path: root/src/makefile/mod.rs

use std::cell::{Cell, RefCell};
use std::collections::HashMap;
use std::env;
use std::fmt;
use std::fs::File;
use std::io::{BufRead, BufReader, Error as IoError};
use std::path::Path;
use std::rc::Rc;

use crate::args::Args;
use lazy_static::lazy_static;
use regex::Regex;

mod command_line;
mod conditional;
mod functions;
mod inference_rules;
mod r#macro;
mod pattern;
mod target;
mod token;

use command_line::CommandLine;
use conditional::{ConditionalLine, ConditionalState};
use inference_rules::InferenceRule;
use r#macro::{MacroSet, MacroSource};
use target::Target;
use token::{tokenize, Token, TokenString};

enum LineType {
    Rule,
    Macro,
    Unknown,
}

impl LineType {
    fn of(line_tokens: &TokenString) -> Self {
        for token in line_tokens.tokens() {
            if let Token::Text(text) = token {
                let colon_idx = text.find(':');
                let equals_idx = text.find('=');
                match (colon_idx, equals_idx) {
                    (Some(_), None) => {
                        return Self::Rule;
                    }
                    (Some(c), Some(e)) if c < e => {
                        return Self::Rule;
                    }
                    (None, Some(_)) => {
                        return Self::Macro;
                    }
                    (Some(c), Some(e)) if e < c => {
                        return Self::Macro;
                    }
                    _ => {}
                }
            }
        }
        Self::Unknown
    }
}

fn inference_match<'a>(
    targets: &[&'a str],
    prerequisites: &[String],
) -> Option<regex::Captures<'a>> {
    lazy_static! {
        static ref INFERENCE_RULE: Regex =
            Regex::new(r"^(?P<s2>(\.[^/.]+)?)(?P<s1>\.[^/.]+)$").unwrap();
        static ref SPECIAL_TARGET: Regex = Regex::new(r"^\.[A-Z]+$").unwrap();
    }

    let inference_match = INFERENCE_RULE.captures(targets[0]);
    let special_target_match = SPECIAL_TARGET.captures(targets[0]);

    let inference_rule = targets.len() == 1
        && prerequisites.is_empty()
        && inference_match.is_some()
        && special_target_match.is_none();
    if inference_rule {
        inference_match
    } else {
        None
    }
}

pub(crate) struct Makefile<'a> {
    inference_rules: Vec<InferenceRule>,
    macros: MacroSet<'static, 'static>,
    targets: RefCell<HashMap<String, Rc<RefCell<Target>>>>,
    pub(crate) first_non_special_target: Option<String>,
    args: &'a Args,
    // TODO borrow warnings from Python version
}

impl<'a> Makefile<'a> {
    pub(crate) fn new(args: &'a Args) -> Self {
        let mut inference_rules = vec![];
        let mut macros = MacroSet::new();
        let mut targets = HashMap::new();
        let first_non_special_target = None;

        if !args.no_builtin_rules {
            inference_rules.extend(builtin_inference_rules());
            macros.add_builtins();
            targets.extend(
                builtin_targets()
                    .into_iter()
                    .map(|t| (t.name.clone(), Rc::new(RefCell::new(t)))),
            );
        }

        macros.add_env();

        for r#macro in args.macros() {
            if let [name, value] = *r#macro.splitn(2, '=').collect::<Vec<_>>() {
                macros.set(
                    name.into(),
                    MacroSource::CommandLineOrMakeflags,
                    TokenString::text(value),
                );
            }
        }

        Makefile {
            inference_rules,
            macros,
            targets: RefCell::new(targets),
            first_non_special_target,
            args,
        }
    }

    pub(crate) fn and_read_file(&mut self, path: impl AsRef<Path>) {
        let file = File::open(path);
        // TODO handle errors
        let file = file.expect("couldn't open makefile!");
        let file_reader = BufReader::new(file);
        self.and_read(file_reader);
    }

    pub(crate) fn and_read(&mut self, source: impl BufRead) {
        let mut lines_iter = source.lines().enumerate().peekable();
        let mut conditional_stack: Vec<ConditionalState> = vec![];
        while let Some((line_number, line)) = lines_iter.next() {
            // TODO handle I/O errors at all
            let mut line = line.expect("failed to read line of makefile!");

