Add API to Record Compile Invocation

This commit adds API to record the compilation invocations in order to
emit compile_commands.json a.k.a. JSON compilation database.

Fixes #497

Author: schrieveslaach

Cargo.toml

@@ -19,6 +19,7 @@ edition = "2018"
 
 [dependencies]
 jobserver = { version = "0.1.16", optional = true }
+tinyjson = "2.3.0"
 
 [features]
 parallel = ["jobserver"]

src/json_compilation_database.rs

@@ -0,0 +1,111 @@
+use std::collections::HashMap;
+use std::path::{Path, PathBuf};
+use std::process::Command;
+use tinyjson::JsonValue;
+
+/// An entry for creating a [JSON Compilation Database](https://clang.llvm.org/docs/JSONCompilationDatabase.html).
+pub struct CompileCommand {
+    directory: PathBuf,
+    arguments: Vec<String>,
+    file: PathBuf,
+    output: PathBuf,
+}
+
+impl CompileCommand {
+    pub(crate) fn new(cmd: &Command, src: PathBuf, output: PathBuf) -> Self {
+        let mut arguments = Vec::with_capacity(cmd.get_args().len() + 1);
+
+        let program = String::from(cmd.get_program().to_str().unwrap());
+        arguments.push(
+            crate::which(&program)
+                .map(|p| p.to_string_lossy().into_owned())
+                .map(|p| p.to_string())
+                .unwrap_or(program),
+        );
+        arguments.extend(
+            cmd.get_args()
+                .flat_map(std::ffi::OsStr::to_str)
+                .map(String::from),
+        );
+
+        Self {
+            // TODO: is the assumption correct?
+            directory: std::env::current_dir().unwrap(),
+            arguments,
+            file: src,
+            output,
+        }
+    }
+
+    /// The working directory of the compilation. All paths specified in the command or file fields
+    /// must be either absolute or relative to this directory.
+    pub fn directory(&self) -> &PathBuf {
+        &self.directory
+    }
+
+    /// The name of the output created by this compilation step. This field is optional. It can be
+    /// used to distinguish different processing modes of the same input file.
+    pub fn output(&self) -> &PathBuf {
+        &self.output
+    }
+
+    /// The main translation unit source processed by this compilation step. This is used by tools
+    /// as the key into the compilation database. There can be multiple command objects for the
+    /// same file, for example if the same source file is compiled with different configurations.
+    pub fn file(&self) -> &PathBuf {
+        &self.file
+    }
+
+    /// The compile command argv as list of strings. This should run the compilation step for the
+    /// translation unit file. arguments[0] should be the executable name, such as clang++.
+    /// Arguments should not be escaped, but ready to pass to execvp().
+    pub fn arguments(&self) -> &Vec<String> {
+        &self.arguments
+    }
+}
+
+/// Stores the provided list of [compile commands](crate::CompileCommand) as [JSON
+/// Compilation Database](https://clang.llvm.org/docs/JSONCompilationDatabase.html).
+pub fn store_json_compilation_database<'a, C, P>(commands: C, path: P)
+where
+    C: IntoIterator<Item = &'a CompileCommand>,
+    P: AsRef<Path>,
+{
+    let db = JsonValue::Array(
+        commands
+            .into_iter()
+            .map(|command| command.into())
+            .collect::<Vec<JsonValue>>(),
+    );
+
+    std::fs::write(path, db.stringify().unwrap()).unwrap();
+}
+
+impl<'a> std::convert::From<&CompileCommand> for JsonValue {
+    fn from(compile_command: &CompileCommand) -> Self {
+        JsonValue::Object(HashMap::from([
+            (
+                String::from("directory"),
+                JsonValue::String(compile_command.directory.to_string_lossy().to_string()),
+            ),
+            (
+                String::from("file"),
+                JsonValue::String(compile_command.file.to_string_lossy().to_string()),
+            ),
+            (
+                String::from("output"),
+                JsonValue::String(compile_command.output.to_string_lossy().to_string()),
+            ),
+            (
+                String::from("arguments"),
+                JsonValue::Array(
+                    compile_command
+                        .arguments
+                        .iter()
+                        .map(|arg| JsonValue::String(arg.to_string()))
+                        .collect::<Vec<_>>(),
+                ),
+            ),
+        ]))
+    }
+}

src/lib.rs

@@ -56,6 +56,8 @@
 #![allow(deprecated)]
 #![deny(missing_docs)]
 
+pub use crate::json_compilation_database::store_json_compilation_database;
+pub use crate::json_compilation_database::CompileCommand;
 use std::collections::HashMap;
 use std::env;
 use std::ffi::{OsStr, OsString};
@@ -81,6 +83,7 @@ mod setup_config;
 #[cfg(windows)]
 mod vs_instances;
 
+mod json_compilation_database;
 pub mod windows_registry;
 
 /// A builder for compilation of a native library.
@@ -943,8 +946,17 @@ impl Build {
 
