Version 3.4.4

* Copy strings without escape sequences in them

* Add support for more encodings from encoding_rs

Also use encoding_rs instead of CStr for UTF-8.

AGENTS.md

@@ -0,0 +1,44 @@
+# Repository Guidelines
+
+## Project Structure & Module Organization
+
+- `objdiff-core`: core library with diffing logic and shared components
+- `objdiff-cli`: CLI/TUI with `ratatui`
+- `objdiff-gui`: GUI with `eframe`/`egui`
+- `objdiff-wasm`: WebAssembly bindings
+
+objdiff has three main frontends: GUI, CLI/TUI, and [web](https://github.com/encounter/objdiff-web) (utilizing `objdiff-wasm`).
+
+`objdiff-gui` is the most fully-featured and is the primary target for new features.
+
+`objdiff-cli` has an interactive TUI `diff` mode (powered by `ratatui`) and a non-interactive `report` mode to generate detailed progress reports (e.g. for <https://decomp.dev>).
+
+`objdiff-wasm` is compiled into a [WebAssembly Component](https://component-model.bytecodealliance.org/). The [API](objdiff-wasm/wit/objdiff.wit) is defined using [WIT](https://component-model.bytecodealliance.org/design/wit.html). The web interface is separate, but mirrors the GUI in functionality.
+
+## Build, Test, and Development Commands
+
+Run `cargo build` for a debug build and `cargo build --release` for an optimized build.
+The CLI can be exercised with `cargo run --release --bin objdiff-cli -- --help`.
+
+The WASM build is not included in the workspace by default due to differences in toolchain and target; build it with `npm -C objdiff-wasm run build`. (nightly required)
+In general, do NOT use `--workspace`, as it will include the WASM crate and likely fail. Formatting and linting commands are exempt.
+
+Pre-commit tasks:
+
+- `cargo test` to run the test suite
+- `cargo +nightly fmt --all` to format code (nightly required)
+- `cargo +nightly clippy --all-targets --all-features --workspace -- -D warnings` to lint code (nightly required)
+- `cargo deny check` (`cargo install --locked cargo-deny` if needed) if dependencies change
+
+## Testing Guidelines
+
+Favor focused unit tests adjacent to the code, and integration scenarios in `objdiff-core/tests`.
+Integration tests utilize snapshots with the `insta` crate; run `curl -LsSf https://insta.rs/install.sh | sh` if needed.
+To generate updated snapshots, use `cargo insta test`, review diffs manually, then accept changes with `cargo insta accept` (or simply `mv` to accept specific snapshots).
+
+## Configuration Tips
+
+- `config.schema.json`: JSON schema for the user-facing project config file (`objdiff.json`)
+- `objdiff-core/config-schema.json`: Internal options schema to generate `DiffObjConfig` (see `objdiff-core/config_gen.rs`)
+
+The project configuration (`objdiff.json`) is intended to be generated by the projects' build system. It includes paths to the target (expected) and base (current) object files, along with build commands and file watch patterns. See `README.md` for a summary of key options.

Cargo.toml

@@ -14,7 +14,7 @@ default-members = [
 resolver = "3"
 
 [workspace.package]
-version = "3.2.0"
+version = "3.4.4"
 authors = ["Luke Street <luke@street.dev>"]
 edition = "2024"
 license = "MIT OR Apache-2.0"

config.schema.json

@@ -111,6 +111,25 @@
       "items": {
         "$ref": "#/$defs/progress_category"
       }
+    },
+    "options": {
+      "type": "object",
+      "description": "Diff configuration options that should be applied automatically when the project is loaded.",
+      "additionalProperties": {
+        "oneOf": [
+          {
+            "type": "boolean"
+          },
+          {
+            "type": "string"
+          }
+        ]
+      },
+      "examples": [
+        {
+          "demangler": "gnu_legacy"
+        }
+      ]
     }
   },
   "$defs": {
@@ -156,6 +175,20 @@
           "additionalProperties": {
             "type": "string"
           }
+        },
+        "options": {
+          "type": "object",
+          "description": "Diff configuration options that should be applied when this unit is active.",
+          "additionalProperties": {
+            "oneOf": [
+              {
+                "type": "boolean"
+              },
+              {
+                "type": "string"
+              }
+            ]
+          }
         }
       }
     },

deny.toml

@@ -75,6 +75,7 @@ ignore = [
     #{ 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" },
+    { id = "RUSTSEC-2025-0119", reason = "Unmaintained number_prefix 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.

objdiff-cli/Cargo.toml

@@ -16,19 +16,19 @@ publish = false
 anyhow = "1.0"
 argp = "0.4"
 crossterm = "0.29"
-enable-ansi-support = "0.2"
+enable-ansi-support = "0.3"
 memmap2 = "0.9"
 objdiff-core = { path = "../objdiff-core", features = ["all"] }
 prost = "0.14"
 ratatui = "0.29"
-rayon = "1.10"
+rayon = "1.11"
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 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 = "0.11"
+typed-path = "0.12"
 
