Improve automatic symbol pairing for nameless literals (#247)

* Improve automatic symbol pairing for nameless literals

* Fix data reloc diffing when the reloc points to an in-function static symbol

* Only pair up literals that match perfectly

* Clippy

* Do two separate passes when pairing up literals

* Fix partially-matching literal pairups not working right

* Remove duplicate $ splitting code

* Implement $ splitting for section names too

* Minor cleanup

.cargo/config.toml

@@ -1,5 +1,4 @@
-[target.x86_64-pc-windows-msvc]
-linker = "rust-lld"
-
-[target.aarch64-pc-windows-msvc]
-linker = "rust-lld"
+# statically link the C runtime so the executable does not depend on
+# that shared/dynamic library.
+[target.'cfg(all(target_env = "msvc", target_os = "windows"))']
+rustflags = ["-C", "target-feature=+crt-static"]

.github/workflows/build.yaml

@@ -36,9 +36,9 @@ jobs:
       - name: Cache Rust workspace
         uses: Swatinem/rust-cache@v2
       - name: Cargo check
-        run: cargo check --all-targets --all-features
+        run: cargo check --all-targets --all-features --workspace
       - name: Cargo clippy
-        run: cargo clippy --all-targets --all-features
+        run: cargo clippy --all-targets --all-features --workspace
 
   fmt:
     name: Format
@@ -92,7 +92,7 @@ jobs:
       - name: Cache Rust workspace
         uses: Swatinem/rust-cache@v2
       - name: Cargo test
-        run: cargo test --release --features all
+        run: cargo test --release --all-features --workspace
 
   build-cli:
     name: Build objdiff-cli
@@ -146,13 +146,14 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
+      - name: Install uv
+        if: matrix.build == 'zigbuild'
+        uses: astral-sh/setup-uv@v6
       - name: Install cargo-zigbuild
         if: matrix.build == 'zigbuild'
         run: |
-          python3 -m venv .venv
-          . .venv/bin/activate
-          echo PATH=$PATH >> $GITHUB_ENV
-          pip install ziglang==0.13.0.post1 cargo-zigbuild==0.19.8
+          uv tool install cargo-zigbuild==0.20.1 --with-executables-from ziglang==0.15.1
+          echo "CARGO_ZIGBUILD_ZIG_PATH=$(uv tool dir)/cargo-zigbuild/bin/python-zig" >> $GITHUB_ENV
       - name: Setup Rust toolchain
         uses: dtolnay/rust-toolchain@stable
         with:
@@ -213,13 +214,14 @@ jobs:
           sudo apt-get -y install ${{ matrix.packages }}
       - name: Checkout
         uses: actions/checkout@v4
+      - name: Install uv
+        if: matrix.build == 'zigbuild'
+        uses: astral-sh/setup-uv@v6
       - name: Install cargo-zigbuild
         if: matrix.build == 'zigbuild'
         run: |
-          python3 -m venv .venv
-          . .venv/bin/activate
-          echo PATH=$PATH >> $GITHUB_ENV
-          pip install ziglang==0.13.0.post1 cargo-zigbuild==0.19.8
+          uv tool install cargo-zigbuild==0.20.1 --with-executables-from ziglang==0.15.1
+          echo "CARGO_ZIGBUILD_ZIG_PATH=$(uv tool dir)/cargo-zigbuild/bin/python-zig" >> $GITHUB_ENV
       - name: Setup Rust toolchain
         uses: dtolnay/rust-toolchain@stable
         with:

.pre-commit-config.yaml

@@ -24,7 +24,7 @@ repos:
         description: Run cargo clippy on all project files.
         language: system
         entry: cargo
-        args: ["+nightly", "clippy", "--all-targets", "--all-features"]
+        args: ["+nightly", "clippy", "--all-targets", "--all-features", "--workspace"]
         pass_filenames: false
       - id: cargo-deny
         name: cargo deny

Cargo.toml

@@ -5,18 +5,28 @@ members = [
     "objdiff-gui",
     "objdiff-wasm",
 ]
+default-members = [
+    "objdiff-cli",
+    "objdiff-core",
+    "objdiff-gui",
+    # Exclude objdiff-wasm by default
+]
 resolver = "3"
 
+[workspace.package]
+version = "3.0.1"
+authors = ["Luke Street <luke@street.dev>"]
+edition = "2024"
+license = "MIT OR Apache-2.0"
+repository = "https://github.com/encounter/objdiff"
+rust-version = "1.88"
+
 [profile.release-lto]
 inherits = "release"
 lto = "fat"
 strip = "debuginfo"
 codegen-units = 1
 
-[workspace.package]
-version = "3.0.0-beta.6"
-authors = ["Luke Street <luke@street.dev>"]
-edition = "2024"
-license = "MIT OR Apache-2.0"
-repository = "https://github.com/encounter/objdiff"
-rust-version = "1.85"
+[profile.release-min]
+inherits = "release-lto"
+opt-level = "z"

README.md

@@ -16,11 +16,12 @@ Features:
 
 Supports:
 
-- PowerPC 750CL (GameCube, Wii)
-- MIPS (N64, PS1, PS2, PSP)
-- x86 (COFF only at the moment)
 - ARM (GBA, DS, 3DS)
-- ARM64 (Switch, experimental)
+- ARM64 (Switch)
+- MIPS (N64, PS1, PS2, PSP)
+- PowerPC (GameCube, Wii, PS3, Xbox 360)
+- SuperH (Saturn, Dreamcast)
+- x86, x86_64 (PC)
 
 See [Usage](#usage) for more information.
 
@@ -89,22 +90,31 @@ file as well. You can then add `objdiff.json` to your `.gitignore` to prevent it
   "build_base": true,
   "watch_patterns": [
     "*.c",
+    "*.cc",
     "*.cp",
     "*.cpp",
     "*.cxx",
+    "*.c++",
     "*.h",
+    "*.hh",
     "*.hp",
     "*.hpp",
     "*.hxx",
+    "*.h++",
+    "*.pch",
+    "*.pch++",
+    "*.inc",
     "*.s",
     "*.S",
     "*.asm",
-    "*.inc",
     "*.py",
     "*.yml",
     "*.txt",
     "*.json"
   ],
+  "ignore_patterns": [
+    "build/**/*"
+  ],
   "units": [
     {
       "name": "main/MetroTRK/mslsupp",
@@ -140,6 +150,10 @@ It's unlikely you'll want to disable this, unless you're using an external tool
 If any of these files change, objdiff will automatically rebuild the objects and re-compare them.  
 If not specified, objdiff will use the default patterns listed above.
 
