Merge pull request #792 from rust-lang-nursery/custom-preprocessor

WIP: Custom Preprocessors

10 files changed, 555 insertions, 125 deletions

diff --git a/book-example/src/for_developers/preprocessors.md b/book-example/src/for_developers/preprocessors.md
index 03c915bb..e8bcaf0a 100644
--- a/book-example/src/for_developers/preprocessors.md
+++ b/book-example/src/for_developers/preprocessors.md
@@ -11,68 +11,71 @@ the book. Possible use cases are:
   mathjax equivalents
 
 
-## Implementing a Preprocessor
+## Hooking Into MDBook
+
+MDBook uses a fairly simple mechanism for discovering third party plugins.
+A new table is added to `book.toml` (e.g. `preprocessor.foo` for the `foo`
+preprocessor) and then `mdbook` will try to invoke the `mdbook-foo` program as
+part of the build process.
+
+While preprocessors can be hard-coded to specify which backend it should be run
+for (e.g. it doesn't make sense for MathJax to be used for non-HTML renderers)
+with the `preprocessor.foo.renderer` key.
+
+```toml
+[book]
+title = "My Book"
+authors = ["Michael-F-Bryan"]
+
+[preprocessor.foo]
+# The command can also be specified manually
+command = "python3 /path/to/foo.py"
+# Only run the `foo` preprocessor for the HTML and EPUB renderer
+renderer = ["html", "epub"]
+```
 
-A preprocessor is represented by the `Preprocessor` trait.
+In typical unix style, all inputs to the plugin will be written to `stdin` as
+JSON and `mdbook` will read from `stdout` if it is expecting output.
 
-```rust
-pub trait Preprocessor {
-    fn name(&self) -> &str;
-    fn run(&self, ctx: &PreprocessorContext, book: Book) -> Result<Book>;
-    fn supports_renderer(&self, _renderer: &str) -> bool {
-        true
-    }
-}
-```
+The easiest way to get started is by creating your own implementation of the
+`Preprocessor` trait (e.g. in `lib.rs`) and then creating a shell binary which
+translates inputs to the correct `Preprocessor` method. For convenience, there
+is [an example no-op preprocessor] in the `examples/` directory which can easily
+be adapted for other preprocessors.
 
-Where the `PreprocessorContext` is defined as
+<details>
+<summary>Example no-op preprocessor</summary>
 
 ```rust
-pub struct PreprocessorContext {
-    pub root: PathBuf,
-    pub config: Config,
-    /// The `Renderer` this preprocessor is being used with.
-    pub renderer: String,
-}
-```
-
-The `renderer` value allows you react accordingly, for example, PDF or HTML.
+// nop-preprocessors.rs
 
-## A complete Example
+{{#include ../../../examples/nop-preprocessor.rs}}
+```
+</details>
 
-The magic happens within the `run(...)` method of the
-[`Preprocessor`][preprocessor-docs] trait implementation.
+## Hints For Implementing A Preprocessor
 
-As direct access to the chapters is not possible, you will probably end up
-iterating them using `for_each_mut(...)`:
+By pulling in `mdbook` as a library, preprocessors can have access to the
+existing infrastructure for dealing with books.
 
-```rust
-book.for_each_mut(|item: &mut BookItem| {
-    if let BookItem::Chapter(ref mut chapter) = *item {
-      eprintln!("{}: processing chapter '{}'", self.name(), chapter.name);
-      res = Some(
-          match Deemphasize::remove_emphasis(&mut num_removed_items, chapter) {
-              Ok(md) => {
-                  chapter.content = md;
-                  Ok(())
-              }
-              Err(err) => Err(err),
-          },
-      );
-  }
-});
-```
+For example, a custom preprocessor could use the
+[`CmdPreprocessor::parse_input()`] function to deserialize the JSON written to
+`stdin`. Then each chapter of the `Book` can be mutated in-place via
+[`Book::for_each_mut()`], and then written to `stdout` with the `serde_json`
+crate.
 
-The `chapter.content` is just a markdown formatted string, and you will have to
-process it in some way. Even though it's entirely possible to implement some
-sort of manual find & replace operation, if that feels too unsafe you can use
-[`pulldown-cmark`][pc] to parse the string into events and work on them instead.
+Chapters can be accessed either directly (by recursively iterating over
+chapters) or via the `Book::for_each_mut()` convenience method.
 
-Finally you can use [`pulldown-cmark-to-cmark`][pctc] to transform these events
-back to a string.
+The `chapter.content` is just a string which happens to be markdown. While it's
+entirely possible to use regular expressions or do a manual find & replace,
+you'll probably want to process the input into something more computer-friendly.
+The [`pulldown-cmark`][pc] crate implements a production-quality event-based
+Markdown parser, with the [`pulldown-cmark-to-cmark`][pctc] allowing you to
+translate events back into markdown text.
 
-The following code block shows how to remove all emphasis from markdown, and do
-so safely.
+The following code block shows how to remove all emphasis from markdown,
+without accidentally breaking the document.
 
