sdk: Allow any byte stream for hashing dictionary entries

Fatshark has a few weird string fields, where they provide a length
field, but then sometimes write a shorter, NUL-terminated string into
that same field and adding padding up to the "advertised" length.
To properly read those strings, we can't rely on just the length field
anymore, but need to check for a NUL, too.

crates/dtmt/src/cmd/build.rs

@@ -55,6 +55,7 @@ pub(crate) fn command_definition() -> Command {
         )
 }
 
+/// Try to find a `dtmt.cfg` in the given directory or traverse up the parents.
 #[tracing::instrument]
 async fn find_project_config(dir: Option<PathBuf>) -> Result<ModConfig> {
     let (path, mut file) = if let Some(path) = dir {
@@ -102,6 +103,8 @@ async fn find_project_config(dir: Option<PathBuf>) -> Result<ModConfig> {
     Ok(cfg)
 }
 
+/// Iterate over the paths in the given `Package` and
+/// compile each file by its file type.
 #[tracing::instrument(skip_all)]
 async fn compile_package_files(pkg: &Package, cfg: &ModConfig) -> Result<Vec<BundleFile>> {
     let root = Arc::new(&cfg.dir);
@@ -148,6 +151,8 @@ async fn compile_package_files(pkg: &Package, cfg: &ModConfig) -> Result<Vec<Bun
     results.into_iter().collect()
 }
 
+/// Read a `.package` file, collect the referenced files
+/// and compile all of them into a bundle.
 #[tracing::instrument]
 async fn build_package(
     cfg: &ModConfig,
@@ -176,6 +181,8 @@ async fn build_package(
     Ok(bundle)
 }
 
+/// Cleans the path of internal parent (`../`) or self (`./`) components,
+/// and ensures that it is relative.
 fn normalize_file_path<P: AsRef<Path>>(path: P) -> Result<PathBuf> {
     let path = path.as_ref();
 

crates/dtmt/src/cmd/bundle/list.rs

@@ -36,6 +36,18 @@ enum OutputFormat {
     Text,
 }
 
+fn format_byte_size(size: usize) -> String {
+    if size < 1024 {
+        format!("{} Bytes", size)
+    } else if size < 1024 * 1024 {
+        format!("{} kB", size / 1024)
+    } else if size < 1024 * 1024 * 1024 {
+        format!("{} MB", size / (1024 * 1024))
+    } else {
+        format!("{} GB", size / (1024 * 1024 * 1024))
+    }
+}
+
 #[tracing::instrument(skip(ctx))]
 async fn print_bundle_contents<P>(ctx: &sdk::Context, path: P, fmt: OutputFormat) -> Result<()>
 where
@@ -50,7 +62,11 @@ where
 
     match fmt {
         OutputFormat::Text => {
-            println!("Bundle: {}", bundle.name().display());
+            println!(
+                "Bundle: {} ({:016x})",
+                bundle.name().display(),
+                bundle.name()
+            );
 
             for f in bundle.files().iter() {
                 if f.variants().len() != 1 {
@@ -63,9 +79,10 @@ where
 
                 let v = &f.variants()[0];
                 println!(
-                    "\t{}.{}: {} bytes",
+                    "\t{}.{}: {} ({})",
                     f.base_name().display(),
                     f.file_type().ext_name(),
+                    format_byte_size(v.size()),
                     v.size()
                 );
             }

lib/dtmt-shared/src/log.rs

@@ -84,7 +84,7 @@ pub fn create_tracing_subscriber() {
         EnvFilter::try_from_default_env().unwrap_or_else(|_| EnvFilter::try_new("info").unwrap());
 
     let (dev_stdout_layer, prod_stdout_layer, filter_layer) = if cfg!(debug_assertions) {
-        let fmt_layer = fmt::layer().pretty();
+        let fmt_layer = fmt::layer().pretty().with_writer(std::io::stderr);
         (Some(fmt_layer), None, None)
     } else {
         // Creates a layer that
@@ -93,6 +93,7 @@ pub fn create_tracing_subscriber() {
         // - does not print spans/targets
         // - only prints time, not date
         let fmt_layer = fmt::layer()
+            .with_writer(std::io::stderr)
             .event_format(Formatter)
             .fmt_fields(debug_fn(format_fields));
 