+`ignore_patterns` _(optional)_: A list of glob patterns to explicitly ignore when watching for changes.  
+([Supported syntax](https://docs.rs/globset/latest/globset/#syntax))  
+If not specified, objdiff will use the default patterns listed above.
+
 `units` _(optional)_: If specified, objdiff will display a list of objects in the sidebar for easy navigation.
 
 > `name` _(optional)_: The name of the object in the UI. If not specified, the object's `path` will be used.

config.schema.json

@@ -57,23 +57,39 @@
       },
       "default": [
         "*.c",
+        "*.cc",
         "*.cp",
         "*.cpp",
         "*.cxx",
+        "*.c++",
         "*.h",
+        "*.hh",
         "*.hp",
         "*.hpp",
         "*.hxx",
+        "*.h++",
+        "*.pch",
+        "*.pch++",
+        "*.inc",
         "*.s",
         "*.S",
         "*.asm",
-        "*.inc",
         "*.py",
         "*.yml",
         "*.txt",
         "*.json"
       ]
     },
+    "ignore_patterns": {
+      "type": "array",
+      "description": "List of glob patterns to explicitly ignore when watching for changes.\nFiles matching these patterns will not trigger a rebuild.\nSupported syntax: https://docs.rs/globset/latest/globset/#syntax",
+      "items": {
+        "type": "string"
+      },
+      "default": [
+        "build/**/*"
+      ]
+    },
     "objects": {
       "type": "array",
       "description": "Use units instead.",
@@ -175,6 +191,10 @@
           "type": "boolean",
           "description": "If true, objdiff will run the build command with the context file as an argument to generate it.",
           "default": false
+        },
+        "preset_id": {
+          "type": "number",
+          "description": "The decomp.me preset ID to use for the scratch.\nCompiler and flags in the config will take precedence over the preset, but the preset is useful for organizational purposes."
         }
       },
       "required": [

deny.toml

@@ -74,6 +74,7 @@ ignore = [
     #"a-crate-that-is-yanked@0.1.1", # you can also ignore yanked crate versions if you wish
     #{ crate = "a-crate-that-is-yanked@0.1.1", reason = "you can specify why you are ignoring the yanked crate" },
     { id = "RUSTSEC-2024-0436", reason = "Unmaintained paste crate is an indirect dependency" },
+    { id = "RUSTSEC-2025-0052", reason = "Unmaintained async-std crate is an indirect dependency" },
 ]
 # If this is true, then cargo deny will use the git executable to fetch advisory database.
 # If this is false, then it uses a built-in git library.
@@ -103,6 +104,7 @@ allow = [
     "0BSD",
     "OFL-1.1",
     "Ubuntu-font-1.0",
+    "CDLA-Permissive-2.0",
 ]
 # The confidence threshold for detecting a license from license text.
 # The higher the value, the more closely the license text must be to the
@@ -240,8 +242,8 @@ allow-git = []
 [sources.allow-org]
 # github.com organizations to allow git sources for
 github = [
-    "enarx", # flagset
     "encounter",
+    "gimli-rs",  # gimli
 ]
 # gitlab.com organizations to allow git sources for
 gitlab = []

objdiff-cli/Cargo.toml

@@ -15,11 +15,11 @@ publish = false
 [dependencies]
 anyhow = "1.0"
 argp = "0.4"
-crossterm = "0.28"
+crossterm = "0.29"
 enable-ansi-support = "0.2"
 memmap2 = "0.9"
 objdiff-core = { path = "../objdiff-core", features = ["all"] }
-prost = "0.13"
+prost = "0.14"
 ratatui = "0.29"
 rayon = "1.10"
 serde = { version = "1.0", features = ["derive"] }
@@ -28,7 +28,7 @@ supports-color = "3.0"
 time = { version = "0.3", features = ["formatting", "local-offset"] }
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }
-typed-path = "0.10"
+typed-path = "0.11"
 
 [target.'cfg(target_env = "musl")'.dependencies]
 mimalloc = "0.1"

objdiff-cli/src/cmd/diff.rs

@@ -19,7 +19,6 @@ use crossterm::{
     },
 };
 use objdiff_core::{
-    bindings::diff::DiffResult,
     build::{
         BuildConfig, BuildStatus,
         watcher::{Watcher, create_watcher},
@@ -28,7 +27,7 @@ use objdiff_core::{
         ProjectConfig, ProjectObject, ProjectObjectMetadata, build_globset,
         path::{check_path_buf, platform_path, platform_path_serde_option},
     },
-    diff::{self, DiffObjConfig, MappingConfig, ObjectDiff},
+    diff::{DiffObjConfig, MappingConfig, ObjectDiff},
     jobs::{
         Job, JobQueue, JobResult,
         objdiff::{ObjDiffConfig, start_build},
@@ -40,10 +39,7 @@ use typed_path::{Utf8PlatformPath, Utf8PlatformPathBuf};
 
 use crate::{
     cmd::apply_config_args,
-    util::{
-        output::{OutputFormat, write_output},
-        term::crossterm_panic_handler,
-    },
+    util::term::crossterm_panic_handler,
     views::{EventControlFlow, EventResult, UiView, function_diff::FunctionDiffUi},
 };
 
@@ -63,12 +59,6 @@ pub struct Args {
     #[argp(option, short = 'u')]
     /// Unit name within project
     unit: Option<String>,
-    #[argp(option, short = 'o', from_str_fn(platform_path))]
-    /// Output file (one-shot mode) ("-" for stdout)
-    output: Option<Utf8PlatformPathBuf>,
-    #[argp(option)]
-    /// Output format (json, json-pretty, proto) (default: json)
-    format: Option<String>,
     #[argp(positional)]
     /// Function symbol to diff
     symbol: Option<String>,
@@ -171,11 +161,7 @@ pub fn run(args: Args) -> Result<()> {
             _ => bail!("Either target and base or project and unit must be specified"),
         };
 
