compiler/rustc_ast_pretty/src/pp.rs - rust-lang/rust - Git at Google

 //! This pretty-printer is a direct reimplementation of Philip Karlton's
 //! Mesa pretty-printer, as described in the appendix to
 //! Derek C. Oppen, "Pretty Printing" (1979),
 //! Stanford Computer Science Department STAN-CS-79-770,
 //! <http://i.stanford.edu/pub/cstr/reports/cs/tr/79/770/CS-TR-79-770.pdf>.
 //!
 //! The algorithm's aim is to break a stream into as few lines as possible
 //! while respecting the indentation-consistency requirements of the enclosing
 //! block, and avoiding breaking at silly places on block boundaries, for
 //! example, between "x" and ")" in "x)".
 //!
 //! I am implementing this algorithm because it comes with 20 pages of
 //! documentation explaining its theory, and because it addresses the set of
 //! concerns I've seen other pretty-printers fall down on. Weirdly. Even though
 //! it's 32 years old. What can I say?
 //!
 //! Despite some redundancies and quirks in the way it's implemented in that
 //! paper, I've opted to keep the implementation here as similar as I can,
 //! changing only what was blatantly wrong, a typo, or sufficiently
 //! non-idiomatic rust that it really stuck out.
 //!
 //! In particular you'll see a certain amount of churn related to INTEGER vs.
 //! CARDINAL in the Mesa implementation. Mesa apparently interconverts the two
 //! somewhat readily? In any case, I've used usize for indices-in-buffers and
 //! ints for character-sizes-and-indentation-offsets. This respects the need
 //! for ints to "go negative" while carrying a pending-calculation balance, and
 //! helps differentiate all the numbers flying around internally (slightly).
 //!
 //! I also inverted the indentation arithmetic used in the print stack, since
 //! the Mesa implementation (somewhat randomly) stores the offset on the print
 //! stack in terms of margin-col rather than col itself. I store col.
 //!
 //! I also implemented a small change in the String token, in that I store an
 //! explicit length for the string. For most tokens this is just the length of
 //! the accompanying string. But it's necessary to permit it to differ, for
 //! encoding things that are supposed to "go on their own line" -- certain
 //! classes of comment and blank-line -- where relying on adjacent
 //! hardbreak-like Break tokens with long blankness indication doesn't actually
 //! work. To see why, consider when there is a "thing that should be on its own
 //! line" between two long blocks, say functions. If you put a hardbreak after
 //! each function (or before each) and the breaking algorithm decides to break
 //! there anyways (because the functions themselves are long) you wind up with
 //! extra blank lines. If you don't put hardbreaks you can wind up with the
 //! "thing which should be on its own line" not getting its own line in the
 //! rare case of "really small functions" or such. This re-occurs with comments
 //! and explicit blank lines. So in those cases we use a string with a payload
 //! we want isolated to a line and an explicit length that's huge, surrounded
 //! by two zero-length breaks. The algorithm will try its best to fit it on a
 //! line (which it can't) and so naturally place the content on its own line to
 //! avoid combining it with other lines and making matters even worse.
 //!
 //! # Explanation
 //!
 //! In case you do not have the paper, here is an explanation of what's going
 //! on.
 //!
 //! There is a stream of input tokens flowing through this printer.
 //!
 //! The printer buffers up to 3N tokens inside itself, where N is linewidth.
 //! Yes, linewidth is chars and tokens are multi-char, but in the worst
 //! case every token worth buffering is 1 char long, so it's ok.
 //!
 //! Tokens are String, Break, and Begin/End to delimit blocks.
 //!
 //! Begin tokens can carry an offset, saying "how far to indent when you break
 //! inside here", as well as a flag indicating "consistent" or "inconsistent"
 //! breaking. Consistent breaking means that after the first break, no attempt
 //! will be made to flow subsequent breaks together onto lines. Inconsistent
 //! is the opposite. Inconsistent breaking example would be, say:
 //!
 //! ```ignore (illustrative)
 //! foo(hello, there, good, friends)
 //! ```
 //!
 //! breaking inconsistently to become
 //!
 //! ```ignore (illustrative)
 //! foo(hello, there,
 //!     good, friends);
 //! ```
 //!
 //! whereas a consistent breaking would yield:
 //!
 //! ```ignore (illustrative)
 //! foo(hello,
 //!     there,
 //!     good,
 //!     friends);
 //! ```
 //!
 //! That is, in the consistent-break blocks we value vertical alignment
 //! more than the ability to cram stuff onto a line. But in all cases if it
 //! can make a block a one-liner, it'll do so.
 //!
 //! Carrying on with high-level logic:
 //!
 //! The buffered tokens go through a ring-buffer, 'tokens'. The 'left' and
 //! 'right' indices denote the active portion of the ring buffer as well as
 //! describing hypothetical points-in-the-infinite-stream at most 3N tokens
 //! apart (i.e., "not wrapped to ring-buffer boundaries"). The paper will switch
 //! between using 'left' and 'right' terms to denote the wrapped-to-ring-buffer
 //! and point-in-infinite-stream senses freely.
 //!
 //! There is a parallel ring buffer, `size`, that holds the calculated size of
 //! each token. Why calculated? Because for Begin/End pairs, the "size"
 //! includes everything between the pair. That is, the "size" of Begin is
 //! actually the sum of the sizes of everything between Begin and the paired
 //! End that follows. Since that is arbitrarily far in the future, `size` is
 //! being rewritten regularly while the printer runs; in fact most of the
 //! machinery is here to work out `size` entries on the fly (and give up when
 //! they're so obviously over-long that "infinity" is a good enough
 //! approximation for purposes of line breaking).
 //!
 //! The "input side" of the printer is managed as an abstract process called
 //! SCAN, which uses `scan_stack`, to manage calculating `size`. SCAN is, in
 //! other words, the process of calculating 'size' entries.
 //!
 //! The "output side" of the printer is managed by an abstract process called
 //! PRINT, which uses `print_stack`, `margin` and `space` to figure out what to
 //! do with each token/size pair it consumes as it goes. It's trying to consume
 //! the entire buffered window, but can't output anything until the size is >=
 //! 0 (sizes are set to negative while they're pending calculation).
 //!
 //! So SCAN takes input and buffers tokens and pending calculations, while
 //! PRINT gobbles up completed calculations and tokens from the buffer. The
 //! theory is that the two can never get more than 3N tokens apart, because
 //! once there's "obviously" too much data to fit on a line, in a size
 //! calculation, SCAN will write "infinity" to the size and let PRINT consume
 //! it.
 //!
 //! In this implementation (following the paper, again) the SCAN process is the
 //! methods called `Printer::scan_*`, and the 'PRINT' process is the
 //! method called `Printer::print`.

