Fix generate-copyright tool.

LLVM copyrights are now condensed to those reported in the .reuse/dep5 file.

Author: jonathanpallant

Diff for: .reuse/dep5

@@ -1,8 +1,6 @@
 # WARNING: this metadata is currently incomplete, do not rely on it yet.
 
 Format: https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
-Files-Excluded:
- src/llvm-project
 
 # Note that we're explicitly listing the individual files at the root of the
 # repository rather than just having `Files: *`. This is explicitly done to
@@ -39,13 +37,8 @@ Files: compiler/*
 Copyright: The Rust Project Developers (see https://thanks.rust-lang.org)
 License: MIT or Apache-2.0
 
-Files: compiler/rustc_codegen_cranelift/src/cranelift_native.rs
-Copyright: The Cranelift Project Developers
-           The Rust Project Developers (see https://thanks.rust-lang.org)
-License: Apache-2.0 WITH LLVM-exception AND (Apache-2.0 OR MIT)
-
 Files: compiler/rustc_llvm/llvm-wrapper/SymbolWrapper.cpp
-Copyright: LLVM authors
+Copyright: 2003-2019 University of Illinois at Urbana-Champaign.
            The Rust Project Developers (see https://thanks.rust-lang.org)
 License: Apache-2.0 WITH LLVM-exception AND (Apache-2.0 OR MIT)
 
@@ -85,19 +78,38 @@ Files: src/librustdoc/html/static/css/normalize.css
 Copyright: Nicolas Gallagher and Jonathan Neal
 License: MIT
 
-Files: src/librustdoc/html/static/css/themes/ayu.css
-Copyright: Ike Ku, Jessica Stokes, Leon Guan
+Files: src/librustdoc/html/static/css/rustdoc.css
+Copyright: 2016 Ike Ku
            The Rust Project Developers (see https://thanks.rust-lang.org)
 License: MIT OR Apache-2.0
 
 Files: src/doc/rustc-dev-guide/mermaid.min.js
-Copyright: Knut Sveidqvist
+Copyright: 2014-2021 Knut Sveidqvist
 License: MIT
 
-# Note that the LLVM submodule here only specifies the NCSA as the license for
-# LLVM, even though LLVM is in the process of a relicensing to Apache 2.0 with
-# the LLVM Exception. That's because relicensed files have a SPDX header with
-# the correct license, so it will automatically be included here.
+Files: library/backtrace/*
+Copyright: 2014 Alex Crichton
+           The Rust Project Developers (see https://thanks.rust-lang.org)
+License: MIT OR Apache-2.0
+
+Files: src/doc/embedded-book/*
+Copyright: Rust on Embedded Devices Working Group
+           The Rust Project Developers (see https://thanks.rust-lang.org)
+License: MIT OR Apache-2.0 OR CC-BY-SA-4.0
+
+Files: src/doc/rust-by-example/*
+Copyright: 2014 Jorge Aparicio
+           The Rust Project Developers (see https://thanks.rust-lang.org)
+License: MIT OR Apache-2.0
+
+# Reuse cannot process the LLVM source tree, and so the copyrights for the LLVM
+# submodule are written out here manually. The collect-licence-metadata tool
+# has a specific exception coded within it to ignore ./src/llvm-project so
+# any time LLVM is updated, please revisit this section. The copyrights are
+# taken from the relevant LLVM sub-folders: llvm, lld, lldb, compiler-rt and libunwind.
+
 Files: src/llvm-project/*
-Copyright: The LLVM Authors
-License: NCSA
+Copyright: 2003-2019 by the contributors listed in [CREDITS.TXT](https://github.com/rust-lang/llvm-project/blob/7738295178045041669876bf32b0543ec8319a5c/llvm/CREDITS.TXT)
+           2010 Apple Inc
+           2003-2019 University of Illinois at Urbana-Champaign.
+License: NCSA OR Apache-2.0 WITH LLVM-exception

Diff for: src/tools/collect-license-metadata/src/main.rs

@@ -6,6 +6,16 @@ use crate::licenses::LicensesInterner;
 use anyhow::Error;
 use std::path::PathBuf;
 