-    if let Some(output) = &args.output {
-        run_oneshot(&args, output, target_path.as_deref(), base_path.as_deref())
-    } else {
-        run_interactive(args, target_path, base_path, project_config)
-    }
+    run_interactive(args, target_path, base_path, project_config)
 }
 
 fn build_config_from_args(args: &Args) -> Result<(DiffObjConfig, MappingConfig)> {
@@ -194,32 +180,6 @@ fn build_config_from_args(args: &Args) -> Result<(DiffObjConfig, MappingConfig)>
     Ok((diff_config, mapping_config))
 }
 
-fn run_oneshot(
-    args: &Args,
-    output: &Utf8PlatformPath,
-    target_path: Option<&Utf8PlatformPath>,
-    base_path: Option<&Utf8PlatformPath>,
-) -> Result<()> {
-    let output_format = OutputFormat::from_option(args.format.as_deref())?;
-    let (diff_config, mapping_config) = build_config_from_args(args)?;
-    let target = target_path
-        .map(|p| {
-            obj::read::read(p.as_ref(), &diff_config).with_context(|| format!("Loading {}", p))
-        })
-        .transpose()?;
-    let base = base_path
-        .map(|p| {
-            obj::read::read(p.as_ref(), &diff_config).with_context(|| format!("Loading {}", p))
-        })
-        .transpose()?;
-    let result =
-        diff::diff_objs(target.as_ref(), base.as_ref(), None, &diff_config, &mapping_config)?;
-    let left = target.as_ref().and_then(|o| result.left.as_ref().map(|d| (o, d)));
-    let right = base.as_ref().and_then(|o| result.right.as_ref().map(|d| (o, d)));
-    write_output(&DiffResult::new(left, right), Some(output), output_format)?;
-    Ok(())
-}
-
 pub struct AppState {
     pub jobs: JobQueue,
     pub waker: Arc<TermWaker>,
@@ -382,10 +342,12 @@ fn run_interactive(
     };
     if let (Some(project_dir), Some(project_config)) = (&state.project_dir, &state.project_config) {
         let watch_patterns = project_config.build_watch_patterns()?;
+        let ignore_patterns = project_config.build_ignore_patterns()?;
         state.watcher = Some(create_watcher(
             state.modified.clone(),
             project_dir.as_ref(),
             build_globset(&watch_patterns)?,
+            build_globset(&ignore_patterns)?,
             Waker::from(state.waker.clone()),
         )?);
     }
@@ -399,7 +361,7 @@ fn run_interactive(
         stdout(),
         EnterAlternateScreen,
         EnableMouseCapture,
-        SetTitle(format!("{} - objdiff", symbol_name)),
+        SetTitle(format!("{symbol_name} - objdiff")),
     )?;
     let backend = CrosstermBackend::new(stdout());
     let mut terminal = Terminal::new(backend)?;

objdiff-cli/src/cmd/report.rs

@@ -85,6 +85,9 @@ pub fn run(args: Args) -> Result<()> {
 fn generate(args: GenerateArgs) -> Result<()> {
     let mut diff_config = diff::DiffObjConfig {
         function_reloc_diffs: diff::FunctionRelocDiffs::None,
+        combine_data_sections: true,
+        combine_text_sections: true,
+        ppc_calculate_pool_relocations: false,
         ..Default::default()
     };
     apply_config_args(&mut diff_config, &args.config)?;
@@ -174,16 +177,14 @@ fn report_object(
         .target_path
         .as_ref()
         .map(|p| {
-            obj::read::read(p.as_ref(), diff_config)
-                .with_context(|| format!("Failed to open {}", p))
+            obj::read::read(p.as_ref(), diff_config).with_context(|| format!("Failed to open {p}"))
         })
         .transpose()?;
     let base = object
         .base_path
         .as_ref()
         .map(|p| {
-            obj::read::read(p.as_ref(), diff_config)
-                .with_context(|| format!("Failed to open {}", p))
+            obj::read::read(p.as_ref(), diff_config).with_context(|| format!("Failed to open {p}"))
         })
         .transpose()?;
     let result =
@@ -225,6 +226,7 @@ fn report_object(
                 demangled_name: None,
                 virtual_address: section.virtual_address,
             }),
+            address: None,
         });
 
         match section.kind {
@@ -246,13 +248,12 @@ fn report_object(
             {
                 continue;
             }
-            if let Some(existing_functions) = &mut existing_functions {
-                if (symbol.flags.contains(SymbolFlag::Global)
+            if let Some(existing_functions) = &mut existing_functions
+                && (symbol.flags.contains(SymbolFlag::Global)
                     || symbol.flags.contains(SymbolFlag::Weak))
-                    && !existing_functions.insert(symbol.name.clone())
-                {
-                    continue;
-                }
+                && !existing_functions.insert(symbol.name.clone())
+            {
+                continue;
             }
             let match_percent = symbol_diff.match_percent.unwrap_or_else(|| {
                 // Support cases where we don't have a target object,
@@ -272,6 +273,7 @@ fn report_object(
                     demangled_name: symbol.demangled_name.clone(),
                     virtual_address: symbol.virtual_address,
                 }),
+                address: symbol.address.checked_sub(section.address),
             });
             if match_percent == 100.0 {
                 measures.matched_functions += 1;
@@ -279,6 +281,16 @@ fn report_object(
             measures.total_functions += 1;
         }
     }
+    sections.sort_by(|a, b| a.name.cmp(&b.name));
+    let reverse_fn_order = object.metadata.reverse_fn_order.unwrap_or(false);
+    functions.sort_by(|a, b| {
+        if reverse_fn_order {
+            b.address.unwrap_or(0).cmp(&a.address.unwrap_or(0))
+        } else {
+            a.address.unwrap_or(u64::MAX).cmp(&b.address.unwrap_or(u64::MAX))
+        }
+        .then_with(|| a.size.cmp(&b.size))
+    });
     if metadata.complete.unwrap_or(false) {
         measures.complete_code = measures.total_code;
         measures.complete_data = measures.total_data;
@@ -418,8 +430,8 @@ fn read_report(path: &Utf8PlatformPath) -> Result<Report> {
         std::io::stdin().read_to_end(&mut data)?;
         return Report::parse(&data).with_context(|| "Failed to load report from stdin");
     }
-    let file = File::open(path).with_context(|| format!("Failed to open {}", path))?;
+    let file = File::open(path).with_context(|| format!("Failed to open {path}"))?;
     let mmap =
-        unsafe { memmap2::Mmap::map(&file) }.with_context(|| format!("Failed to map {}", path))?;
-    Report::parse(mmap.as_ref()).with_context(|| format!("Failed to load report {}", path))
+        unsafe { memmap2::Mmap::map(&file) }.with_context(|| format!("Failed to map {path}"))?;
+    Report::parse(mmap.as_ref()).with_context(|| format!("Failed to load report {path}"))
 }