 mod convenience;
 mod ring;

 use std::borrow::Cow;
 use std::collections::VecDeque;
 use std::{cmp, iter};

 use ring::RingBuffer;

 /// How to break. Described in more detail in the module docs.
 #[derive(Clone, Copy, PartialEq)]
 pub enum Breaks {
     Consistent,
     Inconsistent,
 }

 #[derive(Clone, Copy, PartialEq)]
 enum IndentStyle {
     /// Vertically aligned under whatever column this block begins at.
     ///
     ///     fn demo(arg1: usize,
     ///             arg2: usize) {}
     Visual,
     /// Indented relative to the indentation level of the previous line.
     ///
     ///     fn demo(
     ///         arg1: usize,
     ///         arg2: usize,
     ///     ) {}
     Block { offset: isize },
 }

 #[derive(Clone, Copy, Default, PartialEq)]
 pub(crate) struct BreakToken {
     offset: isize,
     blank_space: isize,
     pre_break: Option<char>,
 }

 #[derive(Clone, Copy, PartialEq)]
 pub(crate) struct BeginToken {
     indent: IndentStyle,
     breaks: Breaks,
 }

 #[derive(PartialEq)]
 pub(crate) enum Token {
     // In practice a string token contains either a `&'static str` or a
     // `String`. `Cow` is overkill for this because we never modify the data,
     // but it's more convenient than rolling our own more specialized type.
     String(Cow<'static, str>),
     Break(BreakToken),
     Begin(BeginToken),
     End,
 }

 #[derive(Copy, Clone)]
 enum PrintFrame {
     Fits,
     Broken { indent: usize, breaks: Breaks },
 }

 const SIZE_INFINITY: isize = 0xffff;