            // handle escaped newlines
            while line.ends_with('\\') {
                line.pop();
                line.push(' ');
                if let Some((_, x)) = lines_iter.next() {
                    line.push_str(x.expect("failed to read line of makefile!").trim_start())
                }
            }

            // handle comments
            lazy_static! {
                static ref COMMENT: Regex = Regex::new("#.*$").unwrap();
            }
            let line = COMMENT.replace(&line, "").into_owned();

            // skip lines if we need to
            if conditional_stack
                .last()
                .map_or(false, ConditionalState::skipping)
            {
                continue;
            }

            // handle include lines
            if let Some(line) = line.strip_prefix("include ") {
                // remove extra leading space
                let line = line.trim_start();
                let line = self.expand_macros(&tokenize(line), None);
                let fields = line.split_whitespace();
                // POSIX says we only have to handle a single filename, but GNU make
                // handles arbitrarily many filenames, and it's not like that's more work
                for field in fields {
                    self.and_read_file(field);
                }
            } else if let Some(line) = ConditionalLine::from(&line, |t| self.expand_macros(t, None))
            {
                line.action(
                    conditional_stack.last(),
                    |name| self.macros.is_defined(name),
                    |t| self.expand_macros(t, None),
                )
                .apply_to(&mut conditional_stack);
            } else if line.trim().is_empty() {
                // handle blank lines
                continue;
            } else {
                // unfortunately, rules vs macros can't be determined until after
                // macro tokenizing. so that's suboptimal.

                // TODO errors
                let line_tokens: TokenString = line.parse().unwrap();

                let line_type = LineType::of(&line_tokens);

                match line_type {
                    LineType::Rule => self.read_rule(&line_tokens, &mut lines_iter),
                    LineType::Macro => self.read_macro(&line_tokens),
                    LineType::Unknown => {
                        panic!(
                            "error: line {}: unknown line {:?}",
                            line_number, line_tokens
                        );
                    }
                }
            }
        }
    }

    fn read_rule(
        &mut self,
        line_tokens: &TokenString,
        lines_iter: &mut impl Iterator<Item = (usize, Result<String, IoError>)>,
    ) {
        let mut lines_iter = lines_iter.peekable();
        let (targets, not_targets) = line_tokens.split_once(':').unwrap();
        let targets = self.expand_macros(&targets, None);
        let targets = targets.split_whitespace().collect::<Vec<_>>();
        let (prerequisites, mut commands) = match not_targets.split_once(';') {
            Some((prerequisites, mut command)) => {
                while command.ends_with("\\") && lines_iter.peek().is_some() {
                    command.strip_suffix("\\");
                    command.extend(tokenize(&lines_iter.next().unwrap().1.unwrap()));
                }
                (prerequisites, vec![command])
            }
            None => (not_targets, vec![]),
        };
        let prerequisites = self.expand_macros(&prerequisites, None);
        let prerequisites = prerequisites
            .split_whitespace()
            .map(|x| x.into())
            .collect::<Vec<String>>();

        while let Some((_, x)) = lines_iter.next_if(|(_, x)| {
            x.as_ref()
                .ok()
                .map_or(false, |line| line.starts_with('\t') || line.is_empty())
        }) {
            let mut line = x.unwrap();
            if !line.is_empty() {
                line.remove(0);
            }
            if line.is_empty() {
                continue;
            }
            while line.ends_with('\\') {
                match lines_iter.next() {
                    Some((_, Ok(next_line))) => {
                        let next_line = next_line.strip_prefix("\t").unwrap_or(&next_line);
                        line.push('\n');
                        line.push_str(next_line);
                    }
                    _ => break,
                }
            }
            commands.push(line.parse().unwrap());
        }

        let commands = commands
            .into_iter()
            .map(CommandLine::from)
            .collect::<Vec<_>>();

        if targets.is_empty() {
            return;
        }

        // we don't know yet if it's a target rule or an inference rule
        let inference_match = inference_match(&targets, &prerequisites);

        if let Some(inference_match) = inference_match {
            let new_rule = InferenceRule {
                product: inference_match.name("s1").unwrap().as_str().to_string(),
                prereq: inference_match.name("s2").unwrap().as_str().to_string(),
                commands,
            };