objdiff-cli/src/views/function_diff.rs

@@ -87,7 +87,7 @@ impl UiView for FunctionDiffUi {
             .and_then(|(_, _, d)| d.match_percent)
         {
             line_r.spans.push(Span::styled(
-                format!("{:.2}% ", percent),
+                format!("{percent:.2}% "),
                 Style::new().fg(match_percent_color(percent)),
             ));
         }
@@ -97,7 +97,7 @@ impl UiView for FunctionDiffUi {
             .and_then(|t| t.format(&state.time_format).ok())
             .unwrap_or_else(|| "N/A".to_string());
         line_r.spans.push(Span::styled(
-            format!("Last reload: {}", reload_time),
+            format!("Last reload: {reload_time}"),
             Style::new().fg(Color::White),
         ));
         line_r.spans.push(Span::styled(
@@ -170,32 +170,31 @@ impl UiView for FunctionDiffUi {
 
         let mut prev_text = None;
         let mut prev_margin_text = None;
-        if self.three_way {
-            if let Some((obj, symbol_idx, symbol_diff)) =
+        if self.three_way
+            && let Some((obj, symbol_idx, symbol_diff)) =
                 get_symbol(state.prev_obj.as_ref(), self.prev_sym)
-            {
-                let mut text = Text::default();
-                let rect = content_chunks[4].inner(Margin::new(0, 1));
-                self.print_sym(
-                    &mut text,
-                    obj,
-                    symbol_idx,
-                    symbol_diff,
-                    &state.diff_obj_config,
-                    rect,
-                    &self.right_highlight,
-                    result,
-                    true,
-                );
-                max_width = max_width.max(text.width());
-                prev_text = Some(text);
+        {
+            let mut text = Text::default();
+            let rect = content_chunks[4].inner(Margin::new(0, 1));
+            self.print_sym(
+                &mut text,
+                obj,
+                symbol_idx,
+                symbol_diff,
+                &state.diff_obj_config,
+                rect,
+                &self.right_highlight,
+                result,
+                true,
+            );
+            max_width = max_width.max(text.width());
+            prev_text = Some(text);
 
-                // Render margin
-                let mut text = Text::default();
-                let rect = content_chunks[3].inner(Margin::new(1, 1));
-                self.print_margin(&mut text, symbol_diff, rect);
-                prev_margin_text = Some(text);
-            }
+            // Render margin
+            let mut text = Text::default();
+            let rect = content_chunks[3].inner(Margin::new(1, 1));
+            self.print_margin(&mut text, symbol_diff, rect);
+            prev_margin_text = Some(text);
         }
 
         let max_scroll_x =
@@ -450,11 +449,11 @@ impl UiView for FunctionDiffUi {
 
     fn reload(&mut self, state: &AppState) -> Result<()> {
         let left_sym =
-            state.left_obj.as_ref().and_then(|(o, _)| find_function(o, &self.symbol_name));
+            state.left_obj.as_ref().and_then(|(o, _)| o.symbol_by_name(&self.symbol_name));
         let right_sym =
-            state.right_obj.as_ref().and_then(|(o, _)| find_function(o, &self.symbol_name));
+            state.right_obj.as_ref().and_then(|(o, _)| o.symbol_by_name(&self.symbol_name));
         let prev_sym =
-            state.prev_obj.as_ref().and_then(|(o, _)| find_function(o, &self.symbol_name));
+            state.prev_obj.as_ref().and_then(|(o, _)| o.symbol_by_name(&self.symbol_name));
         self.num_rows = match (
             get_symbol(state.left_obj.as_ref(), left_sym),
             get_symbol(state.right_obj.as_ref(), right_sym),
@@ -538,7 +537,7 @@ impl FunctionDiffUi {
                 let label_text = match segment.text {
                     DiffText::Basic(text) => text.to_string(),
                     DiffText::Line(num) => format!("{num} "),
-                    DiffText::Address(addr) => format!("{:x}:", addr),
+                    DiffText::Address(addr) => format!("{addr:x}:"),
                     DiffText::Opcode(mnemonic, _op) => format!("{mnemonic} "),
                     DiffText::Argument(arg) => arg.to_string(),
                     DiffText::BranchDest(addr) => format!("{addr:x}"),
@@ -546,7 +545,7 @@ impl FunctionDiffUi {
                         sym.demangled_name.as_ref().unwrap_or(&sym.name).clone()
                     }
                     DiffText::Addend(addend) => match addend.cmp(&0i64) {
-                        Ordering::Greater => format!("+{:#x}", addend),
+                        Ordering::Greater => format!("+{addend:#x}"),
                         Ordering::Less => format!("-{:#x}", -addend),
                         _ => String::new(),
                     },
@@ -561,15 +560,18 @@ impl FunctionDiffUi {
                 let len = label_text.len();
                 let highlighted =
                     highlight_kind != HighlightKind::None && *highlight == highlight_kind;
-                if let Some((cx, cy)) = result.click_xy {
-                    if cx >= sx && cx < sx + len as u16 && cy == sy {
-                        new_highlight = Some(highlight_kind);
-                    }
+                if let Some((cx, cy)) = result.click_xy
+                    && cx >= sx
+                    && cx < sx + len as u16
+                    && cy == sy
+                {
+                    new_highlight = Some(highlight_kind);
                 }
                 let mut style = Style::new().fg(match segment.color {
                     DiffTextColor::Normal => Color::Gray,
                     DiffTextColor::Dim => Color::DarkGray,
                     DiffTextColor::Bright => Color::White,
+                    DiffTextColor::DataFlow => Color::LightCyan,
                     DiffTextColor::Replace => Color::Cyan,
                     DiffTextColor::Delete => Color::Red,
                     DiffTextColor::Insert => Color::Green,
@@ -650,12 +652,3 @@ fn get_symbol(
     let sym = sym?;
     Some((obj, sym, &diff.symbols[sym]))
 }
-
-fn find_function(obj: &Object, name: &str) -> Option<usize> {
-    for (symbol_idx, symbol) in obj.symbols.iter().enumerate() {
-        if symbol.name == name {
-            return Some(symbol_idx);
-        }
-    }
-    None
-}