 /// Target line width.
 const MARGIN: isize = 78;
 /// Every line is allowed at least this much space, even if highly indented.
 const MIN_SPACE: isize = 60;

 pub struct Printer {
     out: String,
     /// Number of spaces left on line
     space: isize,
     /// Ring-buffer of tokens and calculated sizes
     buf: RingBuffer<BufEntry>,
     /// Running size of stream "...left"
     left_total: isize,
     /// Running size of stream "...right"
     right_total: isize,
     /// Pseudo-stack, really a ring too. Holds the
     /// primary-ring-buffers index of the Begin that started the
     /// current block, possibly with the most recent Break after that
     /// Begin (if there is any) on top of it. Stuff is flushed off the
     /// bottom as it becomes irrelevant due to the primary ring-buffer
     /// advancing.
     scan_stack: VecDeque<usize>,
     /// Stack of blocks-in-progress being flushed by print
     print_stack: Vec<PrintFrame>,
     /// Level of indentation of current line
     indent: usize,
     /// Buffered indentation to avoid writing trailing whitespace
     pending_indentation: isize,
     /// The token most recently popped from the left boundary of the
     /// ring-buffer for printing
     last_printed: Option<Token>,
 }

 struct BufEntry {
     token: Token,
     size: isize,
 }

 // Boxes opened with methods like `Printer::{cbox,ibox}` must be closed with
 // `Printer::end`. Failure to do so can result in bad indenting, or in extreme
 // cases, cause no output to be produced at all.
 //
 // Box opening and closing used to be entirely implicit, which was hard to
 // understand and easy to get wrong. This marker type is now returned from the
 // box opening methods and forgotten by `Printer::end`. Any marker that isn't
 // forgotten will trigger a panic in `drop`. (Closing a box more than once
 // isn't possible because `BoxMarker` doesn't implement `Copy` or `Clone`.)
 //
 // Note: it would be better to make open/close mismatching impossible and avoid
 // the need for this marker type altogether by having functions like
 // `with_ibox` that open a box, call a closure, and then close the box. That
 // would work for simple cases, but box lifetimes sometimes interact with
 // complex control flow and across function boundaries in ways that are
 // difficult to handle with such a technique.
 #[must_use]
 pub struct BoxMarker;

 impl !Clone for BoxMarker {}
 impl !Copy for BoxMarker {}

 impl Drop for BoxMarker {
     fn drop(&mut self) {
         panic!("BoxMarker not ended with `Printer::end()`");
     }
 }

 impl Printer {
     pub fn new() -> Self {
         Printer {
             out: String::new(),
             space: MARGIN,
             buf: RingBuffer::new(),
             left_total: 0,
             right_total: 0,
             scan_stack: VecDeque::new(),
             print_stack: Vec::new(),
             indent: 0,
             pending_indentation: 0,
             last_printed: None,
         }
     }

     pub(crate) fn last_token(&self) -> Option<&Token> {
         self.last_token_still_buffered().or_else(|| self.last_printed.as_ref())
     }

     pub(crate) fn last_token_still_buffered(&self) -> Option<&Token> {
         self.buf.last().map(|last| &last.token)
     }

     /// Be very careful with this!
     pub(crate) fn replace_last_token_still_buffered(&mut self, token: Token) {
         self.buf.last_mut().unwrap().token = token;
     }

     fn scan_eof(&mut self) {
         if !self.scan_stack.is_empty() {
             self.check_stack(0);
             self.advance_left();
         }
     }

     // This is is where `BoxMarker`s are produced.
     fn scan_begin(&mut self, token: BeginToken) -> BoxMarker {
         if self.scan_stack.is_empty() {
             self.left_total = 1;
             self.right_total = 1;
             self.buf.clear();
         }
         let right = self.buf.push(BufEntry { token: Token::Begin(token), size: -self.right_total });
         self.scan_stack.push_back(right);
         BoxMarker
     }

     // This is is where `BoxMarker`s are consumed.
     fn scan_end(&mut self, b: BoxMarker) {
         if self.scan_stack.is_empty() {
             self.print_end();
         } else {
             let right = self.buf.push(BufEntry { token: Token::End, size: -1 });
             self.scan_stack.push_back(right);
         }
         std::mem::forget(b)
     }

     fn scan_break(&mut self, token: BreakToken) {
         if self.scan_stack.is_empty() {
             self.left_total = 1;
             self.right_total = 1;
             self.buf.clear();
         } else {
             self.check_stack(0);
         }
         let right = self.buf.push(BufEntry { token: Token::Break(token), size: -self.right_total });
         self.scan_stack.push_back(right);
         self.right_total += token.blank_space;
     }