 ```rust
 fn remove_emphasis(
@@ -107,3 +110,6 @@ For everything else, have a look [at the complete example][example].
 [pc]: https://crates.io/crates/pulldown-cmark
 [pctc]: https://crates.io/crates/pulldown-cmark-to-cmark
 [example]: https://github.com/rust-lang-nursery/mdBook/blob/master/examples/de-emphasize.rs
+[an example no-op preprocessor]: https://github.com/rust-lang-nursery/mdBook/blob/master/examples/nop-preprocessor.rs
+[`CmdPreprocessor::parse_input()`]: https://docs.rs/mdbook/latest/mdbook/preprocess/trait.Preprocessor.html#method.parse_input
+[`Book::for_each_mut()`]: https://docs.rs/mdbook/latest/mdbook/book/struct.Book.html#method.for_each_mut
diff --git a/examples/de-emphasize.rs b/examples/de-emphasize.rs
index 88c1b3a4..933e5c45 100644
--- a/examples/de-emphasize.rs
+++ b/examples/de-emphasize.rs
@@ -1,4 +1,6 @@
-//! This program removes all forms of emphasis from the markdown of the book.
+//! An example preprocessor for removing all forms of emphasis from a markdown
+//! book.
+
 extern crate mdbook;
 extern crate pulldown_cmark;
 extern crate pulldown_cmark_to_cmark;
@@ -6,31 +8,13 @@ extern crate pulldown_cmark_to_cmark;
 use mdbook::book::{Book, BookItem, Chapter};
 use mdbook::errors::{Error, Result};
 use mdbook::preprocess::{Preprocessor, PreprocessorContext};
-use mdbook::MDBook;
 use pulldown_cmark::{Event, Parser, Tag};
 use pulldown_cmark_to_cmark::fmt::cmark;
 
-use std::env::{args, args_os};
-use std::ffi::OsString;
-use std::process;
-
 const NAME: &str = "md-links-to-html-links";
 
-fn do_it(book: OsString) -> Result<()> {
-    let mut book = MDBook::load(book)?;
-    book.with_preprecessor(Deemphasize);
-    book.build()
-}
-
 fn main() {
-    if args_os().count() != 2 {
-        eprintln!("USAGE: {} <book>", args().next().expect("executable"));
-        return;
-    }
-    if let Err(e) = do_it(args_os().skip(1).next().expect("one argument")) {
-        eprintln!("{}", e);
-        process::exit(1);
-    }
+    panic!("This example is intended to be part of a library");
 }
 
