refactor makefile reading into a separate module

1 files changed, 648 insertions, 0 deletions

diff --git a/src/makefile/input.rs b/src/makefile/input.rs
new file mode 100644
index 0000000..fda81d0
--- /dev/null
+++ b/src/makefile/input.rs
@@ -0,0 +1,648 @@
+use std::cell::{Cell, RefCell};
+use std::collections::HashMap;
+use std::fs::File;
+use std::io::{BufRead, BufReader};
+use std::iter::Peekable;
+use std::path::Path;
+use std::rc::Rc;
+
+use eyre::{bail, eyre, Context, Result};
+use lazy_static::lazy_static;
+use regex::Regex;
+
+use crate::args::Args;
+
+use super::command_line::CommandLine;
+#[cfg(feature = "full")]
+use super::conditional::{Line as ConditionalLine, State as ConditionalState};
+use super::inference_rules::InferenceRule;
+use super::r#macro::{Set as MacroSet, Source as MacroSource};
+use super::target::Target;
+use super::token::{tokenize, Token, TokenString};
+
+enum LineType {
+    Rule,
+    Macro,
+    Unknown,
+}
+
+impl LineType {
+    fn of(line_tokens: &TokenString) -> Self {
+        #[cfg(feature = "full")]
+        if line_tokens.starts_with("define ") {
+            return Self::Macro;
+        }
+        for token in line_tokens.tokens() {
+            if let Token::Text(text) = token {
+                let colon_idx = text.find(':');
+                #[cfg(not(feature = "full"))]
+                let equals_idx = text.find('=');
+                #[cfg(feature = "full")]
+                let equals_idx = ["=", ":=", "::=", "?=", "+="]
+                    .iter()
+                    .filter_map(|p| text.find(p))
+                    .min();
+                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 struct MakefileReader<'a> {
+    inference_rules: Vec<InferenceRule>,
+    macros: MacroSet<'static, 'static>,
+    targets: RefCell<HashMap<String, Rc<RefCell<Target>>>>,
+    pub first_non_special_target: Option<String>,
+    args: &'a Args,
+    // TODO borrow warnings from Python version
+}
+
+impl<'a> MakefileReader<'a> {
+    pub 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),
+                );
+            }
+        }
+
+        MakefileReader {
+            inference_rules,
+            macros,
+            targets: RefCell::new(targets),
+            first_non_special_target,
+            args,
+        }
+    }
+
+    pub fn and_read_file(&mut self, path: impl AsRef<Path>) -> Result<()> {
+        let file = File::open(path);
+        // TODO handle errors
+        let file = file.context("couldn't open makefile!")?;
+        let file_reader = BufReader::new(file);
+        self.and_read(file_reader)
+    }
+
+    pub fn and_read(&mut self, source: impl BufRead) -> Result<()> {
+        let mut lines_iter = source
+            .lines()
+            .enumerate()
+            .map(|(number, line)| (number.saturating_add(1), line))
+            .map(|(line, x)| {
+                (
+                    line,
+                    x.with_context(|| format!("failed to read line {} of makefile", line)),
+                )
+            })
+            .peekable();
+        #[cfg(feature = "full")]
+        let mut conditional_stack: Vec<ConditionalState> = vec![];
+        while let Some((line_number, line)) = lines_iter.next() {
+            let mut line = line?;
+
+            // handle escaped newlines
+            while line.ends_with('\\') {
+                line.pop();
+                line.push(' ');
+                if let Some((_, x)) = lines_iter.next() {
+                    line.push_str(x?.trim_start())
+                }
+            }
+
+            // handle comments
+            lazy_static! {
+                static ref COMMENT: Regex = Regex::new("#.*$").unwrap();
+            }
+            let line = COMMENT.replace(&line, "").into_owned();
+
+            #[cfg(feature = "full")]
+            if let Some(line) = ConditionalLine::from(&line, |t| self.expand_macros(t))? {
+                line.action(
+                    conditional_stack.last(),
+                    |name| self.macros.is_defined(name),
+                    |t| self.expand_macros(t),
+                )?
+                .apply_to(&mut conditional_stack);
+                continue;
+            }
+
+            // skip lines if we need to
+            #[cfg(feature = "full")]
+            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)?)?;
+                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)?;
+                }
+                continue;
+            }
+
+            if line.trim().is_empty() {
+                // handle blank lines
+                continue;
+            }
+            // unfortunately, rules vs macros can't be determined until after
+            // macro tokenizing. so that's suboptimal.