            self.inference_rules.retain(|existing_rule| {
                (&existing_rule.prereq, &existing_rule.product)
                    != (&new_rule.prereq, &new_rule.product)
            });
            self.inference_rules.push(new_rule);
        } else {
            for target in targets {
                if self.first_non_special_target.is_none() && !target.starts_with('.') {
                    self.first_non_special_target = Some(target.into());
                }
                let mut targets = self.targets.borrow_mut();
                match targets.get_mut(target) {
                    Some(old_target)
                        if commands.is_empty()
                            && !(target == ".SUFIXES" && prerequisites.is_empty()) =>
                    {
                        let mut old_target = old_target.borrow_mut();
                        let new_prerequisites = prerequisites
                            .iter()
                            .filter(|x| !old_target.prerequisites.contains(x))
                            .cloned()
                            .collect::<Vec<_>>();
                        old_target.prerequisites.extend(new_prerequisites);
                    }
                    _ => {
                        let new_target = Target {
                            name: target.into(),
                            prerequisites: prerequisites.clone(),
                            commands: commands.clone(),
                            already_updated: Cell::new(false),
                        };
                        targets.insert(target.into(), Rc::new(RefCell::new(new_target)));
                    }
                }
            }
        }
    }

    fn read_macro(&mut self, line_tokens: &TokenString) {
        let (name, mut value) = line_tokens.split_once('=').unwrap();
        let name = self.expand_macros(&name, None);
        // GNUisms are annoying, but popular
        let mut expand_value = false;
        let mut skip_if_defined = false;
        let mut append = false;
        let name = if let Some(real_name) = name.strip_suffix("::") {
            expand_value = true;
            real_name
        } else if let Some(real_name) = name.strip_suffix(":") {
            expand_value = true;
            real_name
        } else if let Some(real_name) = name.strip_suffix("?") {
            skip_if_defined = true;
            real_name
        } else if let Some(real_name) = name.strip_suffix("+") {
            append = true;
            real_name
        } else {
            &name
        };

        let name = name.trim_end();
        value.trim_start();

        let value = if expand_value {
            TokenString::text(self.expand_macros(&value, None))
        } else {
            value
        };

        match self.macros.get(name) {
            // We always let command line or MAKEFLAGS macros override macros from the file.
            Some((MacroSource::CommandLineOrMakeflags, _)) => return,
            // We let environment variables override macros from the file only if the command-line argument to do that was given
            Some((MacroSource::Environment, _)) if self.args.environment_overrides => return,
            _ if skip_if_defined => return,
            _ => {}
        }

        let value = match self.macros.pop(name) {
            Some((_, mut old_value)) if append => {
                // TODO eagerly expand if appending to eagerly-expanded macro
                old_value.extend(TokenString::text(" "));
                old_value.extend(value);
                old_value
            }
            _ => value,
        };
        self.macros.set(name.into(), MacroSource::File, value);
    }

    fn special_target_has_prereq(&self, target: &str, name: &str) -> bool {
        let targets = self.targets.borrow();
        match targets.get(target) {
            Some(target) => {
                let target = target.borrow();
                target.prerequisites.is_empty() || target.prerequisites.iter().any(|e| e == name)
            }
            None => false,
        }
    }

    pub(crate) fn get_target(&self, name: &str) -> Rc<RefCell<Target>> {
        // TODO implement .POSIX
        let follow_gnu = true;

        let vpath_options = match self.macros.get("VPATH") {
            Some((_, vpath)) if follow_gnu => {
                let vpath = self.expand_macros(vpath, None);
                env::split_paths(&vpath).collect()
            }
            _ => vec![],
        };

        let targets = self.targets.borrow();
        let mut new_target = None;
        let exists_but_infer_anyway = if follow_gnu {
            targets
                .get(name)
                .map_or(false, |target| target.borrow().commands.is_empty())
        } else {
            false
        };
        if !targets.contains_key(name) || exists_but_infer_anyway {
            // When no target rule is found to update a target, the inference rules shall
            // be checked. The suffix of the target to be built...
            let suffix = Path::new(name)
                .extension()
                .map_or_else(String::new, |ext| format!(".{}", ext.to_string_lossy()));
            // is compared to the list of suffixes specified by the .SUFFIXES special
            // targets. If the .s1 suffix is found in .SUFFIXES...
            if self.special_target_has_prereq(".SUFFIXES", &suffix) || suffix.is_empty() {
                // the inference rules shall be searched in the order defined...
                'rules: for rule in self
                    .inference_rules
                    .iter()
                    // for the first .s2.s1 rule...
                    .filter(|rule| rule.product == suffix)
                {
                    // whose prerequisite file ($*.s2) exists.
                    let prereq_path =
                        Path::new(name).with_extension(rule.prereq.trim_start_matches('.'));
                    if let Some(prereq) = std::iter::once(prereq_path.clone())
                        .chain(
                            if prereq_path.is_absolute() {
                                None
                            } else {
                                Some(vpath_options.iter().map(|vpath| vpath.join(&prereq_path)))
                            }
                            .into_iter()
                            .flatten(),
                        )
                        .find(|prereq| prereq.exists())
                    {
                        new_target = Some(Target {
                            name: name.into(),
                            prerequisites: vec![prereq.to_string_lossy().into()],
                            commands: rule.commands.clone(),
                            already_updated: Cell::new(false),
                        });
                        break 'rules;
                    }
                }
            }
        }