 struct Deemphasize;
diff --git a/examples/nop-preprocessor.rs b/examples/nop-preprocessor.rs
new file mode 100644
index 00000000..d4615ef6
--- /dev/null
+++ b/examples/nop-preprocessor.rs
@@ -0,0 +1,112 @@
+extern crate clap;
+extern crate mdbook;
+extern crate serde_json;
+
+use clap::{App, Arg, ArgMatches, SubCommand};
+use mdbook::book::Book;
+use mdbook::errors::Error;
+use mdbook::preprocess::{CmdPreprocessor, Preprocessor, PreprocessorContext};
+use std::io;
+use std::process;
+use nop_lib::Nop;
+
+pub fn make_app() -> App<'static, 'static> {
+    App::new("nop-preprocessor")
+        .about("A mdbook preprocessor which does precisely nothing")
+        .subcommand(
+            SubCommand::with_name("supports")
+                .arg(Arg::with_name("renderer").required(true))
+                .about("Check whether a renderer is supported by this preprocessor"))
+}
+
+fn main() {
+    let matches = make_app().get_matches();
+
+    // Users will want to construct their own preprocessor here
+    let preprocessor = Nop::new();
+
+    if let Some(sub_args) = matches.subcommand_matches("supports") {
+        handle_supports(&preprocessor, sub_args);
+    } else {
+        if let Err(e) = handle_preprocessing(&preprocessor) {
+            eprintln!("{}", e);
+            process::exit(1);
+        }
+    }
+}
+
+fn handle_preprocessing(pre: &dyn Preprocessor) -> Result<(), Error> {
+    let (ctx, book) = CmdPreprocessor::parse_input(io::stdin())?;
+
+    if ctx.mdbook_version != mdbook::MDBOOK_VERSION {
+        // We should probably use the `semver` crate to check compatibility
+        // here...
+        eprintln!(
+            "Warning: The {} plugin was built against version {} of mdbook, \
+             but we're being called from version {}",
+            pre.name(),
+            mdbook::MDBOOK_VERSION,
+            ctx.mdbook_version
+        );
+    }
+
+    let processed_book = pre.run(&ctx, book)?;
+    serde_json::to_writer(io::stdout(), &processed_book)?;
+
+    Ok(())
+}
+
+fn handle_supports(pre: &dyn Preprocessor, sub_args: &ArgMatches) -> ! {
+    let renderer = sub_args.value_of("renderer").expect("Required argument");
+    let supported = pre.supports_renderer(&renderer);
+
+    // Signal whether the renderer is supported by exiting with 1 or 0.
+    if supported {
+        process::exit(0);
+    } else {
+        process::exit(1);
+    }
+}
+
+/// The actual implementation of the `Nop` preprocessor. This would usually go
+/// in your main `lib.rs` file.
+mod nop_lib {
+    use super::*;
+
+    /// A no-op preprocessor.
+    pub struct Nop;
+
+    impl Nop {
+        pub fn new() -> Nop {
+            Nop
+        }
+    }
+
+    impl Preprocessor for Nop {
+        fn name(&self) -> &str {
+            "nop-preprocessor"
+        }
+
+        fn run(
+            &self,
+            ctx: &PreprocessorContext,
+            book: Book,
+        ) -> Result<Book, Error> {
+            // In testing we want to tell the preprocessor to blow up by setting a
+            // particular config value
+            if let Some(nop_cfg) = ctx.config.get_preprocessor(self.name()) {
+                if nop_cfg.contains_key("blow-up") {
+                    return Err("Boom!!1!".into());
+                }
+            }
+
+            // we *are* a no-op preprocessor after all
+            Ok(book)
+        }
+
+        fn supports_renderer(&self, renderer: &str) -> bool {
+            renderer != "not-supported"
+        }
+    }
+}
+
diff --git a/src/book/mod.rs b/src/book/mod.rs
index f1e3daf4..3ffbde6f 100644
--- a/src/book/mod.rs
+++ b/src/book/mod.rs
@@ -20,7 +20,8 @@ use tempfile::Builder as TempFileBuilder;
 use toml::Value;
 
 use errors::*;
-use preprocess::{IndexPreprocessor, LinkPreprocessor, Preprocessor, PreprocessorContext};
+use preprocess::{IndexPreprocessor, LinkPreprocessor, Preprocessor,
+    PreprocessorContext, CmdPreprocessor};
 use renderer::{CmdRenderer, HtmlHandlebars, RenderContext, Renderer};
 use utils;
 