     fn scan_string(&mut self, string: Cow<'static, str>) {
         if self.scan_stack.is_empty() {
             self.print_string(&string);
         } else {
             let len = string.len() as isize;
             self.buf.push(BufEntry { token: Token::String(string), size: len });
             self.right_total += len;
             self.check_stream();
         }
     }

     pub(crate) fn offset(&mut self, offset: isize) {
         if let Some(BufEntry { token: Token::Break(token), .. }) = &mut self.buf.last_mut() {
             token.offset += offset;
         }
     }

     fn check_stream(&mut self) {
         while self.right_total - self.left_total > self.space {
             if *self.scan_stack.front().unwrap() == self.buf.index_of_first() {
                 self.scan_stack.pop_front().unwrap();
                 self.buf.first_mut().unwrap().size = SIZE_INFINITY;
             }
             self.advance_left();
             if self.buf.is_empty() {
                 break;
             }
         }
     }

     fn advance_left(&mut self) {
         while self.buf.first().unwrap().size >= 0 {
             let left = self.buf.pop_first().unwrap();

             match &left.token {
                 Token::String(string) => {
                     self.left_total += string.len() as isize;
                     self.print_string(string);
                 }
                 Token::Break(token) => {
                     self.left_total += token.blank_space;
                     self.print_break(*token, left.size);
                 }
                 Token::Begin(token) => self.print_begin(*token, left.size),
                 Token::End => self.print_end(),
             }

             self.last_printed = Some(left.token);

             if self.buf.is_empty() {
                 break;
             }
         }
     }

     fn check_stack(&mut self, mut depth: usize) {
         while let Some(&index) = self.scan_stack.back() {
             let entry = &mut self.buf[index];
             match entry.token {
                 Token::Begin(_) => {
                     if depth == 0 {
                         break;
                     }
                     self.scan_stack.pop_back().unwrap();
                     entry.size += self.right_total;
                     depth -= 1;
                 }
                 Token::End => {
                     // paper says + not =, but that makes no sense.
                     self.scan_stack.pop_back().unwrap();
                     entry.size = 1;
                     depth += 1;
                 }
                 _ => {
                     self.scan_stack.pop_back().unwrap();
                     entry.size += self.right_total;
                     if depth == 0 {
                         break;
                     }
                 }
             }
         }
     }

     fn get_top(&self) -> PrintFrame {
         *self
             .print_stack
             .last()
             .unwrap_or(&PrintFrame::Broken { indent: 0, breaks: Breaks::Inconsistent })
     }

     fn print_begin(&mut self, token: BeginToken, size: isize) {
         if size > self.space {
             self.print_stack.push(PrintFrame::Broken { indent: self.indent, breaks: token.breaks });
             self.indent = match token.indent {
                 IndentStyle::Block { offset } => {
                     usize::try_from(self.indent as isize + offset).unwrap()
                 }
                 IndentStyle::Visual => (MARGIN - self.space) as usize,
             };
         } else {
             self.print_stack.push(PrintFrame::Fits);
         }
     }

     fn print_end(&mut self) {
         if let PrintFrame::Broken { indent, .. } = self.print_stack.pop().unwrap() {
             self.indent = indent;
         }
     }

     fn print_break(&mut self, token: BreakToken, size: isize) {
         let fits = match self.get_top() {
             PrintFrame::Fits => true,
             PrintFrame::Broken { breaks: Breaks::Consistent, .. } => false,
             PrintFrame::Broken { breaks: Breaks::Inconsistent, .. } => size <= self.space,
         };
         if fits {
             self.pending_indentation += token.blank_space;
             self.space -= token.blank_space;
         } else {
             if let Some(pre_break) = token.pre_break {
                 self.out.push(pre_break);
             }
             self.out.push('\n');
             let indent = self.indent as isize + token.offset;
             self.pending_indentation = indent;
             self.space = cmp::max(MARGIN - indent, MIN_SPACE);
         }
     }

     fn print_string(&mut self, string: &str) {
         // Write the pending indent. A more concise way of doing this would be:
         //
         //   write!(self.out, "{: >n$}", "", n = self.pending_indentation as usize)?;
         //
         // But that is significantly slower. This code is sufficiently hot, and indents can get
         // sufficiently large, that the difference is significant on some workloads.
         self.out.reserve(self.pending_indentation as usize);
         self.out.extend(iter::repeat(' ').take(self.pending_indentation as usize));
         self.pending_indentation = 0;

