Rollup merge of #68791 - mark-i-m:proper-linkcheck, r=ehuss,JohnTitor

implement proper linkchecker hardening

r? @JohnTitor

This implements proper linkcheck filtering... we might need to fiddle with a bit to adjust what is or isn't filtered, but this seems to work reasonable locally.

Cargo.lock

@@ -3067,6 +3067,7 @@ name = "rustbook"
 version = "0.1.0"
 dependencies = [
  "clap",
+ "codespan",
  "codespan-reporting",
  "failure",
  "mdbook",

src/doc/rustc-guide

@@ -1 +1 @@
-Subproject commit 92baf7293dd2d418d2ac4b141b0faa822075d9f7
+Subproject commit 5bd60bc51efaec04e69e2e18b59678e2af066433

src/tools/rustbook/Cargo.toml

@@ -6,13 +6,14 @@ license = "MIT OR Apache-2.0"
 edition = "2018"
 
 [features]
-linkcheck = ["mdbook-linkcheck", "codespan-reporting"]
+linkcheck = ["mdbook-linkcheck", "codespan-reporting", "codespan"]
 
 [dependencies]
 clap = "2.25.0"
 failure = "0.1"
 mdbook-linkcheck = { version = "0.5.0", optional = true }
 # Keep in sync with mdbook-linkcheck.
+codespan = { version = "0.5", optional = true }
 codespan-reporting = { version = "0.5", optional = true }
 
 

src/tools/rustbook/src/main.rs

@@ -8,11 +8,6 @@ use clap::{App, AppSettings, ArgMatches, SubCommand};
 use mdbook::errors::Result as Result3;
 use mdbook::MDBook;
 
-#[cfg(feature = "linkcheck")]
-use failure::Error;
-#[cfg(feature = "linkcheck")]
-use mdbook::renderer::RenderContext;
-
 fn main() {
     let d_message = "-d, --dest-dir=[dest-dir]
 'The output directory for your book{n}(Defaults to ./book when omitted)'";
@@ -53,8 +48,18 @@ fn main() {
         ("linkcheck", Some(sub_matches)) => {
             #[cfg(feature = "linkcheck")]
             {
-                if let Err(err) = linkcheck(sub_matches) {
-                    eprintln!("Error: {}", err);
+                let (diags, files) = linkcheck(sub_matches).expect("Error while linkchecking.");
+                if !diags.is_empty() {
+                    let color = codespan_reporting::term::termcolor::ColorChoice::Auto;
+                    let mut writer =
+                        codespan_reporting::term::termcolor::StandardStream::stderr(color);
+                    let cfg = codespan_reporting::term::Config::default();
+
+                    for diag in diags {
+                        codespan_reporting::term::emit(&mut writer, &cfg, &files, &diag)
+                            .expect("Unable to emit linkcheck error.");
+                    }
+
                     std::process::exit(101);
                 }
             }
@@ -73,14 +78,55 @@ fn main() {
 }
 
 #[cfg(feature = "linkcheck")]
-pub fn linkcheck(args: &ArgMatches<'_>) -> Result<(), Error> {
+pub fn linkcheck(
+    args: &ArgMatches<'_>,
+) -> Result<(Vec<codespan_reporting::diagnostic::Diagnostic>, codespan::Files), failure::Error> {
+    use mdbook_linkcheck::Reason;
+
     let book_dir = get_book_dir(args);
+    let src_dir = book_dir.join("src");
     let book = MDBook::load(&book_dir).unwrap();
-    let cfg = book.config;
-    let render_ctx = RenderContext::new(&book_dir, book.book, cfg, &book_dir);
-    let cache_file = render_ctx.destination.join("cache.json");
-    let color = codespan_reporting::term::termcolor::ColorChoice::Auto;
-    mdbook_linkcheck::run(&cache_file, color, &render_ctx)
+    let linkck_cfg = mdbook_linkcheck::get_config(&book.config)?;
+    let mut files = codespan::Files::new();
+    let target_files = mdbook_linkcheck::load_files_into_memory(&book.book, &mut files);
+    let cache = mdbook_linkcheck::Cache::default();
+
+    let (links, incomplete) = mdbook_linkcheck::extract_links(target_files, &files);
+
+    let outcome =
+        mdbook_linkcheck::validate(&links, &linkck_cfg, &src_dir, &cache, &files, incomplete)?;
+
+    let mut is_real_error = false;
+
+    for link in outcome.invalid_links.iter() {
+        match &link.reason {
+            Reason::FileNotFound | Reason::TraversesParentDirectories => {
+                is_real_error = true;
+            }
+            Reason::UnsuccessfulServerResponse(status) => {
+                if status.is_client_error() {
+                    is_real_error = true;
+                } else {
+                    eprintln!("Unsuccessful server response for link `{}`", link.link.uri);
+                }
+            }
+            Reason::Client(err) => {
+                if err.is_timeout() {
+                    eprintln!("Timeout for link `{}`", link.link.uri);
+                } else if err.is_server_error() {
+                    eprintln!("Server error for link `{}`", link.link.uri);
+                } else {
+                    is_real_error = true;
+                }
+            }
+        }
+    }
+
+    if is_real_error {
+        Ok((outcome.generate_diagnostics(&files, linkck_cfg.warning_policy), files))
+    } else {
+        Ok((vec![], files))
+    }
 }
 
 // Build command implementation