
// maybe need to check for [ here
2021-02-26 15:10:24 -08:00
*node = ParserNode::ConfigName;
2021-02-18 09:49:47 -08:00
let (i, name) = config_name(i)?;
2021-02-25 08:11:26 -08:00
2021-02-26 16:30:04 -08:00
items.push(Event::Key(Cow::Borrowed(name.into())));
2021-02-25 08:11:26 -08:00
2021-02-26 16:30:04 -08:00
let (i, whitespace) = opt(take_spaces)(i)?;
2021-02-25 08:11:26 -08:00
2021-02-19 16:18:02 -08:00
if let Some(whitespace) = whitespace {
2021-02-26 16:30:04 -08:00
items.push(Event::Whitespace(Cow::Borrowed(whitespace.into())));
2021-02-19 16:18:02 -08:00
}
2021-02-26 16:30:04 -08:00
*node = ParserNode::ConfigValue;
let (i, _) = config_value(i, items)?;
Ok((i, ()))
2021-02-18 09:49:47 -08:00
}
/// Parses the config name of a config pair. Assumes the input has already been
/// trimmed of any leading whitespace.
2021-02-26 15:14:40 -08:00
fn config_name(i: &[u8]) -> IResult<&[u8], &[u8]> {
2021-02-18 09:49:47 -08:00
if i.is_empty() {
return Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::NonEmpty,
}));
}
2021-02-23 16:47:24 -08:00
if !(i[0] as char).is_alphabetic() {
2021-02-18 09:49:47 -08:00
return Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Alpha,
}));
}
2021-02-23 16:47:24 -08:00
take_while(|c: u8| (c as char).is_alphanumeric() || c == b'-')(i)
2021-02-18 09:49:47 -08:00
}
2021-02-26 16:30:04 -08:00
fn config_value<'a, 'b>(i: &'a [u8], events: &'b mut Vec<Event<'a>>) -> IResult<&'a [u8], ()> {
2021-02-18 09:49:47 -08:00
if let (i, Some(_)) = opt(char('='))(i)? {
2021-02-23 08:30:48 -08:00
events.push(Event::KeyValueSeparator);
2021-02-19 16:18:02 -08:00
let (i, whitespace) = opt(take_spaces)(i)?;
if let Some(whitespace) = whitespace {
2021-02-23 16:47:24 -08:00
events.push(Event::Whitespace(Cow::Borrowed(whitespace.into())));
2021-02-19 16:18:02 -08:00
}
2021-02-26 16:30:04 -08:00
let (i, _) = value_impl(i, events)?;
Ok((i, ()))
2021-02-18 09:49:47 -08:00
} else {
2021-02-26 16:30:04 -08:00
events.push(Event::Value(Cow::Borrowed("".into())));
Ok((i, ()))
2021-02-18 09:49:47 -08:00
}
}
2021-02-26 16:30:04 -08:00
fn value_impl<'a, 'b>(i: &'a [u8], events: &'b mut Vec<Event<'a>>) -> IResult<&'a [u8], ()> {
2021-02-18 09:49:47 -08:00
let mut parsed_index: usize = 0;
let mut offset: usize = 0;
let mut was_prev_char_escape_char = false;
// This is required to ignore comment markers if they're in a quote.
let mut is_in_quotes = false;
// Used to determine if we return a Value or Value{Not,}Done
let mut partial_value_found = false;
2021-02-23 16:47:24 -08:00
for (index, c) in i.iter().enumerate() {
2021-02-18 09:49:47 -08:00
if was_prev_char_escape_char {
was_prev_char_escape_char = false;
match c {
// We're escaping a newline, which means we've found a
// continuation.