@@ -356,36 +357,48 @@ fn is_default_preprocessor(pre: &Preprocessor) -> bool {
 
 /// Look at the `MDBook` and try to figure out what preprocessors to run.
 fn determine_preprocessors(config: &Config) -> Result<Vec<Box<Preprocessor>>> {
-    let preprocessor_keys = config.get("preprocessor")
-        .and_then(|value| value.as_table())
-        .map(|table| table.keys());
-
-    let mut preprocessors = if config.build.use_default_preprocessors {
-        default_preprocessors()
-    } else {
-        Vec::new()
-    };
-
-    let preprocessor_keys = match preprocessor_keys {
-        Some(keys) => keys,
-        // If no preprocessor field is set, default to the LinkPreprocessor and
-        // IndexPreprocessor. This allows you to disable default preprocessors
-        // by setting "preprocess" to an empty list.
-        None => return Ok(preprocessors),
-    };
-
-    for key in preprocessor_keys {
-        match key.as_ref() {
-            "links" => preprocessors.push(Box::new(LinkPreprocessor::new())),
-            "index" => preprocessors.push(Box::new(IndexPreprocessor::new())),
-            _ => bail!("{:?} is not a recognised preprocessor", key),
+    let mut preprocessors = Vec::new();
+
+    if config.build.use_default_preprocessors {
+        preprocessors.extend(default_preprocessors());
+    }
+
+    if let Some(preprocessor_table) =
+        config.get("preprocessor").and_then(|v| v.as_table())
+    {
+        for key in preprocessor_table.keys() {
+            match key.as_ref() {
+                "links" => {
+                    preprocessors.push(Box::new(LinkPreprocessor::new()))
+                }
+                "index" => {
+                    preprocessors.push(Box::new(IndexPreprocessor::new()))
+                }
+                name => preprocessors.push(interpret_custom_preprocessor(
+                    name,
+                    &preprocessor_table[name],
+                )),
+            }
         }
     }
 
     Ok(preprocessors)
 }
 
-fn interpret_custom_renderer(key: &str, table: &Value) -> Box<Renderer> {
+fn interpret_custom_preprocessor(
+    key: &str,
+    table: &Value,
+) -> Box<CmdPreprocessor> {
+    let command = table
+        .get("command")
+        .and_then(|c| c.as_str())
+        .map(|s| s.to_string())
+        .unwrap_or_else(|| format!("mdbook-{}", key));
+
+    Box::new(CmdPreprocessor::new(key.to_string(), command.to_string()))
+}
+
+fn interpret_custom_renderer(key: &str, table: &Value) -> Box<CmdRenderer> {
     // look for the `command` field, falling back to using the key
     // prepended by "mdbook-"
     let table_dot_command = table
@@ -393,7 +406,8 @@ fn interpret_custom_renderer(key: &str, table: &Value) -> Box<Renderer> {
         .and_then(|c| c.as_str())
         .map(|s| s.to_string());
 
-    let command = table_dot_command.unwrap_or_else(|| format!("mdbook-{}", key));
+    let command =
+        table_dot_command.unwrap_or_else(|| format!("mdbook-{}", key));
 
     Box::new(CmdRenderer::new(key.to_string(), command.to_string()))
 }
@@ -492,7 +506,7 @@ mod tests {
     }
 
     #[test]
-    fn config_complains_if_unimplemented_preprocessor() {
+    fn can_determine_third_party_preprocessors() {
         let cfg_str: &'static str = r#"
         [book]
         title = "Some Book"
@@ -509,9 +523,28 @@ mod tests {
         // make sure the `preprocessor.random` table exists
         assert!(cfg.get_preprocessor("random").is_some());
 
-        let got = determine_preprocessors(&cfg);
+        let got = determine_preprocessors(&cfg).unwrap();
+
+        assert!(got.into_iter().any(|p| p.name() == "random"));
+    }
+
+    #[test]
+    fn preprocessors_can_provide_their_own_commands() {
+        let cfg_str = r#"
+        [preprocessor.random]
+        command = "python random.py"
+        "#;
+
+        let cfg = Config::from_str(cfg_str).unwrap();
+
+        // make sure the `preprocessor.random` table exists
+        let random = cfg.get_preprocessor("random").unwrap();
+        let random = interpret_custom_preprocessor(
+            "random",
+            &Value::Table(random.clone()),
+        );
 
-        assert!(got.is_err());
+        assert_eq!(random.cmd(), "python random.py");
     }
 
     #[test]
diff --git a/src/lib.rs b/src/lib.rs
index 2c91e283..028a2ba8 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -114,6 +114,12 @@ pub mod renderer;
 pub mod theme;
 pub mod utils;
 
