retex-engine 1.0.0-rc.2

//! The chief executive (*TeX: The Program* §1029–§1054) and the semantic
//! nest (§211–§219).
//!
//! `main_control` reads tokens and dispatches on the current mode and
//! command. Assignments go to [`crate::prefixed`]. List-building commands
//! go to the typesetting modules.

use std::rc::Rc;

use crate::Engine;
use crate::cmd::Cmd;
use crate::hash::NO_CS;
use crate::input::{Level, TokKind};
use crate::nodes::NodeRef;
use crate::primitives_gen as pg;
use crate::save::GroupCode;
use crate::token::Token;

/// A mode of TeX's semantic nest.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum Mode {
    /// No mode (during `\write` expansion, …).
    None,
    /// Vertical mode.
    Vertical,
    /// Internal vertical mode.
    InternalVertical,
    /// Horizontal mode.
    Horizontal,
    /// Restricted horizontal mode.
    RestrictedHorizontal,
    /// Display math mode.
    DisplayMath,
    /// Non-display math mode.
    Math,
}

/// The kind of a mode, ignoring whether it is inner.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum ModeKind {
    /// No mode.
    None,
    /// (Internal) vertical.
    Vertical,
    /// (Restricted) horizontal.
    Horizontal,
    /// (Display) math.
    Math,
}

impl Mode {
    /// The mode kind (`abs(mode)` in TeX).
    pub fn abs_kind(self) -> ModeKind {
        match self {
            Mode::None => ModeKind::None,
            Mode::Vertical | Mode::InternalVertical => ModeKind::Vertical,
            Mode::Horizontal | Mode::RestrictedHorizontal => ModeKind::Horizontal,
            Mode::DisplayMath | Mode::Math => ModeKind::Math,
        }
    }

    /// Whether the mode is inner (`mode < 0` in TeX).
    pub fn is_inner(self) -> bool {
        matches!(
            self,
            Mode::InternalVertical | Mode::RestrictedHorizontal | Mode::Math
        )
    }

    /// TeX's `mode > 0`.
    pub fn is_privileged(self) -> bool {
        matches!(self, Mode::Vertical | Mode::Horizontal | Mode::DisplayMath)
    }
}

/// The context in which a box is being built (`box_context`).
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum BoxContext {
    /// Shift by the given amount and append (`\moveleft`, `\raise`, or plain
    /// appending with 0).
    Shift(i32),
    /// `\setbox n=`.
    SetBox(u32),
    /// `\global\setbox n=`.
    GlobalSetBox(u32),
    /// `\shipout`.
    Shipout,
    /// Leaders of the given glue subtype.
    Leaders(u16),
    /// Handed to Lua (`token.scan_list`, see `Engine::scan_list`).
    Lua,
}

/// One level of the semantic nest (`list_state_record`).
#[derive(Clone, Debug)]
pub struct ListState {
    /// The mode.
    pub mode: Mode,
    /// The list head (a temp node).
    pub head: NodeRef,
    /// The last node of the list.
    pub tail: NodeRef,
    /// `\prevgraf`.
    pub prev_graf: i32,
    /// Line where this mode was entered (negative if the mode was entered
    /// by `\halign` or `\valign`).
    pub mode_line: i32,
    /// `\prevdepth` (vertical modes).
    pub prev_depth: i32,
    /// `\spacefactor` (horizontal modes).
    pub space_factor: i32,
    /// Current language (horizontal modes, `clang`).
    pub clang: i32,
    /// Incompleat noad (math modes).
    pub incompleat_noad: NodeRef,
    /// Math style (math modes).
    pub math_style: i32,
    /// The most recent `\left` or `\middle` fence of a math left group
    /// (e-TeX's `delim_ptr`).
    pub delim_ptr: NodeRef,
}

/// The `\prevdepth` value that suppresses interline glue
/// (`ignore_depth`).
pub const IGNORE_DEPTH: i32 = -65_536_000;

impl Engine {
    /// The current list state.
    pub fn cur_list(&self) -> &ListState {
        self.nest.last().expect("nest is never empty")
    }

    /// The current list state, mutably.
    pub fn cur_list_mut(&mut self) -> &mut ListState {
        self.nest.last_mut().expect("nest is never empty")
    }