         self.out.push_str(string);
         self.space -= string.len() as isize;
     }
 }
	//! This pretty-printer is a direct reimplementation of Philip Karlton's
	//! Mesa pretty-printer, as described in the appendix to
	//! Derek C. Oppen, "Pretty Printing" (1979),
	//! Stanford Computer Science Department STAN-CS-79-770,
	//! <http://i.stanford.edu/pub/cstr/reports/cs/tr/79/770/CS-TR-79-770.pdf>.
	//!
	//! The algorithm's aim is to break a stream into as few lines as possible
	//! while respecting the indentation-consistency requirements of the enclosing
	//! block, and avoiding breaking at silly places on block boundaries, for
	//! example, between "x" and ")" in "x)".
	//!
	//! I am implementing this algorithm because it comes with 20 pages of
	//! documentation explaining its theory, and because it addresses the set of
	//! concerns I've seen other pretty-printers fall down on. Weirdly. Even though
	//! it's 32 years old. What can I say?
	//!
	//! Despite some redundancies and quirks in the way it's implemented in that
	//! paper, I've opted to keep the implementation here as similar as I can,
	//! changing only what was blatantly wrong, a typo, or sufficiently
	//! non-idiomatic rust that it really stuck out.
	//!
	//! In particular you'll see a certain amount of churn related to INTEGER vs.
	//! CARDINAL in the Mesa implementation. Mesa apparently interconverts the two
	//! somewhat readily? In any case, I've used usize for indices-in-buffers and
	//! ints for character-sizes-and-indentation-offsets. This respects the need
	//! for ints to "go negative" while carrying a pending-calculation balance, and
	//! helps differentiate all the numbers flying around internally (slightly).
	//!
	//! I also inverted the indentation arithmetic used in the print stack, since
	//! the Mesa implementation (somewhat randomly) stores the offset on the print
	//! stack in terms of margin-col rather than col itself. I store col.
	//!
	//! I also implemented a small change in the String token, in that I store an
	//! explicit length for the string. For most tokens this is just the length of
	//! the accompanying string. But it's necessary to permit it to differ, for
	//! encoding things that are supposed to "go on their own line" -- certain
	//! classes of comment and blank-line -- where relying on adjacent
	//! hardbreak-like Break tokens with long blankness indication doesn't actually
	//! work. To see why, consider when there is a "thing that should be on its own
	//! line" between two long blocks, say functions. If you put a hardbreak after
	//! each function (or before each) and the breaking algorithm decides to break
	//! there anyways (because the functions themselves are long) you wind up with
	//! extra blank lines. If you don't put hardbreaks you can wind up with the
	//! "thing which should be on its own line" not getting its own line in the
	//! rare case of "really small functions" or such. This re-occurs with comments
	//! and explicit blank lines. So in those cases we use a string with a payload
	//! we want isolated to a line and an explicit length that's huge, surrounded
	//! by two zero-length breaks. The algorithm will try its best to fit it on a
	//! line (which it can't) and so naturally place the content on its own line to
	//! avoid combining it with other lines and making matters even worse.
	//!
	//! # Explanation
	//!
	//! In case you do not have the paper, here is an explanation of what's going
	//! on.
	//!
	//! There is a stream of input tokens flowing through this printer.
	//!
	//! The printer buffers up to 3N tokens inside itself, where N is linewidth.
	//! Yes, linewidth is chars and tokens are multi-char, but in the worst
	//! case every token worth buffering is 1 char long, so it's ok.
	//!
	//! Tokens are String, Break, and Begin/End to delimit blocks.
	//!
	//! Begin tokens can carry an offset, saying "how far to indent when you break
	//! inside here", as well as a flag indicating "consistent" or "inconsistent"
	//! breaking. Consistent breaking means that after the first break, no attempt
	//! will be made to flow subsequent breaks together onto lines. Inconsistent
	//! is the opposite. Inconsistent breaking example would be, say:
	//!
	//! ```ignore (illustrative)
	//! foo(hello, there, good, friends)
	//! ```
	//!
	//! breaking inconsistently to become
	//!
	//! ```ignore (illustrative)
	//! foo(hello, there,
	//! good, friends);
	//! ```
	//!
	//! whereas a consistent breaking would yield:
	//!
	//! ```ignore (illustrative)
	//! foo(hello,
	//! there,
	//! good,
	//! friends);
	//! ```
	//!
	//! That is, in the consistent-break blocks we value vertical alignment
	//! more than the ability to cram stuff onto a line. But in all cases if it
	//! can make a block a one-liner, it'll do so.
	//!
	//! Carrying on with high-level logic:
	//!
	//! The buffered tokens go through a ring-buffer, 'tokens'. The 'left' and
	//! 'right' indices denote the active portion of the ring buffer as well as
	//! describing hypothetical points-in-the-infinite-stream at most 3N tokens
	//! apart (i.e., "not wrapped to ring-buffer boundaries"). The paper will switch
	//! between using 'left' and 'right' terms to denote the wrapped-to-ring-buffer
	//! and point-in-infinite-stream senses freely.
	//!
	//! There is a parallel ring buffer, `size`, that holds the calculated size of
	//! each token. Why calculated? Because for Begin/End pairs, the "size"
	//! includes everything between the pair. That is, the "size" of Begin is
	//! actually the sum of the sizes of everything between Begin and the paired
	//! End that follows. Since that is arbitrarily far in the future, `size` is
	//! being rewritten regularly while the printer runs; in fact most of the
	//! machinery is here to work out `size` entries on the fly (and give up when
	//! they're so obviously over-long that "infinity" is a good enough
	//! approximation for purposes of line breaking).
	//!
	//! The "input side" of the printer is managed as an abstract process called
	//! SCAN, which uses `scan_stack`, to manage calculating `size`. SCAN is, in
	//! other words, the process of calculating 'size' entries.
	//!
	//! The "output side" of the printer is managed by an abstract process called
	//! PRINT, which uses `print_stack`, `margin` and `space` to figure out what to
	//! do with each token/size pair it consumes as it goes. It's trying to consume
	//! the entire buffered window, but can't output anything until the size is >=
	//! 0 (sizes are set to negative while they're pending calculation).
	//!
	//! So SCAN takes input and buffers tokens and pending calculations, while
	//! PRINT gobbles up completed calculations and tokens from the buffer. The
	//! theory is that the two can never get more than 3N tokens apart, because
	//! once there's "obviously" too much data to fit on a line, in a size
	//! calculation, SCAN will write "infinity" to the size and let PRINT consume
	//! it.
	//!
	//! In this implementation (following the paper, again) the SCAN process is the
	//! methods called `Printer::scan_*`, and the 'PRINT' process is the
	//! method called `Printer::print`.