+/// The current version of `mdbook`.
+///
+/// This is provided as a way for custom preprocessors and renderers to do
+/// compatibility checks.
+pub const MDBOOK_VERSION: &str = env!("CARGO_PKG_VERSION");
+
 pub use book::BookItem;
 pub use book::MDBook;
 pub use config::Config;
diff --git a/src/preprocess/cmd.rs b/src/preprocess/cmd.rs
new file mode 100644
index 00000000..4398eba6
--- /dev/null
+++ b/src/preprocess/cmd.rs
@@ -0,0 +1,154 @@
+use super::{Preprocessor, PreprocessorContext};
+use book::Book;
+use errors::*;
+use serde_json;
+use shlex::Shlex;
+use std::io::{self, Read, Write};
+use std::process::{Child, Command, Stdio};
+
+/// A custom preprocessor which will shell out to a 3rd-party program.
+///
+/// # Preprocessing Protocol
+///
+/// When the `supports_renderer()` method is executed, `CmdPreprocessor` will
+/// execute the shell command `$cmd supports $renderer`. If the renderer is
+/// supported, custom preprocessors should exit with a exit code of `0`,
+/// any other exit code be considered as unsupported.
+///
+/// The `run()` method is implemented by passing a `(PreprocessorContext, Book)`
+/// tuple to the spawned command (`$cmd`) as JSON via `stdin`. Preprocessors
+/// should then "return" a processed book by printing it to `stdout` as JSON.
+/// For convenience, the `CmdPreprocessor::parse_input()` function can be used
+/// to parse the input provided by `mdbook`.
+///
+/// Exiting with a non-zero exit code while preprocessing is considered an
+/// error. `stderr` is passed directly through to the user, so it can be used
+/// for logging or emitting warnings if desired.
+///
+/// # Examples
+///
+/// An example preprocessor is available in this project's `examples/`
+/// directory.
+#[derive(Debug, Clone, PartialEq)]
+pub struct CmdPreprocessor {
+    name: String,
+    cmd: String,
+}
+
+impl CmdPreprocessor {
+    /// Create a new `CmdPreprocessor`.
+    pub fn new(name: String, cmd: String) -> CmdPreprocessor {
+        CmdPreprocessor { name, cmd }
+    }
+
+    /// A convenience function custom preprocessors can use to parse the input
+    /// written to `stdin` by a `CmdRenderer`.
+    pub fn parse_input<R: Read>(
+        reader: R,
+    ) -> Result<(PreprocessorContext, Book)> {
+        serde_json::from_reader(reader)
+            .chain_err(|| "Unable to parse the input")
+    }
+
+    fn write_input_to_child(
+        &self,
+        child: &mut Child,
+        book: &Book,
+        ctx: &PreprocessorContext,
+    ) {
+        let stdin = child.stdin.take().expect("Child has stdin");
+
+        if let Err(e) = self.write_input(stdin, &book, &ctx) {
+            // Looks like the backend hung up before we could finish
+            // sending it the render context. Log the error and keep going
+            warn!("Error writing the RenderContext to the backend, {}", e);
+        }
+    }
+
+    fn write_input<W: Write>(&self, writer: W, book: &Book, ctx: &PreprocessorContext) -> Result<()> {
+        serde_json::to_writer(writer, &(ctx, book))
+            .map_err(Into::into)
+    }
+
+    /// The command this `Preprocessor` will invoke.
+    pub fn cmd(&self) -> &str {
+        &self.cmd
+    }
+
+    fn command(&self) -> Result<Command> {
+        let mut words = Shlex::new(&self.cmd);
+        let executable = match words.next() {
+            Some(e) => e,
+            None => bail!("Command string was empty"),
+        };
+
+        let mut cmd = Command::new(executable);
+
+        for arg in words {
+            cmd.arg(arg);
+        }
+
+        Ok(cmd)
+    }
+}
+
+impl Preprocessor for CmdPreprocessor {
+    fn name(&self) -> &str {
+        &self.name
+    }
+
+    fn run(&self, ctx: &PreprocessorContext, book: Book) -> Result<Book> {
+        let mut cmd = self.command()?;
+
+        let mut child = cmd
+            .stdin(Stdio::piped())
+            .stdout(Stdio::piped())
+            .stderr(Stdio::inherit())
+            .spawn()