    /// Enters a new semantic level (`push_nest`).
    pub fn push_nest(&mut self, mode: Mode) {
        let head = self.nodes.new_temp();
        let cl = ListState {
            mode,
            head,
            tail: head,
            prev_graf: 0,
            mode_line: self.input.line(),
            prev_depth: IGNORE_DEPTH,
            space_factor: 1000,
            clang: 0,
            incompleat_noad: NodeRef::NULL,
            math_style: 0,
            delim_ptr: NodeRef::NULL,
        };
        self.nest.push(cl);
        self.mode = mode;
    }

    /// Leaves a semantic level (`pop_nest`). The list is not freed.
    pub fn pop_nest(&mut self) {
        if let Some(cl) = self.nest.pop() {
            self.nodes.free(cl.head);
        }
        self.mode = self.cur_list().mode;
    }

    /// Appends node `p` to the current list (`tail_append`).
    pub fn tail_append(&mut self, p: NodeRef) {
        let tail = self.cur_list().tail;
        self.nodes.link(tail, p);
        self.cur_list_mut().tail = p;
    }

    /// Prints a mode name (`print_mode`).
    pub fn print_mode(&mut self, m: Mode) {
        let s = match m {
            Mode::Vertical => "vertical mode",
            Mode::Horizontal => "horizontal mode",
            Mode::DisplayMath => "display math mode",
            Mode::None => "no mode",
            Mode::InternalVertical => "internal vertical mode",
            Mode::RestrictedHorizontal => "restricted horizontal mode",
            Mode::Math => "math mode",
        };
        self.printer.print_str(s);
    }

    /// The chief executive: reads and executes commands until `\end` or
    /// `\dump` (`main_control`).
    pub fn main_control(&mut self) {
        let every_job = self.eqtb.toks_param(pg::assign_toks::EVERYJOB).clone();
        if !every_job.is_empty() {
            self.begin_token_list(every_job, TokKind::EveryJobText);
        }
        self.main_loop();
    }

    /// Executes `list` in a nested main loop until it is used up (LuaTeX's
    /// local control, used by `tex.runtoks`). Lua output pushed before the
    /// call is read first.
    pub fn run_local(&mut self, list: Option<crate::token::TokenList>) {
        let end: crate::token::TokenList = Rc::from(vec![Token::cs(self.cs.end_local)]);
        // The end marker goes below the material to execute.
        let depth = self.input.stack.len();
        self.begin_token_list(end, TokKind::Inserted);
        if let Some(l) = list {
            self.begin_token_list(l, TokKind::Inserted);
        }
        let _ = depth;
        self.local_level += 1;
        self.main_loop();
        self.local_level -= 1;
    }

    /// Like [`run_local`](Self::run_local), but the material is whatever
    /// is already on top of the input stack above the marker inserted
    /// here. Returns the depth at which the marker was placed.
    pub fn push_local_marker(&mut self) {
        let end: crate::token::TokenList = Rc::from(vec![Token::cs(self.cs.end_local)]);
        self.begin_token_list(end, TokKind::Inserted);
    }

    /// LuaTeX's `token.scan_list`: scans a box specification and returns
    /// the box (null if there was none), running the commands of an
    /// `\hbox{…}` or `\vbox{…}` in local control.
    pub fn scan_list(&mut self) -> NodeRef {
        self.lua_scanned_box = None;
        let level = self.save.cur_level;
        self.scan_box(BoxContext::Lua);
        if self.lua_scanned_box.is_none() && self.save.cur_level > level {
            let waiting = std::mem::replace(&mut self.lua_scan_waiting, true);
            self.run_local_loop();
            self.lua_scan_waiting = waiting;
        }
        self.lua_scanned_box.take().unwrap_or(NodeRef::NULL)
    }

    /// Runs the nested main loop until the local-control marker is met.
    pub fn run_local_loop(&mut self) {
        self.local_level += 1;
        self.main_loop();
        self.local_level -= 1;
    }