	mod convenience;
	mod ring;

	use std::borrow::Cow;
	use std::collections::VecDeque;
	use std::{cmp, iter};

	use ring::RingBuffer;

	/// How to break. Described in more detail in the module docs.
	#[derive(Clone, Copy, PartialEq)]
	pub enum Breaks {
	Consistent,
	Inconsistent,
	}

	#[derive(Clone, Copy, PartialEq)]
	enum IndentStyle {
	/// Vertically aligned under whatever column this block begins at.
	///
	/// fn demo(arg1: usize,
	/// arg2: usize) {}
	Visual,
	/// Indented relative to the indentation level of the previous line.
	///
	/// fn demo(
	/// arg1: usize,
	/// arg2: usize,
	/// ) {}
	Block { offset: isize },
	}

	#[derive(Clone, Copy, Default, PartialEq)]
	pub(crate) struct BreakToken {
	offset: isize,
	blank_space: isize,
	pre_break: Option<char>,
	}

	#[derive(Clone, Copy, PartialEq)]
	pub(crate) struct BeginToken {
	indent: IndentStyle,
	breaks: Breaks,
	}

	#[derive(PartialEq)]
	pub(crate) enum Token {
	// In practice a string token contains either a `&'static str` or a
	// `String`. `Cow` is overkill for this because we never modify the data,
	// but it's more convenient than rolling our own more specialized type.
	String(Cow<'static, str>),
	Break(BreakToken),
	Begin(BeginToken),
	End,
	}

	#[derive(Copy, Clone)]
	enum PrintFrame {
	Fits,
	Broken { indent: usize, breaks: Breaks },
	}

	const SIZE_INFINITY: isize = 0xffff;