        if !targets.contains_key(name) && new_target.is_none() {
            // well, inference didn't work. is there a default?
            if let Some(default) = targets.get(".DEFAULT") {
                let commands = default.borrow().commands.clone();
                new_target = Some(Target {
                    name: name.into(),
                    prerequisites: vec![],
                    commands,
                    already_updated: Cell::new(false),
                });
            } else {
                // if it already exists, it counts as up-to-date
                if Path::new(name).exists() {
                    new_target = Some(Target {
                        name: name.into(),
                        prerequisites: vec![],
                        commands: vec![],
                        already_updated: Cell::new(true),
                    });
                }
            }
        }

        drop(targets);
        if let Some(new_target) = new_target {
            self.targets
                .borrow_mut()
                .insert(new_target.name.clone(), Rc::new(RefCell::new(new_target)));
        }

        let targets = self.targets.borrow();
        targets.get(name).expect("Target not found!").clone()
    }

    pub(crate) fn update_target(&self, name: &str) {
        self.get_target(name).borrow().update(self);
    }

    fn expand_macros(&self, text: &TokenString, target: Option<&Target>) -> String {
        let target = target.cloned();
        let lookup_internal = move |name: &str| {
            let target = target
                .as_ref()
                .expect("internal macro but no current target!");
            let macro_pieces = if name.starts_with('@') {
                // The $@ shall evaluate to the full target name of the
                // current target.
                vec![target.name.clone()]
            } else if name.starts_with('?') {
                // The $? macro shall evaluate to the list of prerequisites
                // that are newer than the current target.
                target
                    .prerequisites
                    .iter()
                    .filter(|prereq| {
                        self.get_target(prereq)
                            .borrow()
                            .newer_than(target)
                            .unwrap_or(false)
                    })
                    .cloned()
                    .collect()
            } else if name.starts_with('<') {
                // In an inference rule, the $< macro shall evaluate to the
                // filename whose existence allowed the inference rule to be
                // chosen for the target. In the .DEFAULT rule, the $< macro
                // shall evaluate to the current target name.
                target.prerequisites.clone()
            } else if name.starts_with('*') {
                // The $* macro shall evaluate to the current target name with
                // its suffix deleted.
                vec![Path::new(name).with_extension("").to_string_lossy().into()]
            } else {
                unreachable!()
            };

            let macro_pieces = if name.ends_with('D') {
                macro_pieces
                    .into_iter()
                    .map(|x| {
                        Path::new(&x)
                            .parent()
                            .expect("no parent")
                            .to_string_lossy()
                            .into()
                    })
                    .collect()
            } else if name.ends_with('F') {
                macro_pieces
                    .into_iter()
                    .map(|x| {
                        Path::new(&x)
                            .file_name()
                            .expect("no filename")
                            .to_string_lossy()
                            .into()
                    })
                    .collect()
            } else {
                macro_pieces
            };

            macro_pieces.join(" ")
        };

        self.macros.with_lookup(&lookup_internal).expand(text)
    }
}

impl fmt::Display for Makefile<'_> {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        let header = |f: &mut fmt::Formatter, t: &str| {
            writeln!(f, "{}\n{:=^width$}", t, "", width = t.len())
        };
        header(f, "Inference Rules")?;
        for rule in &self.inference_rules {
            writeln!(f, "{}", rule)?;
        }
        writeln!(f)?;

        header(f, "Macros")?;
        writeln!(f, "{}", &self.macros)?;
        writeln!(f)?;

        header(f, "Targets")?;
        for target in self.targets.borrow().values() {
            writeln!(f, "{}", target.borrow())?;
        }