    /// LuaTeX's fatal error for a node freed twice (Lua code can put a
    /// node in two lists, or link a list into a loop).
    #[inline]
    pub fn check_double_free(&mut self) {
        if let Some((t, p)) = self.nodes.double_free {
            let name = crate::nodes::type_name(t);
            self.luatex_fatal(
                "nodes",
                &format!("attempt to double-free {name} node {p}, ignored"),
            );
        }
    }

    fn main_loop(&mut self) {
        loop {
            // big_switch:
            self.get_x_token();
            loop {
                // reswitch:
                if self.eqtb.int(pg::assign_int::TRACINGCOMMANDS) > 0 {
                    self.show_cur_cmd_chr();
                }
                match self.dispatch() {
                    Next::BigSwitch => {
                        self.check_double_free();
                        break;
                    }
                    Next::Reswitch => continue,
                    Next::Return => return,
                }
            }
        }
    }

    /// One step of `main_control`.
    fn dispatch(&mut self) -> Next {
        use ModeKind as K;
        let kind = self.mode.abs_kind();
        let cmd = self.cur_cmd;
        match (kind, cmd) {
            (K::Horizontal, Cmd::Letter | Cmd::OtherChar | Cmd::CharGiven) => {
                self.main_loop_char();
                Next::Reswitch
            }
            (K::Horizontal, Cmd::CharNum) => {
                self.scan_char_num();
                self.cur_chr = self.cur_val as u32;
                self.main_loop_char();
                Next::Reswitch
            }
            (K::Horizontal, Cmd::Boundary) => {
                self.append_boundary();
                Next::BigSwitch
            }
            (K::Horizontal, Cmd::Spacer) => {
                self.app_space_main();
                Next::BigSwitch
            }
            (K::Horizontal | K::Math, Cmd::ExSpace) => {
                self.app_explicit_space();
                Next::BigSwitch
            }
            // Observed: in local control (`tex.runtoks`), a space in
            // vertical mode appends an interword space, as in horizontal
            // mode.
            (K::Vertical, Cmd::Spacer) if self.local_level > 0 => {
                self.append_normal_space();
                Next::BigSwitch
            }
            (_, Cmd::Relax) | (K::Vertical | K::Math, Cmd::Spacer) => Next::BigSwitch,
            (K::Math, Cmd::Boundary) => Next::BigSwitch,
            (_, Cmd::IgnoreSpaces) => {
                if self.cur_chr == 0 {
                    self.get_nonblank_noncall();
                    Next::Reswitch
                } else {
                    // LuaTeX's \nospaces etc. are parameters; nothing to do.
                    Next::BigSwitch
                }
            }
            (K::Vertical, Cmd::Stop) => {
                if self.its_all_over() {
                    Next::Return
                } else {
                    Next::BigSwitch
                }
            }
            (K::Vertical, Cmd::Vmove)
            | (K::Horizontal | K::Math, Cmd::Hmove)
            | (_, Cmd::LastItem)
            | (_, Cmd::MacParam) => {
                self.report_illegal_case();
                Next::BigSwitch
            }
            (_, c) if c > Cmd::MAX_NON_PREFIXED => {
                self.prefixed_command();
                Next::BigSwitch
            }
            (_, Cmd::BeginGroup) => {
                self.new_save_level(GroupCode::SemiSimple);
                self.reset_local_counts(true);
                Next::BigSwitch
            }
            (_, Cmd::EndGroup) => {
                if self.save.cur_group == GroupCode::SemiSimple {
                    self.end_group_text_dirs();
                    self.unsave();
                } else {
                    self.off_save();
                }
                Next::BigSwitch
            }
            (_, Cmd::LeftBrace) if kind != K::Math => {
                self.new_save_level(GroupCode::Simple);
                self.reset_local_counts(true);
                Next::BigSwitch
            }
            (_, Cmd::RightBrace) => {
                self.handle_right_brace();
                Next::BigSwitch
            }
            (_, Cmd::AfterAssignment) => {
                self.get_token();
                self.after_token = self.cur_tok;
                Next::BigSwitch
            }
            (_, Cmd::AfterGroup) => {
                self.get_token();
                self.save_for_after(self.cur_tok);
                Next::BigSwitch
            }
            (_, Cmd::InStream) => {
                self.open_or_close_in();
                Next::BigSwitch
            }
            (_, Cmd::Message) => {
                self.issue_message();
                Next::BigSwitch
            }
            (_, Cmd::CaseShift) => {
                self.shift_case();
                Next::BigSwitch
            }
            (_, Cmd::Xray) => {
                self.show_whatever();
                Next::BigSwitch
            }
            (_, Cmd::Extension)
                if self.cur_chr == crate::misc::ext::END_LOCAL_CONTROL && self.local_level > 0 =>
            {
                Next::Return
            }
            (_, Cmd::Extension) => {
                self.do_extension();
                Next::BigSwitch
            }
            (_, Cmd::Normal) => {
                self.do_normal();
                Next::BigSwitch
            }
            (_, Cmd::LuaFunctionCall | Cmd::LuaBytecodeCall | Cmd::LuaCall) => {
                let n = if self.cur_cmd == Cmd::LuaFunctionCall
                    || self.cur_cmd == Cmd::LuaBytecodeCall
                {
                    self.scan_int();
                    self.cur_val
                } else {
                    self.cur_chr as i32
                };
                if self.cur_cmd == Cmd::LuaBytecodeCall {
                    self.call_lua_bytecode(n);
                } else {
                    self.call_lua_function(n);
                }
                Next::BigSwitch
            }
            (_, Cmd::PartokenName) => {
                self.partoken_name();
                Next::BigSwitch
            }
            _ => {
                self.typeset_command();
                Next::BigSwitch
            }
        }
    }