	/// Target line width.
	const MARGIN: isize = 78;
	/// Every line is allowed at least this much space, even if highly indented.
	const MIN_SPACE: isize = 60;

	pub struct Printer {
	out: String,
	/// Number of spaces left on line
	space: isize,
	/// Ring-buffer of tokens and calculated sizes
	buf: RingBuffer<BufEntry>,
	/// Running size of stream "...left"
	left_total: isize,
	/// Running size of stream "...right"
	right_total: isize,
	/// Pseudo-stack, really a ring too. Holds the
	/// primary-ring-buffers index of the Begin that started the
	/// current block, possibly with the most recent Break after that
	/// Begin (if there is any) on top of it. Stuff is flushed off the
	/// bottom as it becomes irrelevant due to the primary ring-buffer
	/// advancing.
	scan_stack: VecDeque<usize>,
	/// Stack of blocks-in-progress being flushed by print
	print_stack: Vec<PrintFrame>,
	/// Level of indentation of current line
	indent: usize,
	/// Buffered indentation to avoid writing trailing whitespace
	pending_indentation: isize,
	/// The token most recently popped from the left boundary of the
	/// ring-buffer for printing
	last_printed: Option<Token>,
	}

	struct BufEntry {
	token: Token,
	size: isize,
	}

	// Boxes opened with methods like `Printer::{cbox,ibox}` must be closed with
	// `Printer::end`. Failure to do so can result in bad indenting, or in extreme
	// cases, cause no output to be produced at all.
	//
	// Box opening and closing used to be entirely implicit, which was hard to
	// understand and easy to get wrong. This marker type is now returned from the
	// box opening methods and forgotten by `Printer::end`. Any marker that isn't
	// forgotten will trigger a panic in `drop`. (Closing a box more than once
	// isn't possible because `BoxMarker` doesn't implement `Copy` or `Clone`.)
	//
	// Note: it would be better to make open/close mismatching impossible and avoid
	// the need for this marker type altogether by having functions like
	// `with_ibox` that open a box, call a closure, and then close the box. That
	// would work for simple cases, but box lifetimes sometimes interact with
	// complex control flow and across function boundaries in ways that are
	// difficult to handle with such a technique.
	#[must_use]
	pub struct BoxMarker;

	impl !Clone for BoxMarker {}
	impl !Copy for BoxMarker {}

	impl Drop for BoxMarker {
	fn drop(&mut self) {
	panic!("BoxMarker not ended with `Printer::end()`");
	}
	}

	impl Printer {
	pub fn new() -> Self {
	Printer {
	out: String::new(),
	space: MARGIN,
	buf: RingBuffer::new(),
	left_total: 0,
	right_total: 0,
	scan_stack: VecDeque::new(),
	print_stack: Vec::new(),
	indent: 0,
	pending_indentation: 0,
	last_printed: None,
	}
	}

	pub(crate) fn last_token(&self) -> Option<&Token> {
	self.last_token_still_buffered().or_else(\|\| self.last_printed.as_ref())
	}

	pub(crate) fn last_token_still_buffered(&self) -> Option<&Token> {
	self.buf.last().map(\|last\| &last.token)
	}

	/// Be very careful with this!
	pub(crate) fn replace_last_token_still_buffered(&mut self, token: Token) {
	self.buf.last_mut().unwrap().token = token;
	}

	fn scan_eof(&mut self) {
	if !self.scan_stack.is_empty() {
	self.check_stack(0);
	self.advance_left();
	}
	}

	// This is is where `BoxMarker`s are produced.
	fn scan_begin(&mut self, token: BeginToken) -> BoxMarker {
	if self.scan_stack.is_empty() {
	self.left_total = 1;
	self.right_total = 1;
	self.buf.clear();
	}
	let right = self.buf.push(BufEntry { token: Token::Begin(token), size: -self.right_total });
	self.scan_stack.push_back(right);
	BoxMarker
	}

	// This is is where `BoxMarker`s are consumed.
	fn scan_end(&mut self, b: BoxMarker) {
	if self.scan_stack.is_empty() {
	self.print_end();
	} else {
	let right = self.buf.push(BufEntry { token: Token::End, size: -1 });
	self.scan_stack.push_back(right);
	}
	std::mem::forget(b)
	}

	fn scan_break(&mut self, token: BreakToken) {
	if self.scan_stack.is_empty() {
	self.left_total = 1;
	self.right_total = 1;
	self.buf.clear();
	} else {
	self.check_stack(0);
	}
	let right = self.buf.push(BufEntry { token: Token::Break(token), size: -self.right_total });
	self.scan_stack.push_back(right);
	self.right_total += token.blank_space;
	}