b'\n' => {
partial_value_found = true;
2021-02-23 16:47:24 -08:00
events.push(Event::ValueNotDone(Cow::Borrowed(
i[offset..index - 1].into(),
)));
events.push(Event::Newline(Cow::Borrowed(i[index..index + 1].into())));
2021-02-18 09:49:47 -08:00
offset = index + 1;
parsed_index = 0;
}
b't' | b'\\' | b'n' | b'"' => (),
_ => {
return Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Escaped,
}));
}
}
} else {
match c {
b'\n' => {
parsed_index = index;
break;
}
b';' | b'#' if !is_in_quotes => {
parsed_index = index;
break;
}
b'\\' => was_prev_char_escape_char = true,
b'"' => is_in_quotes = !is_in_quotes,
_ => {}
}
}
}
if parsed_index == 0 {
parsed_index = i.len();
}
// Handle incomplete escape
if was_prev_char_escape_char {
return Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Escaped,
}));
}
// Handle incomplete quotes
if is_in_quotes {
return Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Tag,
}));
}
2021-02-19 16:18:02 -08:00
let (i, remainder_value) = {
let mut new_index = parsed_index;
for index in (offset..parsed_index).rev() {
2021-02-23 16:47:24 -08:00
if !(i[index] as char).is_whitespace() {
2021-02-19 16:18:02 -08:00
new_index = index + 1;
break;
}
}
(&i[new_index..], &i[offset..new_index])
};
2021-02-18 09:49:47 -08:00
if partial_value_found {
2021-02-23 16:47:24 -08:00
events.push(Event::ValueDone(Cow::Borrowed(remainder_value.into())));
2021-02-18 09:49:47 -08:00
} else {
2021-02-23 16:47:24 -08:00
events.push(Event::Value(Cow::Borrowed(remainder_value.into())));
2021-02-18 09:49:47 -08:00
}
2021-02-26 16:30:04 -08:00
Ok((i, ()))
2021-02-18 09:49:47 -08:00
}
2021-02-24 09:11:18 -08:00
fn take_spaces(i: &[u8]) -> IResult<&[u8], &[u8]> {
2021-02-24 20:19:45 -08:00
let (i, v) = take_while(|c| (c as char).is_ascii() && is_space(c))(i)?;
if v.is_empty() {
Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Eof,
}))
} else {
Ok((i, v))
}
2021-02-19 16:18:02 -08:00
}
2021-02-26 15:14:40 -08:00
fn take_newline(i: &[u8]) -> IResult<&[u8], (&[u8], usize)> {
2021-02-24 20:19:45 -08:00
let mut counter = 0;
2021-02-26 15:35:49 -08:00
let (i, v) = take_while(|c| (c as char).is_ascii() && is_newline(c))(i)?;
2021-02-24 20:19:45 -08:00
counter += v.len();
2021-02-19 16:18:02 -08:00
if v.is_empty() {
Err(nom::Err::Error(NomError {
input: i,
code: ErrorKind::Eof,
}))
} else {
2021-02-24 20:19:45 -08:00
Ok((i, (v, counter)))
2021-02-19 16:18:02 -08:00
}
}
2021-02-20 10:49:01 -08:00
2021-02-18 09:49:47 -08:00
#[cfg(test)]
2021-02-20 10:49:01 -08:00
mod comments {
2021-02-18 09:49:47 -08:00
use super::*;
2021-02-24 15:00:15 -08:00
use crate::test_util::{comment as parsed_comment, fully_consumed};
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn semicolon() {
assert_eq!(
2021-02-23 16:47:24 -08:00
comment(b"; this is a semicolon comment").unwrap(),
fully_consumed(parsed_comment(';', " this is a semicolon comment")),
2021-02-20 10:49:01 -08:00
);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn octothorpe() {
assert_eq!(
2021-02-23 16:47:24 -08:00
comment(b"# this is an octothorpe comment").unwrap(),
fully_consumed(parsed_comment('#', " this is an octothorpe comment")),
2021-02-20 10:49:01 -08:00
);
2021-02-19 16:18:02 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn multiple_markers() {
assert_eq!(
2021-02-23 16:47:24 -08:00
comment(b"###### this is an octothorpe comment").unwrap(),
fully_consumed(parsed_comment('#', "##### this is an octothorpe comment")),
2021-02-20 10:49:01 -08:00
);
}
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[cfg(test)]
mod section_headers {
use super::*;
2021-02-24 15:00:15 -08:00
use crate::test_util::{fully_consumed, section_header as parsed_section_header};
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn no_subsection() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(b"[hello]").unwrap(),
fully_consumed(parsed_section_header("hello", None)),
2021-02-20 10:49:01 -08:00
);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn modern_subsection() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello "world"]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (" ", "world"))),
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn escaped_subsection() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello "foo\\bar\""]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (" ", r#"foo\\bar\""#))),
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn deprecated_subsection() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello.world]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (".", "world")))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn empty_legacy_subsection_name() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello.]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (".", "")))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn empty_modern_subsection_name() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello ""]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (" ", "")))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn newline_in_header() {
2021-02-23 16:47:24 -08:00
assert!(section_header(b"[hello\n]").is_err())
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn null_byte_in_header() {
2021-02-23 16:47:24 -08:00
assert!(section_header(b"[hello\0]").is_err())
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn right_brace_in_subsection_name() {
assert_eq!(
2021-02-23 16:47:24 -08:00
section_header(br#"[hello "]"]"#).unwrap(),