+
+            // TODO errors
+            let line_tokens: TokenString = line
+                .parse()
+                .with_context(|| format!("failed to parse line {}", line_number))?;
+
+            let line_type = LineType::of(&line_tokens);
+
+            // before we actually test it, see if it's only visible after expanding macros
+            let (line_tokens, line_type) = if let LineType::Unknown = line_type {
+                let line_tokens = TokenString::text(self.expand_macros(&line_tokens)?);
+                let line_type = LineType::of(&line_tokens);
+                (line_tokens, line_type)
+            } else {
+                (line_tokens, line_type)
+            };
+
+            match line_type {
+                LineType::Rule => self.read_rule(&line_tokens, line_number, &mut lines_iter)?,
+                LineType::Macro => self.read_macro(line_tokens, line_number, &mut lines_iter)?,
+                LineType::Unknown => {
+                    if !line_tokens.is_empty() {
+                        bail!(
+                            "error: line {}: unknown line \"{}\"",
+                            line_number,
+                            line_tokens
+                        );
+                    }
+                }
+            }
+        }
+
+        Ok(())
+    }
+
+    fn read_rule(
+        &mut self,
+        line_tokens: &TokenString,
+        line_number: usize,
+        lines_iter: &mut Peekable<impl Iterator<Item = (usize, Result<String>)>>,
+    ) -> Result<()> {
+        let (targets, not_targets) = line_tokens
+            .split_once(':')
+            .ok_or_else(|| eyre!("read_rule couldn't find a ':' on line {}", line_number))?;
+        let targets = self.expand_macros(&targets)?;
+        let targets = targets.split_whitespace().collect::<Vec<_>>();
+        let (prerequisites, mut commands) = match not_targets.split_once(';') {
+            Some((prerequisites, mut command)) => {
+                while command.ends_with("\\") {
+                    if let Some((_, next_line)) = lines_iter.next() {
+                        command.strip_suffix("\\");
+                        command.extend(tokenize(&next_line?)?);
+                    } else {
+                        break;
+                    }
+                }
+                (prerequisites, vec![command])
+            }
+            None => (not_targets, vec![]),
+        };
+        let prerequisites = self.expand_macros(&prerequisites)?;
+        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?;
+            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()
+                    .with_context(|| format!("failed to parse line {}", line_number))?,
+            );
+        }
+
+        let commands = commands
+            .into_iter()
+            .map(CommandLine::from)
+            .collect::<Vec<_>>();
+
+        if targets.is_empty() {
+            return Ok(());
+        }
+
+        // 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_owned(),
+                prereq: inference_match.name("s2").unwrap().as_str().to_owned(),
+                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)));
+                    }
+                }
+            }
+        }
+
+        Ok(())
+    }
+
+    fn read_macro(
+        &mut self,
+        mut line_tokens: TokenString,
+        line_number: usize,
+        lines_iter: &mut Peekable<impl Iterator<Item = (usize, Result<String>)>>,
+    ) -> Result<()> {
+        let (name, mut value) = if cfg!(feature = "full") && line_tokens.starts_with("define ") {
+            line_tokens.strip_prefix("define ");
+            if line_tokens.ends_with("=") {
+                line_tokens.strip_suffix("=");
+                line_tokens.trim_end();
+            }
+            let mut value = TokenString::empty();
+            for (_, line) in lines_iter {
+                let line = line?;
+                if line == "endef" {
+                    break;
+                }
+                if !value.is_empty() {
+                    value.extend(TokenString::text("\n"));
+                }
+                value.extend(line.parse()?);
+            }
+            (line_tokens, value)
+        } else {
+            line_tokens
+                .split_once('=')
+                .ok_or_else(|| eyre!("read_rule couldn't find a ':' on line {}", line_number))?
+        };
+        let name = self.expand_macros(&name)?;
+        // GNUisms are annoying, but popular
+        let mut expand_value = false;
+        let mut skip_if_defined = false;
+        let mut append = false;
+
+        #[cfg(feature = "full")]
+        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)?)
+        } else {
+            value
+        };
+
+        match self.macros.get(name) {
+            // We always let command line or MAKEFLAGS macros override macros from the file.
+            Some((MacroSource::CommandLineOrMakeflags, _)) => return Ok(()),
+            // 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 Ok(()),
+            _ if skip_if_defined => return Ok(()),
+            _ => {}
+        }
+
+        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);
+
+        Ok(())
+    }
+
+    fn expand_macros(&self, text: &TokenString) -> Result<String> {
+        self.macros.expand(text)
+    }
+}
+
+impl<'a> From<MakefileReader<'a>> for super::Makefile<'a> {
+    fn from(reader: MakefileReader<'a>) -> Self {
+        Self {
+            inference_rules: reader.inference_rules,
+            macros: reader.macros,
+            targets: reader.targets,
+            first_non_special_target: reader.first_non_special_target,
+            args: reader.args,
+        }
+    }
+}
+
+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::*;
+
+    type R = Result<()>;
+
+    fn empty_makefile(args: &Args) -> MakefileReader {
+        MakefileReader {
+            inference_rules: vec![],
+            macros: MacroSet::new(),
+            targets: RefCell::new(HashMap::new()),
+            first_non_special_target: None,
+            args,
+        }
+    }
+
+    #[cfg(feature = "full")]
+    #[test]
+    fn basic_conditionals() -> R {
+        let file = "
+ifeq (1,1)
+worked = yes
+else ifeq (2,2)
+worked = no
+else 
+worked = perhaps
+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"))?,
+            "yes"
+        );
+        Ok(())
+    }
+
+    #[cfg(feature = "full")]
+    #[test]
+    fn define_syntax() -> R {
+        let file = "
+define foo =
+bar
+baz
+endef
+        ";
+        let args = Args::empty();
+        let mut makefile = empty_makefile(&args);
+        makefile.and_read(Cursor::new(file))?;
+        assert_eq!(
+            makefile.expand_macros(&TokenString::r#macro("foo"))?,
+            "bar\nbaz"
+        );
+        Ok(())
+    }
+
+    #[test]
+    #[ignore = "I still haven't implemented `eval` or %-based macro substitution."]
+    fn eval() -> R {
+        // 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"));
+        Ok(())
+    }
+}