Auto merge of #97485 - bjorn3:new_archive_writer, r=wesleywiser

Rewrite LLVM's archive writer in Rust

This allows it to be used by other codegen backends.

Fixes https://github.com/bjorn3/rustc_codegen_cranelift/issues/1155

3 files changed, 5 insertions, 226 deletions

diff --git a/compiler/rustc_codegen_cranelift/Cargo.lock b/compiler/rustc_codegen_cranelift/Cargo.lock
index 3fa9d56cd01..3b406036c35 100644
--- a/compiler/rustc_codegen_cranelift/Cargo.lock
+++ b/compiler/rustc_codegen_cranelift/Cargo.lock
@@ -20,11 +20,6 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c794e162a5eff65c72ef524dfe393eb923c354e350bb78b9c7383df13f3bc142"
 
 [[package]]
-name = "ar"
-version = "0.8.0"
-source = "git+https://github.com/bjorn3/rust-ar.git?branch=do_not_remove_cg_clif_ranlib#de9ab0e56bf3a208381d342aa5b60f9ff2891648"
-
-[[package]]
 name = "arrayvec"
 version = "0.7.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
@@ -324,7 +319,6 @@ dependencies = [
 name = "rustc_codegen_cranelift"
 version = "0.1.0"
 dependencies = [
- "ar",
  "cranelift-codegen",
  "cranelift-frontend",
  "cranelift-jit",
diff --git a/compiler/rustc_codegen_cranelift/Cargo.toml b/compiler/rustc_codegen_cranelift/Cargo.toml
index 09cf5b4a1ed..0fdd5de118c 100644
--- a/compiler/rustc_codegen_cranelift/Cargo.toml
+++ b/compiler/rustc_codegen_cranelift/Cargo.toml
@@ -18,7 +18,6 @@ target-lexicon = "0.12.0"
 gimli = { version = "0.26.0", default-features = false, features = ["write"]}
 object = { version = "0.29.0", default-features = false, features = ["std", "read_core", "write", "archive", "coff", "elf", "macho", "pe"] }
 
-ar = { git = "https://github.com/bjorn3/rust-ar.git", branch = "do_not_remove_cg_clif_ranlib" }
 indexmap = "1.9.1"
 libloading = { version = "0.7.3", optional = true }
 once_cell = "1.10.0"
diff --git a/compiler/rustc_codegen_cranelift/src/archive.rs b/compiler/rustc_codegen_cranelift/src/archive.rs
index f2e3bf16e61..5a29bc18def 100644
--- a/compiler/rustc_codegen_cranelift/src/archive.rs
+++ b/compiler/rustc_codegen_cranelift/src/archive.rs
@@ -1,35 +1,15 @@
-//! Creation of ar archives like for the lib and staticlib crate type
-
-use std::collections::BTreeMap;
-use std::fs::File;
-use std::io::{self, Read, Seek};
 use std::path::{Path, PathBuf};
 
-use rustc_codegen_ssa::back::archive::{ArchiveBuilder, ArchiveBuilderBuilder};
+use rustc_codegen_ssa::back::archive::{
+    get_native_object_symbols, ArArchiveBuilder, ArchiveBuilder, ArchiveBuilderBuilder,
+};
 use rustc_session::Session;
 
-use object::read::archive::ArchiveFile;
-use object::{Object, ObjectSymbol, ReadCache};
-
-#[derive(Debug)]
-enum ArchiveEntry {
-    FromArchive { archive_index: usize, file_range: (u64, u64) },
-    File(PathBuf),
-}
-
 pub(crate) struct ArArchiveBuilderBuilder;
 
 impl ArchiveBuilderBuilder for ArArchiveBuilderBuilder {
     fn new_archive_builder<'a>(&self, sess: &'a Session) -> Box<dyn ArchiveBuilder<'a> + 'a> {
-        Box::new(ArArchiveBuilder {
-            sess,
-            use_gnu_style_archive: sess.target.archive_format == "gnu",
-            // FIXME fix builtin ranlib on macOS
-            no_builtin_ranlib: sess.target.is_like_osx,
-
-            src_archives: vec![],
-            entries: vec![],
-        })
+        Box::new(ArArchiveBuilder::new(sess, get_native_object_symbols))
     }
 