        Ok(())
    }
}

fn builtin_inference_rules() -> Vec<InferenceRule> {
    // This is a terrible idea.
    macro_rules! prepend_dot {
        ($x:tt) => {
            concat!(".", stringify!($x))
        };
        () => {
            ""
        };
    }

    macro_rules! make {
        {$(.$first:tt$(.$second:tt)?:
            $($cmd:literal)+)+} => {
            vec![$(
                InferenceRule {
                    product: prepend_dot!($($second)?).into(),
                    prereq: concat!(".", stringify!($first)).into(),
                    commands: vec![$(CommandLine::from($cmd.parse().unwrap())),+],
                }
            ),+]
        };
    }

    make! {
        .c:
            "$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $<"
        .f:
            "$(FC) $(FFLAGS) $(LDFLAGS) -o $@ $<"
        .sh:
            "cp $< $@"
            "chmod a+x $@"

        .c.o:
            "$(CC) $(CFLAGS) -c $<"
        .f.o:
            "$(FC) $(FFLAGS) -c $<"
        .y.o:
            "$(YACC) $(YFLAGS) $<"
            "$(CC) $(CFLAGS) -c y.tab.c"
            "rm -f y.tab.c"
            "mv y.tab.o $@"
        .l.o:
            "$(LEX) $(LFLAGS) $<"
            "$(CC) $(CFLAGS) -c lex.yy.c"
            "rm -f lex.yy.c"
            "mv lex.yy.o $@"
        .y.c:
            "$(YACC) $(YFLAGS) $<"
            "mv y.tab.c $@"
        .l.c:
            "$(LEX) $(LFLAGS) $<"
            "mv lex.yy.c $@"
        .c.a:
            "$(CC) -c $(CFLAGS) $<"
            "$(AR) $(ARFLAGS) $@ $*.o"
            "rm -f $*.o"
        .f.a:
            "$(FC) -c $(FFLAGS) $<"
            "$(AR) $(ARFLAGS) $@ $*.o"
            "rm -f $*.o"
    }
}
fn builtin_targets() -> Vec<Target> {
    // even i'm not going to do that just for this
    vec![Target {
        name: ".SUFFIXES".into(),
        prerequisites: vec![".o", ".c", ".y", ".l", ".a", ".sh", ".f"]
            .into_iter()
            .map(String::from)
            .collect(),
        commands: vec![],
        already_updated: Cell::new(false),
    }]
}

#[cfg(test)]
mod test {
    use std::io::Cursor;

    use super::*;

    fn empty_makefile(args: &Args) -> Makefile {
        Makefile {
            inference_rules: vec![],
            macros: MacroSet::new(),
            targets: RefCell::new(HashMap::new()),
            first_non_special_target: None,
            args,
        }
    }

    #[test]
    fn basic_conditionals() {
        let file = "
ifeq (1,1)
worked = yes
else ifeq (2,2)
worked = no
endif
        ";
        let args = Args::empty();
        let mut makefile = empty_makefile(&args);
        makefile.and_read(Cursor::new(file));
        assert_eq!(
            makefile.expand_macros(&TokenString::r#macro("worked"), None),
            "yes"
        );
    }

    #[test]
    #[ignore = "I still haven't implemented `eval` or `define` or %-based macro substitution."]
    fn eval() {
        // This, for the record, is a terrible misfeature.
        // If you need this, you probably shouldn't be using Make.
        // But a lot of people are using this and still use Make anyway, so here we go,
        // I guess.

        let file = "
PROGRAMS    = server client

server_OBJS = server.o server_priv.o server_access.o
server_LIBS = priv protocol

client_OBJS = client.o client_api.o client_mem.o
client_LIBS = protocol

# Everything after this is generic

.PHONY: all
all: $(PROGRAMS)

define PROGRAM_template =
 $(1): $$($(1)_OBJS) $$($(1)_LIBS:%=-l%)
 ALL_OBJS   += $$($(1)_OBJS)
endef

$(foreach prog,$(PROGRAMS),$(eval $(call PROGRAM_template,$(prog))))

$(PROGRAMS):
        $(LINK.o) $^ $(LDLIBS) -o $@

clean:
        rm -f $(ALL_OBJS) $(PROGRAMS)
        ";

        let args = Args::empty();
        let mut makefile = empty_makefile(&args);
        makefile.and_read(Cursor::new(file));
        assert!(makefile.targets.borrow().contains_key("server"))
    }
}