fnmain/jotdown
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

inline: store link cowstrs in vec

Browse source 
try to reduce size of Event by placing the cowstr in a shared vec, and
just keeping an index in the event itself

seems to have a significant performance benefit on benchmarks
This commit is contained in:
Noah Hellman 2023-03-23 19:23:16 +01:00
parent a846477cea
commit 2bcc6122ca
2 changed files with 75 additions and 64 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

120
src/inline.rs
View file

@ -24,8 +24,8 @@ pub enum Atom {
Quote { ty: QuoteType, left: bool }, Quote { ty: QuoteType, left: bool },
} }
#[derive(Debug, Clone, PartialEq, Eq)] #[derive(Debug, Copy, Clone, PartialEq, Eq)]
pub enum Container<'s> { pub enum Container {
Span, Span,
Subscript, Subscript,
Superscript, Superscript,
@ -39,14 +39,16 @@ pub enum Container<'s> {
RawFormat, RawFormat,
InlineMath, InlineMath,
DisplayMath, DisplayMath,
ReferenceLink(CowStr<'s>), ReferenceLink(CowStrIndex),
ReferenceImage(CowStr<'s>), ReferenceImage(CowStrIndex),
InlineLink(CowStr<'s>), InlineLink(CowStrIndex),
InlineImage(CowStr<'s>), InlineImage(CowStrIndex),
/// Open delimiter span is URL, closing is '>'. /// Open delimiter span is URL, closing is '>'.
Autolink, Autolink,
} }
type CowStrIndex = u32;
#[derive(Clone, Copy, Debug, PartialEq, Eq)] #[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum QuoteType { pub enum QuoteType {
Single, Single,
@ -54,9 +56,9 @@ pub enum QuoteType {
} }
#[derive(Clone, Debug, PartialEq, Eq)] #[derive(Clone, Debug, PartialEq, Eq)]
pub enum EventKind<'s> { pub enum EventKind {
Enter(Container<'s>), Enter(Container),
Exit(Container<'s>), Exit(Container),
Atom(Atom), Atom(Atom),
Str, Str,
Attributes { container: bool }, Attributes { container: bool },
@ -64,8 +66,8 @@ pub enum EventKind<'s> {
} }
#[derive(Clone, Debug, PartialEq, Eq)] #[derive(Clone, Debug, PartialEq, Eq)]
pub struct Event<'s> { pub struct Event {
pub kind: EventKind<'s>, pub kind: EventKind,
pub span: Span, pub span: Span,
} }
@ -193,9 +195,11 @@ pub struct Parser<'s> {
openers: Vec<(Opener, usize)>, openers: Vec<(Opener, usize)>,
/// Buffer queue for next events. Events are buffered until no modifications due to future /// Buffer queue for next events. Events are buffered until no modifications due to future
/// characters are needed. /// characters are needed.
events: std::collections::VecDeque<Event<'s>>, events: std::collections::VecDeque<Event>,
/// State if inside a verbatim container. /// State if inside a verbatim container.
verbatim: Option<VerbatimState>, verbatim: Option<VerbatimState>,
/// Storage of cow strs, used to reduce size of [`Container`].
pub(crate) store_cowstrs: Vec<CowStr<'s>>,
} }
impl<'s> Parser<'s> { impl<'s> Parser<'s> {
@ -205,6 +209,7 @@ impl<'s> Parser<'s> {
openers: Vec::new(), openers: Vec::new(),
events: std::collections::VecDeque::new(), events: std::collections::VecDeque::new(),
verbatim: None, verbatim: None,
store_cowstrs: Vec::new(),
} }
} }
@ -218,14 +223,15 @@ impl<'s> Parser<'s> {
self.openers.clear(); self.openers.clear();
debug_assert!(self.events.is_empty()); debug_assert!(self.events.is_empty());
debug_assert!(self.verbatim.is_none()); debug_assert!(self.verbatim.is_none());
self.store_cowstrs.clear();
} }
fn push_sp(&mut self, kind: EventKind<'s>, span: Span) -> Option<()> { fn push_sp(&mut self, kind: EventKind, span: Span) -> Option<()> {
self.events.push_back(Event { kind, span }); self.events.push_back(Event { kind, span });
Some(()) Some(())
} }
fn push(&mut self, kind: EventKind<'s>) -> Option<()> { fn push(&mut self, kind: EventKind) -> Option<()> {
self.push_sp(kind, self.input.span) self.push_sp(kind, self.input.span)
} }
@ -275,12 +281,12 @@ impl<'s> Parser<'s> {
} }
self.input.span = span_attr; self.input.span = span_attr;
}; };
let ty_opener = if let EventKind::Enter(ty) = &self.events[event_opener].kind { let ty_opener = if let EventKind::Enter(ty) = self.events[event_opener].kind {