     fn create_dll_import_lib(
@@ -40,200 +20,6 @@ impl ArchiveBuilderBuilder for ArArchiveBuilderBuilder {
         _tmpdir: &Path,
         _is_direct_dependency: bool,
     ) -> PathBuf {
-        bug!("creating dll imports is not supported");
-    }
-}
-
-pub(crate) struct ArArchiveBuilder<'a> {
-    sess: &'a Session,
-    use_gnu_style_archive: bool,
-    no_builtin_ranlib: bool,
-
-    src_archives: Vec<File>,
-    // Don't use `HashMap` here, as the order is important. `rust.metadata.bin` must always be at
-    // the end of an archive for linkers to not get confused.
-    entries: Vec<(Vec<u8>, ArchiveEntry)>,
-}
-
-impl<'a> ArchiveBuilder<'a> for ArArchiveBuilder<'a> {
-    fn add_file(&mut self, file: &Path) {
-        self.entries.push((
-            file.file_name().unwrap().to_str().unwrap().to_string().into_bytes(),
-            ArchiveEntry::File(file.to_owned()),
-        ));
-    }
-
-    fn add_archive(
-        &mut self,
-        archive_path: &Path,
-        mut skip: Box<dyn FnMut(&str) -> bool + 'static>,
-    ) -> std::io::Result<()> {
-        let read_cache = ReadCache::new(std::fs::File::open(&archive_path)?);
-        let archive = ArchiveFile::parse(&read_cache).unwrap();
-        let archive_index = self.src_archives.len();
-
-        for entry in archive.members() {
-            let entry = entry.map_err(|err| io::Error::new(io::ErrorKind::InvalidData, err))?;
-            let file_name = String::from_utf8(entry.name().to_vec())
-                .map_err(|err| io::Error::new(io::ErrorKind::InvalidData, err))?;
-            if !skip(&file_name) {
-                self.entries.push((
-                    file_name.into_bytes(),
-                    ArchiveEntry::FromArchive { archive_index, file_range: entry.file_range() },
-                ));
-            }
-        }
-
-        self.src_archives.push(read_cache.into_inner());
-        Ok(())
-    }
-
-    fn build(mut self: Box<Self>, output: &Path) -> bool {
-        enum BuilderKind {
-            Bsd(ar::Builder<File>),
-            Gnu(ar::GnuBuilder<File>),
-        }
-
-        let sess = self.sess;
-
-        let mut symbol_table = BTreeMap::new();
-
-        let mut entries = Vec::new();
-
-        for (mut entry_name, entry) in self.entries {
-            // FIXME only read the symbol table of the object files to avoid having to keep all
-            // object files in memory at once, or read them twice.
-            let data = match entry {
-                ArchiveEntry::FromArchive { archive_index, file_range } => {
-                    // FIXME read symbols from symtab
-                    let src_read_cache = &mut self.src_archives[archive_index];
-
-                    src_read_cache.seek(io::SeekFrom::Start(file_range.0)).unwrap();
-                    let mut data = std::vec::from_elem(0, usize::try_from(file_range.1).unwrap());
-                    src_read_cache.read_exact(&mut data).unwrap();
-
-                    data
-                }
-                ArchiveEntry::File(file) => std::fs::read(file).unwrap_or_else(|err| {
-                    sess.fatal(&format!(
-                        "error while reading object file during archive building: {}",
-                        err
-                    ));
-                }),
-            };
-
-            if !self.no_builtin_ranlib {
-                if symbol_table.contains_key(&entry_name) {
-                    // The ar crate can't handle creating a symbol table in case of multiple archive
-                    // members with the same name. Work around this by prepending a number until we
-                    // get a unique name.
-                    for i in 1.. {
-                        let new_name = format!("{}_", i)
-                            .into_bytes()
-                            .into_iter()
-                            .chain(entry_name.iter().copied())
-                            .collect::<Vec<_>>();
-                        if !symbol_table.contains_key(&new_name) {
-                            entry_name = new_name;
-                            break;
-                        }
-                    }
-                }
-
-                match object::File::parse(&*data) {
-                    Ok(object) => {
-                        symbol_table.insert(
-                            entry_name.to_vec(),
-                            object
-                                .symbols()
-                                .filter_map(|symbol| {
-                                    if symbol.is_undefined() || symbol.is_local() {
-                                        None
-                                    } else {
-                                        symbol.name().map(|name| name.as_bytes().to_vec()).ok()
-                                    }
-                                })
-                                .collect::<Vec<_>>(),
-                        );
-                    }
-                    Err(err) => {
-                        let err = err.to_string();
-                        if err == "Unknown file magic" {
-                            // Not an object file; skip it.
-                        } else if object::read::archive::ArchiveFile::parse(&*data).is_ok() {
-                            // Nested archive file; skip it.
-                        } else {
-                            sess.fatal(&format!(
-                                "error parsing `{}` during archive creation: {}",
-                                String::from_utf8_lossy(&entry_name),
-                                err
-                            ));
-                        }
-                    }
-                }
-            }
-
-            entries.push((entry_name, data));
-        }
-
-        let mut builder = if self.use_gnu_style_archive {
-            BuilderKind::Gnu(
-                ar::GnuBuilder::new(
-                    File::create(output).unwrap_or_else(|err| {
-                        sess.fatal(&format!(
-                            "error opening destination during archive building: {}",
-                            err
-                        ));
-                    }),
-                    entries.iter().map(|(name, _)| name.clone()).collect(),
-                    ar::GnuSymbolTableFormat::Size32,
-                    symbol_table,
-                )
-                .unwrap(),
-            )
-        } else {
-            BuilderKind::Bsd(
-                ar::Builder::new(
-                    File::create(output).unwrap_or_else(|err| {
-                        sess.fatal(&format!(
-                            "error opening destination during archive building: {}",
-                            err
-                        ));
-                    }),
-                    symbol_table,
-                )
-                .unwrap(),
-            )
-        };
-
-        let any_members = !entries.is_empty();
-
-        // Add all files
-        for (entry_name, data) in entries.into_iter() {
-            let header = ar::Header::new(entry_name, data.len() as u64);
-            match builder {
-                BuilderKind::Bsd(ref mut builder) => builder.append(&header, &mut &*data).unwrap(),
-                BuilderKind::Gnu(ref mut builder) => builder.append(&header, &mut &*data).unwrap(),
-            }
-        }
-
-        // Finalize archive
-        std::mem::drop(builder);
-
-        if self.no_builtin_ranlib {
-            let ranlib = crate::toolchain::get_toolchain_binary(self.sess, "ranlib");
-
-            // Run ranlib to be able to link the archive
-            let status = std::process::Command::new(ranlib)
-                .arg(output)
-                .status()
-                .expect("Couldn't run ranlib");
-
-            if !status.success() {
-                self.sess.fatal(&format!("Ranlib exited with code {:?}", status.code()));
-            }
-        }
-
-        any_members
+        unimplemented!("creating dll imports is not yet supported");
     }
 }