	fn scan_string(&mut self, string: Cow<'static, str>) {
	if self.scan_stack.is_empty() {
	self.print_string(&string);
	} else {
	let len = string.len() as isize;
	self.buf.push(BufEntry { token: Token::String(string), size: len });
	self.right_total += len;
	self.check_stream();
	}
	}

	pub(crate) fn offset(&mut self, offset: isize) {
	if let Some(BufEntry { token: Token::Break(token), .. }) = &mut self.buf.last_mut() {
	token.offset += offset;
	}
	}

	fn check_stream(&mut self) {
	while self.right_total - self.left_total > self.space {
	if *self.scan_stack.front().unwrap() == self.buf.index_of_first() {
	self.scan_stack.pop_front().unwrap();
	self.buf.first_mut().unwrap().size = SIZE_INFINITY;
	}
	self.advance_left();
	if self.buf.is_empty() {
	break;
	}
	}
	}

	fn advance_left(&mut self) {
	while self.buf.first().unwrap().size >= 0 {
	let left = self.buf.pop_first().unwrap();

	match &left.token {
	Token::String(string) => {
	self.left_total += string.len() as isize;
	self.print_string(string);
	}
	Token::Break(token) => {
	self.left_total += token.blank_space;
	self.print_break(*token, left.size);
	}
	Token::Begin(token) => self.print_begin(*token, left.size),
	Token::End => self.print_end(),
	}

	self.last_printed = Some(left.token);

	if self.buf.is_empty() {
	break;
	}
	}
	}

	fn check_stack(&mut self, mut depth: usize) {
	while let Some(&index) = self.scan_stack.back() {
	let entry = &mut self.buf[index];
	match entry.token {
	Token::Begin(_) => {
	if depth == 0 {
	break;
	}
	self.scan_stack.pop_back().unwrap();
	entry.size += self.right_total;
	depth -= 1;
	}
	Token::End => {
	// paper says + not =, but that makes no sense.
	self.scan_stack.pop_back().unwrap();
	entry.size = 1;
	depth += 1;
	}
	_ => {
	self.scan_stack.pop_back().unwrap();
	entry.size += self.right_total;
	if depth == 0 {
	break;
	}
	}
	}
	}
	}

	fn get_top(&self) -> PrintFrame {
	*self
	.print_stack
	.last()
	.unwrap_or(&PrintFrame::Broken { indent: 0, breaks: Breaks::Inconsistent })
	}

	fn print_begin(&mut self, token: BeginToken, size: isize) {
	if size > self.space {
	self.print_stack.push(PrintFrame::Broken { indent: self.indent, breaks: token.breaks });
	self.indent = match token.indent {
	IndentStyle::Block { offset } => {
	usize::try_from(self.indent as isize + offset).unwrap()
	}
	IndentStyle::Visual => (MARGIN - self.space) as usize,
	};
	} else {
	self.print_stack.push(PrintFrame::Fits);
	}
	}

	fn print_end(&mut self) {
	if let PrintFrame::Broken { indent, .. } = self.print_stack.pop().unwrap() {
	self.indent = indent;
	}
	}

	fn print_break(&mut self, token: BreakToken, size: isize) {
	let fits = match self.get_top() {
	PrintFrame::Fits => true,
	PrintFrame::Broken { breaks: Breaks::Consistent, .. } => false,
	PrintFrame::Broken { breaks: Breaks::Inconsistent, .. } => size <= self.space,
	};
	if fits {
	self.pending_indentation += token.blank_space;
	self.space -= token.blank_space;
	} else {
	if let Some(pre_break) = token.pre_break {
	self.out.push(pre_break);
	}
	self.out.push('\n');
	let indent = self.indent as isize + token.offset;
	self.pending_indentation = indent;
	self.space = cmp::max(MARGIN - indent, MIN_SPACE);
	}
	}

	fn print_string(&mut self, string: &str) {
	// Write the pending indent. A more concise way of doing this would be:
	//
	// write!(self.out, "{: >n$}", "", n = self.pending_indentation as usize)?;
	//
	// But that is significantly slower. This code is sufficiently hot, and indents can get
	// sufficiently large, that the difference is significant on some workloads.
	self.out.reserve(self.pending_indentation as usize);
	self.out.extend(iter::repeat(' ').take(self.pending_indentation as usize));
	self.pending_indentation = 0;

	self.out.push_str(string);
	self.space -= string.len() as isize;
	}
	}