debug_assert!(matches!( debug_assert!(matches!(
ty, ty,
Verbatim | RawFormat | InlineMath | DisplayMath Verbatim | RawFormat | InlineMath | DisplayMath
)); ));
ty.clone() ty
} else { } else {
panic!() panic!()
}; };
@ -505,7 +511,7 @@ impl<'s> Parser<'s> {
self.openers.drain(o..); self.openers.drain(o..);
let mut closed = match DelimEventKind::from(opener) { let mut closed = match DelimEventKind::from(opener) {
DelimEventKind::Container(cont) => { DelimEventKind::Container(cont) => {
self.events[e_opener].kind = EventKind::Enter(cont.clone()); self.events[e_opener].kind = EventKind::Enter(cont);
self.push(EventKind::Exit(cont)) self.push(EventKind::Exit(cont))
} }
DelimEventKind::Quote(ty) => { DelimEventKind::Quote(ty) => {
@ -597,13 +603,15 @@ impl<'s> Parser<'s> {
span_spec.of(self.input.src).into() span_spec.of(self.input.src).into()
}; };
let idx = self.store_cowstrs.len() as CowStrIndex;
self.store_cowstrs.push(spec);
let container = match (image, inline) { let container = match (image, inline) {
(false, false) => ReferenceLink(spec.into()), (false, false) => ReferenceLink(idx),
(false, true) => InlineLink(spec.into()), (false, true) => InlineLink(idx),
(true, false) => ReferenceImage(spec.into()), (true, false) => ReferenceImage(idx),
(true, true) => InlineImage(spec.into()), (true, true) => InlineImage(idx),
}; };
self.events[event_span].kind = EventKind::Enter(container.clone()); self.events[event_span].kind = EventKind::Enter(container);
self.events[e_opener - 1] = Event { self.events[e_opener - 1] = Event {
kind: EventKind::Exit(container), kind: EventKind::Exit(container),
span: Span::new( span: Span::new(
@ -741,7 +749,7 @@ impl<'s> Parser<'s> {
self.push(EventKind::Atom(atom)) self.push(EventKind::Atom(atom))
} }
fn merge_str_events(&mut self, span_str: Span) -> Event<'s> { fn merge_str_events(&mut self, span_str: Span) -> Event {
let mut span = span_str; let mut span = span_str;
let should_merge = |e: &Event, span: Span| { let should_merge = |e: &Event, span: Span| {
matches!(e.kind, EventKind::Str | EventKind::Placeholder) matches!(e.kind, EventKind::Str | EventKind::Placeholder)
@ -765,7 +773,7 @@ impl<'s> Parser<'s> {
} }
} }
fn apply_word_attributes(&mut self, span_str: Span) -> Event<'s> { fn apply_word_attributes(&mut self, span_str: Span) -> Event {
if let Some(i) = span_str if let Some(i) = span_str
.of(self.input.src) .of(self.input.src)
.bytes() .bytes()
@ -902,8 +910,8 @@ impl Opener {
} }
} }
enum DelimEventKind<'s> { enum DelimEventKind {
Container(Container<'s>), Container(Container),
Span(SpanType), Span(SpanType),
Quote(QuoteType), Quote(QuoteType),
Link { Link {
@ -913,7 +921,7 @@ enum DelimEventKind<'s> {
}, },
} }
impl<'s> From<Opener> for DelimEventKind<'s> { impl From<Opener> for DelimEventKind {
fn from(d: Opener) -> Self { fn from(d: Opener) -> Self {
match d { match d {
Opener::Span(ty) => Self::Span(ty), Opener::Span(ty) => Self::Span(ty),
@ -940,7 +948,7 @@ impl<'s> From<Opener> for DelimEventKind<'s> {
} }
impl<'s> Iterator for Parser<'s> { impl<'s> Iterator for Parser<'s> {
type Item = Event<'s>; type Item = Event;
fn next(&mut self) -> Option<Self::Item> { fn next(&mut self) -> Option<Self::Item> {
while self.events.is_empty() while self.events.is_empty()
@ -962,12 +970,12 @@ impl<'s> Iterator for Parser<'s> {
// automatically close unclosed verbatim // automatically close unclosed verbatim
if let Some(VerbatimState { event_opener, .. }) = self.verbatim.take() { if let Some(VerbatimState { event_opener, .. }) = self.verbatim.take() {
let ty_opener = if let EventKind::Enter(ty) = &self.events[event_opener].kind { let ty_opener = if let EventKind::Enter(ty) = self.events[event_opener].kind {
debug_assert!(matches!( debug_assert!(matches!(
ty, ty,
Verbatim | RawFormat | InlineMath | DisplayMath Verbatim | RawFormat | InlineMath | DisplayMath
)); ));
ty.clone() ty