fully_consumed(parsed_section_header("hello", (" ", "]")))
2021-02-20 10:49:01 -08:00
);
}
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[cfg(test)]
mod config_name {
use super::*;
2021-02-24 15:00:15 -08:00
use crate::test_util::fully_consumed;
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn just_name() {
2021-02-23 16:47:24 -08:00
assert_eq!(
config_name(b"name").unwrap(),
fully_consumed("name".as_bytes())
);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn must_start_with_alphabetic() {
2021-02-23 16:47:24 -08:00
assert!(config_name(b"4aaa").is_err());
assert!(config_name(b"-aaa").is_err());
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn cannot_be_empty() {
2021-02-23 16:47:24 -08:00
assert!(config_name(b"").is_err())
2021-02-20 10:49:01 -08:00
}
}
2021-02-18 09:49:47 -08:00
2021-02-23 08:30:48 -08:00
#[cfg(test)]
mod section_body {
use super::*;
2021-02-26 16:30:04 -08:00
use crate::test_util::{name_event, value_event, whitespace_event};
2021-02-23 08:30:48 -08:00
#[test]
fn whitespace_is_not_ambigious() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-26 16:30:04 -08:00
let mut vec = vec![];
assert!(section_body(b"a =b", &mut node, &mut vec).is_ok());
2021-02-23 08:30:48 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
vec,
vec![
name_event("a"),
whitespace_event(" "),
Event::KeyValueSeparator,
value_event("b")
]
2021-02-23 08:30:48 -08:00
);
2021-02-26 16:30:04 -08:00
let mut vec = vec![];
assert!(section_body(b"a= b", &mut node, &mut vec).is_ok());
2021-02-23 08:30:48 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
vec,
vec![
name_event("a"),
Event::KeyValueSeparator,
whitespace_event(" "),
value_event("b")
]
2021-02-23 08:30:48 -08:00
);
}
}
2021-02-20 10:49:01 -08:00
#[cfg(test)]
mod value_no_continuation {
use super::*;
2021-02-26 16:30:04 -08:00
use crate::test_util::value_event;
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn no_comment() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
assert_eq!(value_impl(b"hello", &mut events).unwrap().0, b"");
assert_eq!(events, vec![value_event("hello")]);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn no_comment_newline() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
assert_eq!(value_impl(b"hello\na", &mut events).unwrap().0, b"\na");
assert_eq!(events, vec![value_event("hello")]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn semicolon_comment_not_consumed() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"hello;world", &mut events).unwrap().0,
b";world"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event("hello")]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn octothorpe_comment_not_consumed() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"hello#world", &mut events).unwrap().0,
b"#world"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event("hello")]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn values_with_extraneous_whitespace_without_comment() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"hello ", &mut events).unwrap().0,
b" "
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event("hello")]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn values_with_extraneous_whitespace_before_comment() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"hello #world", &mut events)
.unwrap()
.0,
b" #world"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event("hello")]);
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"hello ;world", &mut events)
.unwrap()
.0,
b" ;world"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event("hello")]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn trans_escaped_comment_marker_not_consumed() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(br##"hello"#"world; a"##, &mut events).unwrap().0,
b"; a"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event(r##"hello"#"world"##)]);
2021-02-20 10:49:01 -08:00
}
2021-02-19 16:18:02 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn complex_test() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(br#"value";";ahhhh"#, &mut events).unwrap().0,
b";ahhhh"
2021-02-20 10:49:01 -08:00
);
2021-02-26 16:30:04 -08:00
assert_eq!(events, vec![value_event(r#"value";""#)]);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn garbage_after_continution_is_err() {
2021-02-26 16:30:04 -08:00
assert!(value_impl(b"hello \\afwjdls", &mut vec![]).is_err());
2021-02-20 10:49:01 -08:00
}
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[cfg(test)]
mod value_continuation {
use super::*;
2021-02-26 16:30:04 -08:00
use crate::test_util::{newline_event, value_done_event, value_not_done_event};
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn simple_continuation() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
assert_eq!(value_impl(b"hello\\\nworld", &mut events).unwrap().0, b"");
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
events,
vec![
2021-02-23 16:47:24 -08:00
value_not_done_event("hello"),
newline_event(),
value_done_event("world")
2021-02-26 16:30:04 -08:00
]
2021-02-20 10:49:01 -08:00
);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
#[test]
fn continuation_with_whitespace() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
assert_eq!(
value_impl(b"hello\\\n world", &mut events)
.unwrap()
.0,
b""
);
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
events,
vec![
2021-02-23 16:47:24 -08:00
value_not_done_event("hello"),
newline_event(),
value_done_event(" world")
2021-02-26 16:30:04 -08:00