objdiff-core/Cargo.toml

@@ -41,6 +41,7 @@ any-arch = [
     "dep:regex",
     "dep:similar",
     "dep:syn",
+    "dep:encoding_rs"
 ]
 bindings = [
     "dep:prost",
@@ -61,7 +62,10 @@ config = [
     "dep:semver",
     "dep:typed-path",
 ]
-dwarf = ["dep:gimli"]
+dwarf = [
+    "dep:gimli",
+    "dep:typed-arena",
+]
 serde = [
     "dep:pbjson",
     "dep:pbjson-build",
@@ -77,6 +81,7 @@ std = [
     "prost?/std",
     "serde?/std",
     "similar?/std",
+    "typed-arena?/std",
     "typed-path?/std",
     "dep:filetime",
     "dep:memmap2",
@@ -91,7 +96,7 @@ ppc = [
     "any-arch",
     "dep:cwdemangle",
     "dep:cwextab",
-    "dep:ppc750cl",
+    "dep:powerpc",
     "dep:rlwinmdec",
 ]
 x86 = [
@@ -122,18 +127,18 @@ features = ["all"]
 [dependencies]
 anyhow = { version = "1.0", default-features = false }
 filetime = { version = "0.2", optional = true }
-flagset = { version = "0.4", default-features = false, optional = true, git = "https://github.com/enarx/flagset.git", rev = "a1fe9369b3741e43fec45da1998e83b9d78966a2" }
+flagset = { version = "0.4", default-features = false, optional = true }
 itertools = { version = "0.14", default-features = false, features = ["use_alloc"] }
 log = { version = "0.4", default-features = false, optional = true }
 memmap2 = { version = "0.9", optional = true }
 num-traits = { version = "0.2", default-features = false, optional = true }
-object = { git = "https://github.com/gimli-rs/object", rev = "a74579249e21ab8fcd3a86be588de336f18297cb", default-features = false, features = ["read_core", "elf", "pe"] }
-pbjson = { version = "0.7", default-features = false, optional = true }
-prost = { version = "0.13", default-features = false, features = ["prost-derive"], optional = true }
+object = { version = "0.37", default-features = false, features = ["read_core", "elf", "coff"] }
+pbjson = { version = "0.8", default-features = false, optional = true }
+prost = { version = "0.14", default-features = false, features = ["derive"], optional = true }
 regex = { version = "1.11", default-features = false, features = [], optional = true }
 serde = { version = "1.0", default-features = false, features = ["derive"], optional = true }
-similar = { version = "2.7", default-features = false, features = ["hashbrown"], optional = true, git = "https://github.com/encounter/similar.git", branch = "no_std" }
-typed-path = { version = "0.10", default-features = false, optional = true }
+similar = { git = "https://github.com/encounter/similar.git", branch = "no_std", default-features = false, features = ["hashbrown"], optional = true }
+typed-path = { version = "0.11", default-features = false, optional = true }
 
 # config
 globset = { version = "0.4", default-features = false, optional = true }
@@ -141,16 +146,17 @@ semver = { version = "1.0", default-features = false, optional = true }
 serde_json = { version = "1.0", default-features = false, features = ["alloc"], optional = true }
 
 # dwarf
-gimli = { version = "0.31", default-features = false, features = ["read"], optional = true }
+gimli = { git = "https://github.com/gimli-rs/gimli", rev = "7335f00e7c39fd501511584fefb0ba974117c950", default-features = false, features = ["read"], optional = true }
+typed-arena = { version = "2.0", default-features = false, optional = true }
 
 # ppc
 cwdemangle = { version = "1.0", optional = true }
-cwextab = { version = "1.0", optional = true }
-ppc750cl = { version = "0.3", optional = true }
+cwextab = { version = "1.1", optional = true }
+powerpc = { version = "0.4", optional = true }
 rlwinmdec = { version = "1.1", optional = true }
 
 # mips
-rabbitizer = { version = "2.0.0-alpha.1", default-features = false, features = ["all_extensions"], optional = true }
+rabbitizer = { version = "2.0.0-alpha.4", default-features = false, features = ["all_extensions"], optional = true }
 
 # x86
 cpp_demangle = { version = "0.4", default-features = false, features = ["alloc"], optional = true }
@@ -158,7 +164,7 @@ iced-x86 = { version = "1.21", default-features = false, features = ["decoder",
 msvc-demangler = { version = "0.11", optional = true }
 
 # arm
-unarm = { version = "1.7", optional = true }
+unarm = { version = "1.9", optional = true }
 arm-attr = { version = "0.2", optional = true }
 
 # arm64
@@ -166,15 +172,15 @@ yaxpeax-arch = { version = "0.3", default-features = false, optional = true }
 yaxpeax-arm = { version = "0.3", default-features = false, optional = true }
 
 # build
-notify = { version = "8.0.0", optional = true }
+notify = { version = "8.1.0", optional = true }
 notify-debouncer-full = { version = "0.5.0", optional = true }
 shell-escape = { version = "0.1", optional = true }
-tempfile = { version = "3.19", optional = true }
+tempfile = { version = "3.20", optional = true }
 time = { version = "0.3", optional = true }
-encoding_rs = "0.8.35"
+encoding_rs = { version = "0.8.35", optional = true }
 