lib/sdk/src/binary.rs

@@ -43,6 +43,7 @@ impl<T: FromBinary> FromBinary for Vec<T> {
 }
 
 pub mod sync {
+    use std::ffi::CStr;
     use std::io::{self, Read, Seek, SeekFrom};
 
     use byteorder::{LittleEndian, ReadBytesExt, WriteBytesExt};
@@ -165,25 +166,13 @@ pub mod sync {
         }
 
         fn read_string_len(&mut self, len: usize) -> Result<String> {
-            let mut buf = vec![0; len];
+            let pos = self.stream_position();
-            let res = self
-                .read_exact(&mut buf)
-                .map_err(Report::new)
-                .and_then(|_| {
-                    String::from_utf8(buf).map_err(|err| {
-                        let ascii = String::from_utf8_lossy(err.as_bytes()).to_string();
-                        let bytes = format!("{:?}", err.as_bytes());
-                        Report::new(err)
-                            .with_section(move || bytes.header("Bytes:"))
-                            .with_section(move || ascii.header("ASCII:"))
-                    })
-                });
 
+            let res = read_string_len(self, len);
             if res.is_ok() {
                 return res;
             }
 
-            let pos = self.stream_position();
             if pos.is_ok() {
                 res.with_section(|| {
                     format!("{pos:#X} ({pos})", pos = pos.unwrap()).header("Position: ")
@@ -243,4 +232,22 @@ pub mod sync {
 
         Err(err).with_section(|| format!("{pos:#X} ({pos})").header("Position: "))
     }
+
+    fn read_string_len(mut r: impl Read, len: usize) -> Result<String> {
+        let mut buf = vec![0; len];
+        r.read_exact(&mut buf)
+            .wrap_err_with(|| format!("Failed to read {} bytes", len))?;
+
+        let res = match CStr::from_bytes_until_nul(&buf) {
+            Ok(s) => {
+                let s = s.to_str()?;
+                Ok(s.to_string())
+            }
+            Err(_) => String::from_utf8(buf.clone()).map_err(Report::new),
+        };
+
+        res.wrap_err("Invalid binary for UTF8 string")
+            .with_section(|| format!("{}", String::from_utf8_lossy(&buf)).header("ASCI:"))
+            .with_section(|| format!("{:x?}", buf).header("Bytes:"))
+    }
 }

lib/sdk/src/filetype/package.rs

@@ -14,6 +14,15 @@ use crate::bundle::file::UserFile;
 use crate::bundle::filetype::BundleFileType;
 use crate::murmur::{HashGroup, IdString64, Murmur64};
 
+/// Resolves a relative path that might contain wildcards into a list of
+/// paths that exist on disk and match that wildcard.
+/// This is similar to globbing in Unix shells, but with much less features.
+///
+/// The only wilcard character allowed is `*`, and only at the end of the string,
+/// where it matches all files recursively in that directory.
+///
+/// `t` is an optional extension name, that may be used to force a wildcard
+/// path to only match that file type `t`.
 #[tracing::instrument]
 #[async_recursion]
 async fn resolve_wildcard<P1, P2>(

lib/sdk/src/murmur/dictionary.rs

@@ -147,14 +147,14 @@ impl Dictionary {
         Ok(())
     }
 
-    pub fn add(&mut self, value: String, group: HashGroup) {
+    pub fn add(&mut self, value: impl AsRef<[u8]>, group: HashGroup) {
-        let long = Murmur64::from(murmurhash64::hash(value.as_bytes(), SEED as u64));
+        let long = Murmur64::from(murmurhash64::hash(value.as_ref(), SEED as u64));
-        let short = Murmur32::from(murmurhash64::hash32(value.as_bytes(), SEED));
+        let short = Murmur32::from(murmurhash64::hash32(value.as_ref(), SEED));
 
         let entry = Entry {
             long,
             short,
-            value,
+            value: String::from_utf8_lossy(value.as_ref()).to_string(),
             group,
         };