]
);
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn complex_continuation_with_leftover_comment() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
assert_eq!(
value_impl(b"1 \"\\\"\\\na ; e \"\\\"\\\nd # \"b\t ; c", &mut events)
.unwrap()
.0,
b" # \"b\t ; c"
);
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
events,
vec![
value_not_done_event(r#"1 "\""#),
newline_event(),
value_not_done_event(r#"a ; e "\""#),
newline_event(),
value_done_event("d")
]
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn quote_split_over_two_lines_with_leftover_comment() {
2021-02-26 16:30:04 -08:00
let mut events = vec![];
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-26 16:30:04 -08:00
value_impl(b"\"\\\n;\";a", &mut events).unwrap().0,
";a".as_bytes()
);
assert_eq!(
events,
vec![
value_not_done_event("\""),
newline_event(),
value_done_event(";\"")
]
);
2021-02-18 09:49:47 -08:00
}
2021-02-20 10:49:01 -08:00
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[cfg(test)]
mod section {
use super::*;
2021-02-23 16:47:24 -08:00
use crate::test_util::{
2021-02-24 15:00:15 -08:00
comment_event, fully_consumed, name_event, newline_event,
section_header as parsed_section_header, value_done_event, value_event,
value_not_done_event, whitespace_event,
2021-02-23 16:47:24 -08:00
};
2021-02-20 10:49:01 -08:00
#[test]
fn empty_section() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(b"[test]", &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("test", None),
events: vec![]
},
0
)),
2021-02-20 10:49:01 -08:00
);
}
2021-02-19 16:18:02 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn simple_section() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-23 16:47:24 -08:00
let section_data = br#"[hello]
2021-02-18 09:49:47 -08:00
a = b
c
d = "lol""#;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(section_data, &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("hello", None),
events: vec![
newline_event(),
whitespace_event(" "),
name_event("a"),
whitespace_event(" "),
Event::KeyValueSeparator,
whitespace_event(" "),
value_event("b"),
newline_event(),
whitespace_event(" "),
name_event("c"),
value_event(""),
newline_event(),
whitespace_event(" "),
name_event("d"),
whitespace_event(" "),
Event::KeyValueSeparator,
whitespace_event(" "),
value_event("\"lol\"")
]
},
3
))
2021-02-20 10:49:01 -08:00
)
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn section_single_line() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(b"[hello] c", &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("hello", None),
events: vec![whitespace_event(" "), name_event("c"), value_event("")]
},
0
))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn section_very_commented() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-23 16:47:24 -08:00
let section_data = br#"[hello] ; commentA
2021-02-18 09:49:47 -08:00
a = b # commentB
; commentC
; commentD
c = d"#;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(section_data, &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("hello", None),
events: vec![
whitespace_event(" "),
comment_event(';', " commentA"),
newline_event(),
whitespace_event(" "),
name_event("a"),
whitespace_event(" "),
Event::KeyValueSeparator,
whitespace_event(" "),
value_event("b"),
whitespace_event(" "),
comment_event('#', " commentB"),
newline_event(),
whitespace_event(" "),
comment_event(';', " commentC"),
newline_event(),
whitespace_event(" "),
comment_event(';', " commentD"),
newline_event(),
whitespace_event(" "),
name_event("c"),
whitespace_event(" "),
Event::KeyValueSeparator,
whitespace_event(" "),
value_event("d"),
]
},
4
))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn complex_continuation() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-20 10:49:01 -08:00
// This test is absolute hell. Good luck if this fails.
assert_eq!(
2021-02-24 20:19:45 -08:00
section(
b"[section] a = 1 \"\\\"\\\na ; e \"\\\"\\\nd # \"b\t ; c",
&mut node
)
.unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("section", None),
events: vec![
whitespace_event(" "),
name_event("a"),
whitespace_event(" "),
Event::KeyValueSeparator,
whitespace_event(" "),
value_not_done_event(r#"1 "\""#),
newline_event(),
value_not_done_event(r#"a ; e "\""#),
newline_event(),
value_done_event("d"),
whitespace_event(" "),
comment_event('#', " \"b\t ; c"),
]
},
0
))
2021-02-20 10:49:01 -08:00
);
}
2021-02-18 09:49:47 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn quote_split_over_two_lines() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(b"[section \"a\"] b =\"\\\n;\";a", &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("section", (" ", "a")),
events: vec![
whitespace_event(" "),
name_event("b"),
whitespace_event(" "),
Event::KeyValueSeparator,
value_not_done_event("\""),
newline_event(),
value_done_event(";\""),
comment_event(';', "a"),
]
},
0
))
2021-02-20 10:49:01 -08:00
)
}
2021-02-19 16:18:02 -08:00
2021-02-20 10:49:01 -08:00
#[test]
fn section_handles_extranous_whitespace_before_comment() {
2021-02-26 15:10:24 -08:00
let mut node = ParserNode::SectionHeader;
2021-02-20 10:49:01 -08:00
assert_eq!(
2021-02-24 20:19:45 -08:00
section(b"[s]hello #world", &mut node).unwrap(),
fully_consumed((
ParsedSection {
section_header: parsed_section_header("s", None),
events: vec![
name_event("hello"),
whitespace_event(" "),
value_event(""),
comment_event('#', "world"),
]
},
0
))
2021-02-20 10:49:01 -08:00
);
2021-02-18 09:49:47 -08:00
}
}