+            .chain_err(|| format!("Unable to start the \"{}\" preprocessor. Is it installed?", self.name()))?;
+
+        self.write_input_to_child(&mut child, &book, ctx);
+
+        let output = child
+            .wait_with_output()
+            .chain_err(|| "Error waiting for the preprocessor to complete")?;
+
+        trace!("{} exited with output: {:?}", self.cmd, output);
+        ensure!(output.status.success(), "The preprocessor exited unsuccessfully");
+
+        serde_json::from_slice(&output.stdout).chain_err(|| "Unable to parse the preprocessed book")
+    }
+
+    fn supports_renderer(&self, renderer: &str) -> bool {
+        debug!("Checking if the \"{}\" preprocessor supports \"{}\"", self.name(), renderer);
+
+        let mut cmd = match self.command() {
+            Ok(c) => c,
+            Err(e) => {
+                warn!("Unable to create the command for the \"{}\" preprocessor, {}", self.name(), e);
+                return false;
+            }
+        };
+
+        let outcome = cmd
+            .arg("supports")
+            .arg(renderer)
+            .stdin(Stdio::null())
+            .stdout(Stdio::inherit())
+            .stderr(Stdio::inherit())
+            .status()
+            .map(|status| status.code() == Some(0));
+
+        if let Err(ref e) = outcome {
+            if e.kind() == io::ErrorKind::NotFound {
+                warn!(
+                    "The command wasn't found, is the \"{}\" preprocessor installed?",
+                    self.name
+                );
+                warn!("\tCommand: {}", self.cmd);
+            }
+        }
+
+        outcome.unwrap_or(false)
+    }
+}
diff --git a/src/preprocess/links.rs b/src/preprocess/links.rs
index 870f96a8..3403891a 100644
--- a/src/preprocess/links.rs
+++ b/src/preprocess/links.rs
@@ -69,7 +69,12 @@ where
             Ok(new_content) => {
                 if depth < MAX_LINK_NESTED_DEPTH {
                     if let Some(rel_path) = playpen.link.relative_path(path) {
-                        replaced.push_str(&replace_all(&new_content, rel_path, source, depth + 1));
+                        replaced.push_str(&replace_all(
+                            &new_content,
+                            rel_path,
+                            source,
+                            depth + 1,
+                        ));
                     } else {
                         replaced.push_str(&new_content);
                     }
@@ -83,6 +88,10 @@ where
             }
             Err(e) => {
                 error!("Error updating \"{}\", {}", playpen.link_text, e);
+                for cause in e.iter().skip(1) {
+                    warn!("Caused By: {}", cause);
+                }
+
                 // This should make sure we include the raw `{{# ... }}` snippet
                 // in the page content if there are any errors.
                 previous_end_index = playpen.start_index;
@@ -109,10 +118,18 @@ impl<'a> LinkType<'a> {
         let base = base.as_ref();
         match self {
             LinkType::Escaped => None,
-            LinkType::IncludeRange(p, _) => Some(return_relative_path(base, &p)),
-            LinkType::IncludeRangeFrom(p, _) => Some(return_relative_path(base, &p)),
-            LinkType::IncludeRangeTo(p, _) => Some(return_relative_path(base, &p)),
-            LinkType::IncludeRangeFull(p, _) => Some(return_relative_path(base, &p)),
+            LinkType::IncludeRange(p, _) => {
+                Some(return_relative_path(base, &p))
+            }
+            LinkType::IncludeRangeFrom(p, _) => {
+                Some(return_relative_path(base, &p))
+            }
+            LinkType::IncludeRangeTo(p, _) => {
+                Some(return_relative_path(base, &p))
+            }
+            LinkType::IncludeRangeFull(p, _) => {
+                Some(return_relative_path(base, &p))
+            }
             LinkType::Playpen(p, _) => Some(return_relative_path(base, &p)),
         }
     }
@@ -182,11 +199,15 @@ impl<'a> Link<'a> {
 