+// Some directories have too many slight license differences that'd result in a
+// huge report, and could be considered a standalone project anyway. Those
+// directories are "condensed" into a single licensing block for ease of
+// reading, merging the licensing information.
+//
+// For every `(dir, file)``, every file in `dir` is considered to have the
+// license info of `file`.
+const CONDENSED_DIRECTORIES: &[(&str, &str)] =
+    &[("./src/llvm-project/", "./src/llvm-project/README.md")];
+
 fn main() -> Result<(), Error> {
     let reuse_exe: PathBuf = std::env::var_os("REUSE_EXE").expect("Missing REUSE_EXE").into();
     let dest: PathBuf = std::env::var_os("DEST").expect("Missing DEST").into();

Diff for: src/tools/collect-license-metadata/src/path_tree.rs

@@ -7,11 +7,6 @@ use crate::licenses::{License, LicenseId, LicensesInterner};
 use std::collections::{BTreeMap, BTreeSet};
 use std::path::{Path, PathBuf};
 
-// Some directories have too many slight license differences that'd result in a huge report, and
-// could be considered a standalone project anyway. Those directories are "condensed" into a single
-// licensing block for ease of reading, merging the licensing information.
-const CONDENSED_DIRECTORIED: &[&str] = &["./src/llvm-project/"];
-
 #[derive(serde::Serialize)]
 #[serde(rename_all = "kebab-case", tag = "type")]
 pub(crate) enum Node<L> {
@@ -310,12 +305,17 @@ pub(crate) fn build(mut input: Vec<(PathBuf, LicenseId)>) -> Node<LicenseId> {
     let mut condensed_directories = BTreeMap::new();
     'outer: for (path, license) in input {
         // Files in condensed directories are handled separately.
-        for directory in CONDENSED_DIRECTORIED {
-            if path.starts_with(directory) {
-                condensed_directories
-                    .entry(*directory)
-                    .or_insert_with(BTreeSet::new)
-                    .insert(license);
+        for (condensed_directory, allowed_file) in super::CONDENSED_DIRECTORIES {
+            if path.starts_with(condensed_directory) {
+                if path.as_path() == Path::new(allowed_file) {
+                    // The licence on our allowed file is used to represent the entire directory
+                    condensed_directories
+                        .entry(*condensed_directory)
+                        .or_insert_with(BTreeSet::new)
+                        .insert(license);
+                } else {
+                    // don't add the file
+                }
                 continue 'outer;
             }
         }
@@ -338,9 +338,9 @@ pub(crate) fn build(mut input: Vec<(PathBuf, LicenseId)>) -> Node<LicenseId> {
             name: path.file_name().unwrap().into(),
             licenses: licenses.iter().copied().collect(),
         };
-        for name in path.parent().unwrap_or_else(|| Path::new(".")).components().rev() {
+        for component in path.parent().unwrap_or_else(|| Path::new(".")).components().rev() {
             node = Node::Directory {
-                name: name.as_os_str().into(),
+                name: component.as_os_str().into(),
                 children: vec![node],
                 license: None,
             };

Diff for: src/tools/generate-copyright/src/main.rs

@@ -27,7 +27,7 @@ fn render_recursive(node: &Node, buffer: &mut Vec<u8>, depth: usize) -> Result<(
             }
         }
         Node::Directory { name, children, license } => {
-            render_license(&prefix, std::iter::once(name), std::iter::once(license), buffer)?;
+            render_license(&prefix, std::iter::once(name), license.iter(), buffer)?;
             if !children.is_empty() {
                 writeln!(buffer, "{prefix}")?;
                 writeln!(buffer, "{prefix}*Exceptions:*")?;
@@ -94,7 +94,7 @@ struct Metadata {
 #[serde(rename_all = "kebab-case", tag = "type")]
 pub(crate) enum Node {
     Root { children: Vec<Node> },
-    Directory { name: String, children: Vec<Node>, license: License },
+    Directory { name: String, children: Vec<Node>, license: Option<License> },
     CondensedDirectory { name: String, licenses: Vec<License> },
     File { name: String, license: License },
     Group { files: Vec<String>, directories: Vec<String>, license: License },