    /// Reports a command that is not allowed in this mode
    /// (`report_illegal_case`).
    pub fn report_illegal_case(&mut self) {
        self.you_cant();
        self.help(&[
            "Sorry, but I'm not programmed to handle this case;",
            "I'll just pretend that you didn't ask for it.",
            "If you're in the wrong mode, you might be able to",
            "return to the right one by typing `I}' or `I$' or `I\\par'.",
        ]);
        self.error();
    }

    /// Prints "You can't use … in … mode" (`you_cant`).
    pub fn you_cant(&mut self) {
        self.print_err("You can't use `");
        self.print_cmd_chr(self.cur_cmd, self.cur_chr);
        self.printer.print_str("' in ");
        self.print_mode(self.mode);
    }

    /// Whether the mode is privileged; reports an error if not
    /// (`privileged`).
    pub fn privileged(&mut self) -> bool {
        if self.mode.is_privileged() {
            true
        } else {
            self.report_illegal_case();
            false
        }
    }

    /// Inserts a missing `$` (`insert_dollar_sign`).
    pub fn insert_dollar_sign(&mut self) {
        self.back_input();
        self.cur_tok = Token::char(Cmd::MathShift, '$' as u32);
        self.print_err("Missing $ inserted");
        self.help(&[
            "I've inserted a begin-math/end-math symbol since I think",
            "you left one out. Proceed, with fingers crossed.",
        ]);
        self.ins_error();
    }

    /// `\end` and `\dump`: finish if everything is shipped out
    /// (`its_all_over`).
    fn its_all_over(&mut self) -> bool {
        if self.privileged() {
            let cl = self.cur_list();
            if self.page_is_empty() && cl.head == cl.tail && self.dead_cycles == 0 {
                return true;
            }
            self.back_input();
            let b = self.nodes.new_null_box(true);
            let hsize = self.eqtb.dimen(pg::assign_dimen::HSIZE);
            if let Some(bd) = self.nodes.box_data_mut(b) {
                bd.width = hsize;
            }
            self.tail_append(b);
            let g = self.nodes.new_glue(
                crate::glue::Glue {
                    stretch: crate::glue::Glue::fixed(0).stretch + retex_arith::UNITY,
                    stretch_order: crate::glue::FILL,
                    ..crate::glue::Glue::ZERO
                },
                0,
            );
            self.tail_append(g);
            let p = self.nodes.new_penalty(-0o10000000000, 0);
            self.tail_append(p);
            self.build_page_from("end");
        }
        false
    }