                 match (typ.as_str(), file_arg) {
                     ("include", Some(pth)) => Some(parse_include_path(pth)),
-                    ("playpen", Some(pth)) => Some(LinkType::Playpen(pth.into(), props)),
+                    ("playpen", Some(pth)) => {
+                        Some(LinkType::Playpen(pth.into(), props))
+                    }
                     _ => None,
                 }
             }
-            (Some(mat), None, None) if mat.as_str().starts_with(ESCAPE_CHAR) => {
+            (Some(mat), None, None)
+                if mat.as_str().starts_with(ESCAPE_CHAR) =>
+            {
                 Some(LinkType::Escaped)
             }
             _ => None,
@@ -207,20 +228,65 @@ impl<'a> Link<'a> {
         match self.link {
             // omit the escape char
             LinkType::Escaped => Ok((&self.link_text[1..]).to_owned()),
-            LinkType::IncludeRange(ref pat, ref range) => file_to_string(base.join(pat))
-                .map(|s| take_lines(&s, range.clone()))
-                .chain_err(|| format!("Could not read file for link {}", self.link_text)),
-            LinkType::IncludeRangeFrom(ref pat, ref range) => file_to_string(base.join(pat))
-                .map(|s| take_lines(&s, range.clone()))
-                .chain_err(|| format!("Could not read file for link {}", self.link_text)),
-            LinkType::IncludeRangeTo(ref pat, ref range) => file_to_string(base.join(pat))
-                .map(|s| take_lines(&s, range.clone()))
-                .chain_err(|| format!("Could not read file for link {}", self.link_text)),
-            LinkType::IncludeRangeFull(ref pat, _) => file_to_string(base.join(pat))
-                .chain_err(|| format!("Could not read file for link {}", self.link_text)),
+            LinkType::IncludeRange(ref pat, ref range) => {
+                let target = base.join(pat);
+
+                file_to_string(&target)
+                    .map(|s| take_lines(&s, range.clone()))
+                    .chain_err(|| {
+                        format!(
+                            "Could not read file for link {} ({})",
+                            self.link_text,
+                            target.display(),
+                        )
+                    })
+            }
+            LinkType::IncludeRangeFrom(ref pat, ref range) => {
+                let target = base.join(pat);
+
+                file_to_string(&target)
+                    .map(|s| take_lines(&s, range.clone()))
+                    .chain_err(|| {
+                        format!(
+                            "Could not read file for link {} ({})",
+                            self.link_text,
+                            target.display(),
+                        )
+                    })
+            }
+            LinkType::IncludeRangeTo(ref pat, ref range) => {
+                let target = base.join(pat);
+
+                file_to_string(&target)
+                    .map(|s| take_lines(&s, range.clone()))
+                    .chain_err(|| {
+                        format!(
+                            "Could not read file for link {} ({})",
+                            self.link_text,
+                            target.display(),
+                        )
+                    })
+            }
+            LinkType::IncludeRangeFull(ref pat, _) => {
+                let target = base.join(pat);
+
+                file_to_string(&target).chain_err(|| {
+                    format!("Could not read file for link {} ({})",
+                            self.link_text,
+                            target.display())
+                })
+            }
             LinkType::Playpen(ref pat, ref attrs) => {
-                let contents = file_to_string(base.join(pat))
-                    .chain_err(|| format!("Could not read file for link {}", self.link_text))?;
+                let target = base.join(pat);
+
+                let contents =
+                file_to_string(&target).chain_err(|| {
+                        format!(
+                            "Could not read file for link {} ({})",
+                            self.link_text,
+                            target.display()
+                        )
+                    })?;
                 let ftype = if !attrs.is_empty() { "rust," } else { "rust" };
                 Ok(format!(
                     "```{}{}\n{}\n```\n",
@@ -465,7 +531,10 @@ mod tests {
                 Link {
                     start_index: 38,
                     end_index: 68,
-                    link: LinkType::Playpen(PathBuf::from("file.rs"), vec!["editable"]),
+                    link: LinkType::Playpen(
+                        PathBuf::from("file.rs"),
+                        vec!["editable"]
+                    ),
                     link_text: "{{#playpen file.rs editable }}",
                 },
                 Link {
@@ -475,7 +544,8 @@ mod tests {
                         PathBuf::from("my.rs"),
                         vec!["editable", "no_run", "should_panic"],
                     ),
-                    link_text: "{{#playpen my.rs editable no_run should_panic}}",
+                    link_text:
+                        "{{#playpen my.rs editable no_run should_panic}}",
                 },
             ]
         );
diff --git a/src/preprocess/mod.rs b/src/preprocess/mod.rs
index 5f59c5bf..b7ab1986 100644
--- a/src/preprocess/mod.rs
+++ b/src/preprocess/mod.rs
@@ -2,9 +2,11 @@
 