} else { } else {
panic!() panic!()
}; };
@ -1163,31 +1171,31 @@ mod test {
fn span_tag() { fn span_tag() {
test_parse!( test_parse!(
"[text][tag]", "[text][tag]",
(Enter(ReferenceLink("tag".into())), "["), (Enter(ReferenceLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(ReferenceLink("tag".into())), "][tag]"), (Exit(ReferenceLink(0)), "][tag]"),
); );
test_parse!( test_parse!(
"![text][tag]", "![text][tag]",
(Enter(ReferenceImage("tag".into())), "!["), (Enter(ReferenceImage(0)), "!["),
(Str, "text"), (Str, "text"),
(Exit(ReferenceImage("tag".into())), "][tag]"), (Exit(ReferenceImage(0)), "][tag]"),
); );
test_parse!( test_parse!(
"before [text][tag] after", "before [text][tag] after",
(Str, "before "), (Str, "before "),
(Enter(ReferenceLink("tag".into())), "["), (Enter(ReferenceLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(ReferenceLink("tag".into())), "][tag]"), (Exit(ReferenceLink(0)), "][tag]"),
(Str, " after"), (Str, " after"),
); );
test_parse!( test_parse!(
"[[inner][i]][o]", "[[inner][i]][o]",
(Enter(ReferenceLink("o".into())), "["), (Enter(ReferenceLink(1)), "["),
(Enter(ReferenceLink("i".into())), "["), (Enter(ReferenceLink(0)), "["),
(Str, "inner"), (Str, "inner"),
(Exit(ReferenceLink("i".into())), "][i]"), (Exit(ReferenceLink(0)), "][i]"),
(Exit(ReferenceLink("o".into())), "][o]"), (Exit(ReferenceLink(1)), "][o]"),
); );
} }
@ -1195,15 +1203,15 @@ mod test {
fn span_tag_empty() { fn span_tag_empty() {
test_parse!( test_parse!(
"[text][]", "[text][]",
(Enter(ReferenceLink("text".into())), "["), (Enter(ReferenceLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(ReferenceLink("text".into())), "][]"), (Exit(ReferenceLink(0)), "][]"),
); );
test_parse!( test_parse!(
"![text][]", "![text][]",
(Enter(ReferenceImage("text".into())), "!["), (Enter(ReferenceImage(0)), "!["),
(Str, "text"), (Str, "text"),
(Exit(ReferenceImage("text".into())), "][]"), (Exit(ReferenceImage(0)), "][]"),
); );
} }
@ -1212,12 +1220,12 @@ mod test {
// TODO strip non str from tag? // TODO strip non str from tag?
test_parse!( test_parse!(
"[some _text_][]", "[some _text_][]",
(Enter(ReferenceLink("some text".into())), "["), (Enter(ReferenceLink(0)), "["),
(Str, "some "), (Str, "some "),
(Enter(Emphasis), "_"), (Enter(Emphasis), "_"),
(Str, "text"), (Str, "text"),
(Exit(Emphasis), "_"), (Exit(Emphasis), "_"),
(Exit(ReferenceLink("some text".into())), "][]"), (Exit(ReferenceLink(0)), "][]"),
); );
} }
@ -1226,19 +1234,19 @@ mod test {
test_parse!( test_parse!(
"before [text](url) after", "before [text](url) after",
(Str, "before "), (Str, "before "),
(Enter(InlineLink("url".into())), "["), (Enter(InlineLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(InlineLink("url".into())), "](url)"), (Exit(InlineLink(0)), "](url)"),
(Str, " after"), (Str, " after"),
); );
test_parse!( test_parse!(
"[outer [inner](i)](o)", "[outer [inner](i)](o)",
(Enter(InlineLink("o".into())), "["), (Enter(InlineLink(1)), "["),
(Str, "outer "), (Str, "outer "),
(Enter(InlineLink("i".into())), "["), (Enter(InlineLink(0)), "["),
(Str, "inner"), (Str, "inner"),
(Exit(InlineLink("i".into())), "](i)"), (Exit(InlineLink(0)), "](i)"),
(Exit(InlineLink("o".into())), "](o)"), (Exit(InlineLink(1)), "](o)"),
); );
} }
@ -1257,9 +1265,9 @@ mod test {
fn span_url_attr_closed() { fn span_url_attr_closed() {
test_parse!( test_parse!(
"[text]({.cls})", "[text]({.cls})",
(Enter(InlineLink("{.cls}".into())), "["), (Enter(InlineLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(InlineLink("{.cls}".into())), "]({.cls})"), (Exit(InlineLink(0)), "]({.cls})"),
); );
} }
@ -1268,9 +1276,9 @@ mod test {
test_parse!( test_parse!(
"before [text]() after", "before [text]() after",
(Str, "before "), (Str, "before "),
(Enter(InlineLink("".into())), "["), (Enter(InlineLink(0)), "["),
(Str, "text"), (Str, "text"),
(Exit(InlineLink("".into())), "]()"), (Exit(InlineLink(0)), "]()"),
(Str, " after"), (Str, " after"),
); );
} }