 [target.'cfg(windows)'.dependencies]
-winapi = { version = "0.3", optional = true }
+winapi = { version = "0.3", optional = true, features = ["winbase"] }
 
 # For Linux static binaries, use rustls
 [target.'cfg(target_os = "linux")'.dependencies]
@@ -188,10 +194,10 @@ self_update = { version = "0.42", optional = true }
 
 [build-dependencies]
 heck = { version = "0.5", optional = true }
-pbjson-build = { version = "0.7", optional = true }
+pbjson-build = { version = "0.8", optional = true }
 prettyplease = { version = "0.2", optional = true }
 proc-macro2 = { version = "1.0", optional = true }
-prost-build = { version = "0.13", optional = true }
+prost-build = { version = "0.14", optional = true }
 quote = { version = "1.0", optional = true }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = { version = "1.0" }
@@ -200,4 +206,4 @@ syn = { version = "2.0", optional = true }
 [dev-dependencies]
 # Enable all features for tests
 objdiff-core = { path = ".", features = ["all"] }
-insta = "1.42"
+insta = "1.43"

objdiff-core/README.md

@@ -5,11 +5,12 @@ objdiff-core contains the core functionality of [objdiff](https://github.com/enc
 ## Crate feature flags
 
 - **`all`**: Enables all main features.
+- **`bindings`**: Enables serialization and deserialization of objdiff data structures.
 - **`config`**: Enables objdiff configuration file support.
 - **`dwarf`**: Enables extraction of line number information from DWARF debug sections.
-- **`mips`**: Enables the MIPS backend powered by [rabbitizer](https://github.com/Decompollaborate/rabbitizer). (Note: C library with Rust bindings)
-- **`ppc`**: Enables the PowerPC backend powered by [ppc750cl](https://github.com/encounter/ppc750cl).
-- **`x86`**: Enables the x86 backend powered by [iced-x86](https://crates.io/crates/iced-x86).
-- **`arm`**: Enables the ARM backend powered by [unarm](https://github.com/AetiasHax/unarm).
 - **`arm64`**: Enables the ARM64 backend powered by [yaxpeax-arm](https://github.com/iximeow/yaxpeax-arm).
-- **`bindings`**: Enables serialization and deserialization of objdiff data structures.
+- **`arm`**: Enables the ARM backend powered by [unarm](https://github.com/AetiasHax/unarm).
+- **`mips`**: Enables the MIPS backend powered by [rabbitizer](https://github.com/Decompollaborate/rabbitizer).
+- **`ppc`**: Enables the PowerPC backend powered by [powerpc](https://github.com/encounter/powerpc-rs).
+- **`superh`**: Enables the SuperH backend powered by an included disassembler.
+- **`x86`**: Enables the x86 backend powered by [iced-x86](https://crates.io/crates/iced-x86).

objdiff-core/config-schema.json

@@ -25,6 +25,20 @@
         }
       ]
     },
+    {
+      "id": "analyzeDataFlow",
+      "type": "boolean",
+      "default": false,
+      "name": "(Experimental) Perform data flow analysis",
+      "description": "Use data flow analysis to display known information about register contents where possible"
+    },
+    {
+      "id": "showDataFlow",
+      "type": "boolean",
+      "default": true,
+      "name": "Show data flow",
+      "description": "Show data flow analysis results in place of register name where present"
+    },
     {
       "id": "spaceBetweenArgs",
       "type": "boolean",
@@ -264,7 +278,8 @@
       "id": "ppc",
       "name": "PowerPC",
       "properties": [
-        "ppc.calculatePoolRelocations"
+        "ppc.calculatePoolRelocations",
+        "analyzeDataFlow"
       ]
     },
     {

objdiff-core/config_gen.rs

@@ -60,10 +60,10 @@ pub struct ConfigGroup {
 }
 
 fn build_doc(name: &str, description: Option<&str>) -> TokenStream {
-    let mut doc = format!(" {}", name);
+    let mut doc = format!(" {name}");
     let mut out = quote! { #[doc = #doc] };
     if let Some(description) = description {
-        doc = format!(" {}", description);
+        doc = format!(" {description}");
         out.extend(quote! { #[doc = ""] });
         out.extend(quote! { #[doc = #doc] });
     }
@@ -443,9 +443,9 @@ pub fn generate_diff_config() {
         }
         impl core::fmt::Display for ConfigPropertyValue {
             fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
-                match self {
-                    ConfigPropertyValue::Boolean(value) => write!(f, "{}", value),
-                    ConfigPropertyValue::Choice(value) => write!(f, "{}", value),
+                match *self {
+                    ConfigPropertyValue::Boolean(value) => write!(f, "{value}"),
+                    ConfigPropertyValue::Choice(value) => f.write_str(value),
                 }
             }
         }