     /// Run the compiler, generating the file `output`
     ///
-    /// This will return a result instead of panicing; see compile() for the complete description.
+    /// This will return a result instead of panicing; see [compile()](Build::compile) for the complete description.
     pub fn try_compile(&self, output: &str) -> Result<(), Error> {
+        self.try_recorded_compile(output)?;
+        Ok(())
+    }
+
+    /// Run the compiler, generating the file `output` and provides compile commands for creating
+    /// [JSON Compilation Database](https://clang.llvm.org/docs/JSONCompilationDatabase.html).
+    ///
+    /// This will return a result instead of panicing; see [recorded_compile()](Build::recorded_compile) for the complete description.
+    pub fn try_recorded_compile(&self, output: &str) -> Result<Vec<CompileCommand>, Error> {
         let mut output_components = Path::new(output).components();
         match (output_components.next(), output_components.next()) {
             (Some(Component::Normal(_)), None) => {}
@@ -990,7 +1002,7 @@ impl Build {
 
             objects.push(Object::new(file.to_path_buf(), obj));
         }
-        self.compile_objects(&objects)?;
+        let entries = self.compile_objects(&objects)?;
         self.assemble(lib_name, &dst.join(gnu_lib_name), &objects)?;
 
         if self.get_target()?.contains("msvc") {
@@ -1074,7 +1086,7 @@ impl Build {
             }
         }
 
-        Ok(())
+        Ok(entries)
     }
 
     /// Run the compiler, generating the file `output`
@@ -1120,6 +1132,26 @@ impl Build {
         }
     }
 
+    /// Run the compiler, generating the file `output` and provides compile commands for creating
+    /// [JSON Compilation Database](https://clang.llvm.org/docs/JSONCompilationDatabase.html),
+    ///
+    /// ```no_run
+    /// let compile_commands = cc::Build::new().file("blobstore.c")
+    ///    .recorded_compile("blobstore");
+    ///
+    /// cc::store_json_compilation_database(&compile_commands, "target/compilation_database.json");
+    /// ```
+    ///
+    /// See [compile()](Build::compile) for the further description.
+    pub fn recorded_compile(&self, output: &str) -> Vec<CompileCommand> {
+        match self.try_recorded_compile(output) {
+            Ok(entries) => entries,
+            Err(e) => {
+                fail(&e.message);
+            }
+        }
+    }
+
     #[cfg(feature = "parallel")]
     fn compile_objects<'me>(&'me self, objs: &[Object]) -> Result<(), Error> {
         use std::sync::atomic::{AtomicBool, Ordering::SeqCst};
@@ -1272,14 +1304,15 @@ impl Build {
     }
 
     #[cfg(not(feature = "parallel"))]
-    fn compile_objects(&self, objs: &[Object]) -> Result<(), Error> {
+    fn compile_objects(&self, objs: &[Object]) -> Result<Vec<CompileCommand>, Error> {
+        let mut entries = Vec::new();
         for obj in objs {
-            self.compile_object(obj)?;
+            entries.push(self.compile_object(obj)?);
         }
-        Ok(())
+        Ok(entries)
     }
 
-    fn compile_object(&self, obj: &Object) -> Result<(), Error> {
+    fn compile_object(&self, obj: &Object) -> Result<CompileCommand, Error> {
         let is_asm = obj.src.extension().and_then(|s| s.to_str()) == Some("asm");
         let target = self.get_target()?;
         let msvc = target.contains("msvc");
@@ -1324,7 +1357,7 @@ impl Build {
         }
 
         run(&mut cmd, &name)?;
-        Ok(())
+        Ok(CompileCommand::new(&cmd, obj.src.clone(), obj.dst.clone()))
     }
 
     /// This will return a result instead of panicing; see expand() for the complete description.
@@ -3287,22 +3320,29 @@ fn map_darwin_target_from_rust_to_compiler_architecture(target: &str) -> Option<
     }
 }
 
-fn which(tool: &Path) -> Option<PathBuf> {
+pub(crate) fn which<P>(tool: P) -> Option<PathBuf>
+where
+    P: AsRef<Path>,
+{
     fn check_exe(exe: &mut PathBuf) -> bool {
         let exe_ext = std::env::consts::EXE_EXTENSION;
         exe.exists() || (!exe_ext.is_empty() && exe.set_extension(exe_ext) && exe.exists())
     }
 
-    // If |tool| is not just one "word," assume it's an actual path...
-    if tool.components().count() > 1 {
-        let mut exe = PathBuf::from(tool);
-        return if check_exe(&mut exe) { Some(exe) } else { None };
+    fn non_generic_which(tool: &Path) -> Option<PathBuf> {
+        // If |tool| is not just one "word," assume it's an actual path...
+        if tool.components().count() > 1 {
+            let mut exe = PathBuf::from(tool);
+            return if check_exe(&mut exe) { Some(exe) } else { None };
+        }
+
+        // Loop through PATH entries searching for the |tool|.
+        let path_entries = env::var_os("PATH")?;
+        env::split_paths(&path_entries).find_map(|path_entry| {
+            let mut exe = path_entry.join(tool);
+            return if check_exe(&mut exe) { Some(exe) } else { None };
+        })
     }
 
-    // Loop through PATH entries searching for the |tool|.
-    let path_entries = env::var_os("PATH")?;
-    env::split_paths(&path_entries).find_map(|path_entry| {
-        let mut exe = path_entry.join(tool);
-        return if check_exe(&mut exe) { Some(exe) } else { None };
-    })
+    non_generic_which(tool.as_ref())
 }