19
src/lib.rs
View file

@ -830,14 +830,17 @@ impl<'s> Parser<'s> {
inline::Container::Emphasis => Container::Emphasis, inline::Container::Emphasis => Container::Emphasis,
inline::Container::Strong => Container::Strong, inline::Container::Strong => Container::Strong,
inline::Container::Mark => Container::Mark, inline::Container::Mark => Container::Mark,
inline::Container::InlineLink(url) => { inline::Container::InlineLink(url) => Container::Link(
Container::Link(url, LinkType::Span(SpanLinkType::Inline)) self.inline_parser.store_cowstrs[url as usize].clone(),
} LinkType::Span(SpanLinkType::Inline),
inline::Container::InlineImage(url) => { ),
Container::Image(url, SpanLinkType::Inline) inline::Container::InlineImage(url) => Container::Image(
} self.inline_parser.store_cowstrs[url as usize].clone(),
inline::Container::ReferenceLink(ref tag) SpanLinkType::Inline,
| inline::Container::ReferenceImage(ref tag) => { ),
inline::Container::ReferenceLink(tag)
| inline::Container::ReferenceImage(tag) => {
let tag = &self.inline_parser.store_cowstrs[tag as usize];
let link_def = self let link_def = self
.pre_pass .pre_pass
.link_definitions .link_definitions