objdiff-core/protos/diff.proto

@@ -1,165 +0,0 @@
-syntax = "proto3";
-
-package objdiff.diff;
-
-// A symbol
-message Symbol {
-  // Name of the symbol
-  string name = 1;
-  // Demangled name of the symbol
-  optional string demangled_name = 2;
-  // Symbol address
-  uint64 address = 3;
-  // Symbol size
-  uint64 size = 4;
-  // Bitmask of SymbolFlag
-  uint32 flags = 5;
-}
-
-// Symbol visibility flags
-enum SymbolFlag {
-  SYMBOL_NONE = 0;
-  SYMBOL_GLOBAL = 1;
-  SYMBOL_LOCAL = 2;
-  SYMBOL_WEAK = 4;
-  SYMBOL_COMMON = 8;
-  SYMBOL_HIDDEN = 16;
-}
-
-// A single parsed instruction
-message Instruction {
-  // Instruction address
-  uint64 address = 1;
-  // Instruction size
-  uint32 size = 2;
-  // Instruction opcode
-  uint32 opcode = 3;
-  // Instruction mnemonic
-  string mnemonic = 4;
-  // Instruction formatted string
-  string formatted = 5;
-  // Original (unsimplified) instruction string
-  optional string original = 6;
-  // Instruction arguments
-  repeated Argument arguments = 7;
-  // Instruction relocation
-  optional Relocation relocation = 8;
-  // Instruction branch destination
-  optional uint64 branch_dest = 9;
-  // Instruction line number
-  optional uint32 line_number = 10;
-}
-
-// An instruction argument
-message Argument {
-  oneof value {
-    // Plain text
-    string plain_text = 1;
-    // Value
-    ArgumentValue argument = 2;
-    // Relocation
-    ArgumentRelocation relocation = 3;
-    // Branch destination
-    uint64 branch_dest = 4;
-  }
-}
-
-// An instruction argument value
-message ArgumentValue {
-  oneof value {
-    // Signed integer
-    int64 signed = 1;
-    // Unsigned integer
-    uint64 unsigned = 2;
-    // Opaque value
-    string opaque = 3;
-  }
-}
-
-// Marker type for relocation arguments
-message ArgumentRelocation {
-}
-
-message Relocation {
-  uint32 type = 1;
-  string type_name = 2;
-  RelocationTarget target = 3;
-}
-
-message RelocationTarget {
-  uint32 symbol_index = 1;
-  int64 addend = 2;
-}
-
-message InstructionDiffRow {
-  DiffKind diff_kind = 1;
-  optional Instruction instruction = 2;
-  optional InstructionBranchFrom branch_from = 3;
-  optional InstructionBranchTo branch_to = 4;
-  repeated ArgumentDiff arg_diff = 5;
-}
-
-message ArgumentDiff {
-  optional uint32 diff_index = 1;
-}
-
-enum DiffKind {
-  DIFF_NONE = 0;
-  DIFF_REPLACE = 1;
-  DIFF_DELETE = 2;
-  DIFF_INSERT = 3;
-  DIFF_OP_MISMATCH = 4;
-  DIFF_ARG_MISMATCH = 5;
-}
-
-message InstructionBranchFrom {
-  repeated uint32 instruction_index = 1;
-  uint32 branch_index = 2;
-}
-
-message InstructionBranchTo {
-  uint32 instruction_index = 1;
-  uint32 branch_index = 2;
-}
-
-message SymbolDiff {
-  Symbol symbol = 1;
-  repeated InstructionDiffRow instruction_rows = 2;
-  optional float match_percent = 3;
-  // The symbol index in the _other_ object that this symbol was diffed against
-  optional uint32 target_symbol = 5;
-}
-
-message DataDiff {
-  DiffKind kind = 1;
-  bytes data = 2;
-  // May be larger than data
-  uint64 size = 3;
-}
-
-message SectionDiff {
-  string name = 1;
-  SectionKind kind = 2;
-  uint64 size = 3;
-  uint64 address = 4;
-  reserved 5;
-  repeated DataDiff data = 6;
-  optional float match_percent = 7;
-}
-
-enum SectionKind {
-  SECTION_UNKNOWN = 0;
-  SECTION_TEXT = 1;
-  SECTION_DATA = 2;
-  SECTION_BSS = 3;
-}
-
-message ObjectDiff {
-  repeated SectionDiff sections = 1;
-  repeated SymbolDiff symbols = 2;
-}
-
-message DiffResult {
-  optional ObjectDiff left = 1;
-  optional ObjectDiff right = 2;
-}

objdiff-core/protos/report.proto

@@ -99,6 +99,8 @@ message ReportItem {
   float fuzzy_match_percent = 3;
   // Extra metadata for this item
   optional ReportItemMetadata metadata = 4;
+  // Address of the item (section-relative offset)
+  optional uint64 address = 5;
 }
 
 // Extra metadata for an item

objdiff-core/src/arch/arm.rs

@@ -11,11 +11,11 @@ use object::{Endian as _, Object as _, ObjectSection as _, ObjectSymbol as _, el
 use unarm::{args, arm, thumb};
 
 use crate::{
-    arch::Arch,
+    arch::{Arch, RelocationOverride, RelocationOverrideTarget},
     diff::{ArmArchVersion, ArmR9Usage, DiffObjConfig, display::InstructionPart},
     obj::{
         InstructionRef, Relocation, RelocationFlags, ResolvedInstructionRef, ResolvedRelocation,
-        ScannedInstruction, Section, SectionKind, Symbol, SymbolFlag, SymbolFlagSet, SymbolKind,
+        Section, SectionKind, Symbol, SymbolFlag, SymbolFlagSet, SymbolKind,
     },
 };
 
@@ -187,14 +187,14 @@ impl Arch for ArchArm {
         self.disasm_modes = Self::get_mapping_symbols(sections, symbols);
     }
 
-    fn scan_instructions(
+    fn scan_instructions_internal(
         &self,
         address: u64,
         code: &[u8],
         section_index: usize,
         _relocations: &[Relocation],
         diff_config: &DiffObjConfig,
-    ) -> Result<Vec<ScannedInstruction>> {
+    ) -> Result<Vec<InstructionRef>> {
         let start_addr = address as u32;
         let end_addr = start_addr + code.len() as u32;
 
@@ -219,13 +219,13 @@ impl Arch for ArchArm {
         let mut next_mapping = mappings_iter.next();
 
         let ins_count = code.len() / mode.instruction_size(start_addr);
-        let mut ops = Vec::<ScannedInstruction>::with_capacity(ins_count);
+        let mut ops = Vec::<InstructionRef>::with_capacity(ins_count);
 
         let parse_flags = self.parse_flags(diff_config);
 
         let mut address = start_addr;
         while address < end_addr {
-            while let Some(next) = next_mapping.take_if(|x| address >= x.address) {
+            while let Some(next) = next_mapping.filter(|x| address >= x.address) {
                 // Change mapping
                 mode = next.mapping;
                 next_mapping = mappings_iter.next();
@@ -235,12 +235,10 @@ impl Arch for ArchArm {
             let data = &code[(address - start_addr) as usize..];
             if data.len() < ins_size {
                 // Push the remainder as data
-                ops.push(ScannedInstruction {
-                    ins_ref: InstructionRef {
-                        address: address as u64,
-                        size: data.len() as u8,
-                        opcode: u16::MAX,
-                    },
+                ops.push(InstructionRef {
+                    address: address as u64,
+                    size: data.len() as u8,
+                    opcode: u16::MAX,
                     branch_dest: None,
                 });
                 break;
@@ -256,12 +254,10 @@ impl Arch for ArchArm {
                 }
                 _ => {
                     // Invalid instruction size
-                    ops.push(ScannedInstruction {
-                        ins_ref: InstructionRef {
-                            address: address as u64,
-                            size: ins_size as u8,
-                            opcode: u16::MAX,
-                        },
+                    ops.push(InstructionRef {
+                        address: address as u64,
+                        size: ins_size as u8,
+                        opcode: u16::MAX,
                         branch_dest: None,
                     });
                     address += ins_size as u32;
@@ -325,8 +321,10 @@ impl Arch for ArchArm {
                 unarm::ParseMode::Data => (u16::MAX, None),
             };
 