    /// Handles a group-ending `}` (`handle_right_brace`).
    pub fn handle_right_brace(&mut self) {
        match self.save.cur_group {
            GroupCode::Simple => {
                self.end_group_text_dirs();
                self.unsave();
            }
            GroupCode::BottomLevel => {
                self.print_err("Too many }'s");
                self.help(&[
                    "You've closed more groups than you opened.",
                    "Such booboos are generally harmless, so keep going.",
                ]);
                self.error();
            }
            GroupCode::SemiSimple | GroupCode::MathShift | GroupCode::MathLeft => {
                self.extra_right_brace()
            }
            _ => self.handle_right_brace_typesetting(),
        }
    }

    fn extra_right_brace(&mut self) {
        self.print_err("Extra }, or forgotten ");
        match self.save.cur_group {
            GroupCode::SemiSimple => self.print_esc_str("endgroup"),
            GroupCode::MathShift => self.printer.print_char('$' as u32),
            _ => self.print_esc_str("right"),
        }
        self.help(&[
            "I've deleted a group-closing symbol because it seems to be",
            "spurious, as in `$x}$'. But perhaps the } is legitimate and",
            "you forgot something else, as in `\\hbox{$x}'. In such cases",
            "the way to recover is to insert both the forgotten and the",
            "deleted material, e.g., by typing `I$}'.",
        ]);
        self.error();
        self.input.align_state += 1;
    }

    /// Recovers from a mismatched group end (`off_save`, §1064).
    pub fn off_save(&mut self) {
        if self.save.cur_group == GroupCode::BottomLevel {
            // Drop current token and complain that it was unmatched.
            self.print_err("Extra ");
            self.print_cmd_chr(self.cur_cmd, self.cur_chr);
            self.help(&["Things are pretty mixed up, but I think the worst is over."]);
            self.error();
            return;
        }
        self.back_input();
        let mut list: Vec<Token> = Vec::new();
        self.print_err("Missing ");
        match self.save.cur_group {
            GroupCode::SemiSimple => {
                list.push(Token::cs(self.cs.frozen_end_group));
                self.print_esc_str("endgroup");
            }
            GroupCode::MathShift => {
                list.push(Token::char(Cmd::MathShift, '$' as u32));
                self.printer.print_char('$' as u32);
            }
            GroupCode::MathLeft => {
                list.push(Token::cs(self.cs.frozen_right));
                list.push(Token::other('.' as u32));
                self.print_esc_str("right.");
            }
            _ => {
                list.push(Token::RIGHT_BRACE);
                self.printer.print_char('}' as u32);
            }
        }
        self.printer.print_str(" inserted");
        self.ins_list(Rc::from(list));
        self.help(&[
            "I've inserted something that you may have forgotten.",
            "(See the <inserted text> above.)",
            "With luck, this will get me unwedged. But if you",
            "really didn't forget anything, try typing `2' now; then",
            "my insertion and my current dilemma will both disappear.",
        ]);
        self.error();
    }

    /// `\message` and `\errmessage` (`issue_message`).
    fn issue_message(&mut self) {
        let c = self.cur_chr;
        let toks = self.scan_toks_expanded();
        let s = self.print_to_string(|e| e.show_token_list(&toks, 0, toks.len(), 10_000_000));
        if c == 0 {
            if self.printer.term_offset + s.len() > self.printer.max_print_line.saturating_sub(2) {
                self.printer.print_ln();
            } else if self.printer.term_offset > 0 || self.printer.file_offset > 0 {
                self.printer.print_char(' ' as u32);
            }
            self.printer.print_bytes(&s);
            self.printer.update_terminal();
        } else {
            self.print_err("");
            self.printer.print_bytes(&s);
            let err_help = self.eqtb.toks_param(pg::assign_toks::ERRHELP);
            if !err_help.is_empty() {
                self.err.use_err_help = true;
            } else if self.long_help_seen {
                self.help(&["(That was another \\errmessage.)"]);
            } else {
                if self.err.interaction < crate::error::Interaction::ErrorStop {
                    self.long_help_seen = true;
                }
                self.help(&[
                    "This error message was generated by an \\errmessage",
                    "command, so I can't give any explicit help.",
                    "Pretend that you're Hercule Poirot: Examine all clues,",
                    "and deduce the truth by order and method.",
                ]);
            }
            self.error();
            self.err.use_err_help = false;
        }
    }