 pub use self::index::IndexPreprocessor;
 pub use self::links::LinkPreprocessor;
+pub use self::cmd::CmdPreprocessor;
 
 mod index;
 mod links;
+mod cmd;
 
 use book::Book;
 use config::Config;
@@ -14,6 +16,7 @@ use std::path::PathBuf;
 
 /// Extra information for a `Preprocessor` to give them more context when
 /// processing a book.
+#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
 pub struct PreprocessorContext {
     /// The location of the book directory on disk.
     pub root: PathBuf,
@@ -21,12 +24,21 @@ pub struct PreprocessorContext {
     pub config: Config,
     /// The `Renderer` this preprocessor is being used with.
     pub renderer: String,
+    /// The calling `mdbook` version.
+    pub mdbook_version: String,
+    __non_exhaustive: (),
 }
 
 impl PreprocessorContext {
     /// Create a new `PreprocessorContext`.
     pub(crate) fn new(root: PathBuf, config: Config, renderer: String) -> Self {
-        PreprocessorContext { root, config, renderer }
+        PreprocessorContext {
+            root,
+            config,
+            renderer,
+            mdbook_version: ::MDBOOK_VERSION.to_string(),
+            __non_exhaustive: (),
+        }
     }
 }
 
diff --git a/src/renderer/mod.rs b/src/renderer/mod.rs
index 906f7e27..41176ae0 100644
--- a/src/renderer/mod.rs
+++ b/src/renderer/mod.rs
@@ -26,8 +26,6 @@ use book::Book;
 use config::Config;
 use errors::*;
 
-const MDBOOK_VERSION: &str = env!("CARGO_PKG_VERSION");
-
 /// An arbitrary `mdbook` backend.
 ///
 /// Although it's quite possible for you to import `mdbook` as a library and
@@ -66,6 +64,7 @@ pub struct RenderContext {
     /// renderers to cache intermediate results, this directory is not
     /// guaranteed to be empty or even exist.
     pub destination: PathBuf,
+    __non_exhaustive: (),
 }
 
 impl RenderContext {
@@ -78,9 +77,10 @@ impl RenderContext {
         RenderContext {
             book: book,
             config: config,
-            version: MDBOOK_VERSION.to_string(),
+            version: ::MDBOOK_VERSION.to_string(),
             root: root.into(),
             destination: destination.into(),
+            __non_exhaustive: (),
         }
     }
 
diff --git a/tests/custom_preprocessors.rs b/tests/custom_preprocessors.rs
new file mode 100644
index 00000000..54bfd51d
--- /dev/null
+++ b/tests/custom_preprocessors.rs
@@ -0,0 +1,53 @@
+extern crate mdbook;
+
+mod dummy_book;
+
+use dummy_book::DummyBook;
+use mdbook::preprocess::{CmdPreprocessor, Preprocessor};
+use mdbook::MDBook;
+
+fn example() -> CmdPreprocessor {
+    CmdPreprocessor::new("nop-preprocessor".to_string(), "cargo run --example nop-preprocessor --".to_string())
+}
+
+#[test]
+fn example_supports_whatever() {
+    let cmd = example();
+
+    let got = cmd.supports_renderer("whatever");
+
+    assert_eq!(got, true);
+}
+
+#[test]
+fn example_doesnt_support_not_supported() {
+    let cmd = example();
+
+    let got = cmd.supports_renderer("not-supported");
+
+    assert_eq!(got, false);
+}
+
+#[test]
+fn ask_the_preprocessor_to_blow_up() {
+    let dummy_book = DummyBook::new();
+    let temp = dummy_book.build().unwrap();
+    let mut md = MDBook::load(temp.path()).unwrap();
+    md.with_preprecessor(example());
+
+    md.config.set("preprocessor.nop-preprocessor.blow-up", true).unwrap();
+
+    let got = md.build();
+
+    assert!(got.is_err());
+}
+
+#[test]
+fn process_the_dummy_book() {
+    let dummy_book = DummyBook::new();
+    let temp = dummy_book.build().unwrap();
+    let mut md = MDBook::load(temp.path()).unwrap();
+    md.with_preprecessor(example());