-            ops.push(ScannedInstruction {
-                ins_ref: InstructionRef { address: address as u64, size: ins_size as u8, opcode },
+            ops.push(InstructionRef {
+                address: address as u64,
+                size: ins_size as u8,
+                opcode,
                 branch_dest: branch_dest.map(|x| x as u64),
             });
             address += ins_size as u32;
@@ -358,47 +356,57 @@ impl Arch for ArchArm {
         Ok(())
     }
 
-    fn implcit_addend(
+    fn relocation_override(
         &self,
         _file: &object::File<'_>,
         section: &object::Section,
         address: u64,
-        _relocation: &object::Relocation,
-        flags: RelocationFlags,
-    ) -> Result<i64> {
-        let section_data = section.data()?;
-        let address = address as usize;
-        Ok(match flags {
-            // ARM calls
-            RelocationFlags::Elf(elf::R_ARM_PC24)
-            | RelocationFlags::Elf(elf::R_ARM_XPC25)
-            | RelocationFlags::Elf(elf::R_ARM_CALL) => {
-                let data = section_data[address..address + 4].try_into()?;
-                let addend = self.endianness.read_i32_bytes(data);
-                let imm24 = addend & 0xffffff;
-                (imm24 << 2) << 8 >> 8
+        relocation: &object::Relocation,
+    ) -> Result<Option<RelocationOverride>> {
+        match relocation.flags() {
+            // Handle ELF implicit relocations
+            object::RelocationFlags::Elf { r_type } => {
+                if relocation.has_implicit_addend() {
+                    let section_data = section.data()?;
+                    let address = address as usize;
+                    let addend = match r_type {
+                        // ARM calls
+                        elf::R_ARM_PC24 | elf::R_ARM_XPC25 | elf::R_ARM_CALL => {
+                            let data = section_data[address..address + 4].try_into()?;
+                            let addend = self.endianness.read_i32_bytes(data);
+                            let imm24 = addend & 0xffffff;
+                            (imm24 << 2) << 8 >> 8
+                        }
+
+                        // Thumb calls
+                        elf::R_ARM_THM_PC22 | elf::R_ARM_THM_XPC22 => {
+                            let data = section_data[address..address + 2].try_into()?;
+                            let high = self.endianness.read_i16_bytes(data) as i32;
+                            let data = section_data[address + 2..address + 4].try_into()?;
+                            let low = self.endianness.read_i16_bytes(data) as i32;
+
+                            let imm22 = ((high & 0x7ff) << 11) | (low & 0x7ff);
+                            (imm22 << 1) << 9 >> 9
+                        }
+
+                        // Data
+                        elf::R_ARM_ABS32 => {
+                            let data = section_data[address..address + 4].try_into()?;
+                            self.endianness.read_i32_bytes(data)
+                        }
+
+                        flags => bail!("Unsupported ARM implicit relocation {flags:?}"),
+                    };
+                    Ok(Some(RelocationOverride {
+                        target: RelocationOverrideTarget::Keep,
+                        addend: addend as i64,
+                    }))
+                } else {
+                    Ok(None)
+                }
             }
-
-            // Thumb calls
-            RelocationFlags::Elf(elf::R_ARM_THM_PC22)
-            | RelocationFlags::Elf(elf::R_ARM_THM_XPC22) => {
-                let data = section_data[address..address + 2].try_into()?;
-                let high = self.endianness.read_i16_bytes(data) as i32;
-                let data = section_data[address + 2..address + 4].try_into()?;
-                let low = self.endianness.read_i16_bytes(data) as i32;
-
-                let imm22 = ((high & 0x7ff) << 11) | (low & 0x7ff);
-                (imm22 << 1) << 9 >> 9
-            }
-
-            // Data
-            RelocationFlags::Elf(elf::R_ARM_ABS32) => {
-                let data = section_data[address..address + 4].try_into()?;
-                self.endianness.read_i32_bytes(data)
-            }
-
-            flags => bail!("Unsupported ARM implicit relocation {flags:?}"),
-        } as i64)
+            _ => Ok(None),
+        }
     }
 
     fn demangle(&self, name: &str) -> Option<String> {
@@ -456,6 +464,22 @@ impl Arch for ArchArm {
         }
         flags
     }
+
+    fn infer_function_size(
+        &self,
+        symbol: &Symbol,
+        section: &Section,
+        mut next_address: u64,
+    ) -> Result<u64> {
+        // Trim any trailing 4-byte zeroes from the end (padding)
+        while next_address >= symbol.address + 4
+            && let Some(data) = section.data_range(next_address - 4, 4)
+            && data == [0u8; 4]
+        {
+            next_address -= 4;
+        }
+        Ok(next_address.saturating_sub(symbol.address))
+    }
 }
 
 #[derive(Clone, Copy, Debug)]
@@ -577,7 +601,7 @@ fn push_args(
                     arg_cb(InstructionPart::basic("}"))?;
                 }
                 args::Argument::CoprocNum(value) => {
-                    arg_cb(InstructionPart::opaque(format!("p{}", value)))?;
+                    arg_cb(InstructionPart::opaque(format!("p{value}")))?;
                 }
                 args::Argument::ShiftImm(shift) => {
                     arg_cb(InstructionPart::opaque(shift.op.to_string()))?;