    /// `\lowercase` and `\uppercase` (`shift_case`).
    fn shift_case(&mut self) {
        let b = self.cur_chr;
        let toks = self.scan_general_text();
        let upper = b == pg::def_char_code::UCCODE + UC_LC_OFFSET;
        let mut out: Vec<Token> = Vec::with_capacity(toks.len());
        for &t in toks.iter() {
            if t.is_cs() {
                if let crate::hash::CsKind::Active(c) = *self.hash.kind(t.cs_id()) {
                    let m = self.case_map(c, upper);
                    if m != 0 {
                        out.push(Token::cs(self.hash.active(m)));
                        continue;
                    }
                }
                out.push(t);
            } else {
                let c = t.chr();
                let m = self.case_map(c, upper);
                if m != 0 {
                    out.push(Token::char(t.cmd(), m));
                } else {
                    out.push(t);
                }
            }
        }
        self.back_list(Rc::from(out));
    }

    fn case_map(&self, c: u32, upper: bool) -> u32 {
        (if upper {
            self.eqtb.uc.get(c)
        } else {
            self.eqtb.lc.get(c)
        }) as u32
    }

    /// Cleans up after `main_control` ends (`final_cleanup`).
    pub fn final_cleanup(&mut self) {
        let c = self.cur_chr;
        if c != 1 {
            self.eqtb.int_par[pg::assign_int::NEWLINECHAR as usize].val = -1;
            self.printer.new_line_char = -1;
        }
        if self.job_name.is_none() {
            self.open_log_file();
        }
        while self.input.stack.len() > 1 {
            match self.input.stack.last() {
                Some(Level::Tokens(_)) => self.end_token_list(),
                _ => self.end_file_reading(),
            }
        }
        while self.input.open_parens > 0 {
            // LuaTeX prints no space before the parenthesis.
            self.printer.print_char(')' as u32);
            self.input.open_parens -= 1;
        }
        if self.save.cur_level > crate::eqtb::LEVEL_ONE {
            self.printer.print_nl("(");
            self.print_esc_str("end occurred ");
            self.printer.print_str("inside a group at level ");
            self.printer
                .print_int(i64::from(self.save.cur_level - crate::eqtb::LEVEL_ONE));
            self.printer.print_char(')' as u32);
            self.show_save_groups();
        }
        while !self.cond.stack.is_empty() {
            self.printer.print_nl("(");
            self.print_esc_str("end occurred ");
            self.printer.print_str("when ");
            self.print_cmd_chr(Cmd::IfTest, self.cond.cur_if);
            if self.cond.if_line != 0 {
                self.printer.print_str(" on line ");
                self.printer.print_int(self.cond.if_line.into());
            }
            self.printer.print_str(" was incomplete)");
            self.pop_cond();
        }
        if self.err.history != crate::error::History::Spotless
            && (self.err.history == crate::error::History::WarningIssued
                || self.err.interaction < crate::error::Interaction::ErrorStop)
            && self.printer.selector == crate::print::Selector::TermAndLog
        {
            self.printer.selector = crate::print::Selector::TermOnly;
            self.printer
                .print_nl("(see the transcript file for additional information)");
            self.printer.selector = crate::print::Selector::TermAndLog;
        }
        if c == 1 {
            if self.ini_version {
                self.dump_requested = true;
                return;
            }
            self.printer
                .print_nl("(\\dump is performed only by INITEX)");
        }
    }

    /// LuaTeX's `\partokenname` (pdfTeX extension): sets the name of the
    /// paragraph-ending token.
    fn partoken_name(&mut self) {
        self.get_token();
        if self.cur_cs != NO_CS {
            self.par_token_cs = self.cur_cs;
        }
    }
}

/// `\lowercase` and `\uppercase` carry the `def_char_code` modifiers of
/// `\lccode`/`\uccode` plus this offset (LuaTeX numbering).
pub const UC_LC_OFFSET: u32 = 333_280;

/// What `main_control` does next.
pub enum Next {
    /// Read the next token.
    BigSwitch,
    /// Dispatch the current token again.
    Reswitch,
    /// Leave `main_control`.
    Return,
}