 [target.'cfg(target_env = "musl")'.dependencies]
 mimalloc = "0.1"

objdiff-cli/src/cmd/diff.rs

@@ -24,7 +24,8 @@ use objdiff_core::{
         watcher::{Watcher, create_watcher},
     },
     config::{
-        ProjectConfig, ProjectObject, ProjectObjectMetadata, build_globset,
+        ProjectConfig, ProjectObject, ProjectObjectMetadata, ProjectOptions, apply_project_options,
+        build_globset,
         path::{check_path_buf, platform_path, platform_path_serde_option},
     },
     diff::{DiffObjConfig, MappingConfig, ObjectDiff},
@@ -65,23 +66,14 @@ pub struct Args {
     #[argp(option, short = 'c')]
     /// Configuration property (key=value)
     config: Vec<String>,
-    #[argp(option, short = 'm')]
-    /// Symbol mapping (target=base)
-    mapping: Vec<String>,
-    #[argp(option)]
-    /// Left symbol name for selection
-    selecting_left: Option<String>,
-    #[argp(option)]
-    /// Right symbol name for selection
-    selecting_right: Option<String>,
 }
 
 pub fn run(args: Args) -> Result<()> {
-    let (target_path, base_path, project_config) =
+    let (target_path, base_path, project_config, unit_options) =
         match (&args.target, &args.base, &args.project, &args.unit) {
             (Some(_), Some(_), None, None)
             | (Some(_), None, None, None)
-            | (None, Some(_), None, None) => (args.target.clone(), args.base.clone(), None),
+            | (None, Some(_), None, None) => (args.target.clone(), args.base.clone(), None, None),
             (None, None, p, u) => {
                 let project = match p {
                     Some(project) => project.clone(),
@@ -106,28 +98,32 @@ pub fn run(args: Args) -> Result<()> {
                     .base_dir
                     .as_ref()
                     .map(|p| project.join(p.with_platform_encoding()));
-                let objects = project_config
+                let units = project_config.units.as_deref().unwrap_or_default();
-                    .units
+                let objects = units
                     .iter()
-                    .flatten()
+                    .enumerate()
-                    .map(|o| {
+                    .map(|(idx, o)| {
-                        ObjectConfig::new(
+                        (
-                            o,
+                            ObjectConfig::new(
-                            &project,
+                                o,
-                            target_obj_dir.as_deref(),
+                                &project,
-                            base_obj_dir.as_deref(),
+                                target_obj_dir.as_deref(),
+                                base_obj_dir.as_deref(),
+                            ),
+                            idx,
                         )
                     })
                     .collect::<Vec<_>>();
-                let object = if let Some(u) = u {
+                let (object, unit_idx) = if let Some(u) = u {
                     objects
                         .iter()
-                        .find(|obj| obj.name == *u)
+                        .find(|(obj, _)| obj.name == *u)
+                        .map(|(obj, idx)| (obj, *idx))
                         .ok_or_else(|| anyhow!("Unit not found: {}", u))?
                 } else if let Some(symbol_name) = &args.symbol {
                     let mut idx = None;
                     let mut count = 0usize;
-                    for (i, obj) in objects.iter().enumerate() {
+                    for (i, (obj, unit_idx)) in objects.iter().enumerate() {
                         if obj
                             .target_path
                             .as_deref()
@@ -135,7 +131,7 @@ pub fn run(args: Args) -> Result<()> {
                             .transpose()?
                             .unwrap_or(false)
                         {
-                            idx = Some(i);
+                            idx = Some((i, *unit_idx));
                             count += 1;
                             if count > 1 {
                                 break;
@@ -144,7 +140,7 @@ pub fn run(args: Args) -> Result<()> {
                     }
                     match (count, idx) {
                         (0, None) => bail!("Symbol not found: {}", symbol_name),
-                        (1, Some(i)) => &objects[i],
+                        (1, Some((i, unit_idx))) => (&objects[i].0, unit_idx),
                         (2.., Some(_)) => bail!(
                             "Multiple instances of {} were found, try specifying a unit",
                             symbol_name
@@ -154,30 +150,31 @@ pub fn run(args: Args) -> Result<()> {
                 } else {
                     bail!("Must specify one of: symbol, project and unit, target and base objects")
                 };
+                let unit_options = units.get(unit_idx).and_then(|u| u.options().cloned());
                 let target_path = object.target_path.clone();
                 let base_path = object.base_path.clone();
-                (target_path, base_path, Some(project_config))
+                (target_path, base_path, Some(project_config), unit_options)
             }
             _ => bail!("Either target and base or project and unit must be specified"),
         };
 
-    run_interactive(args, target_path, base_path, project_config)
+    run_interactive(args, target_path, base_path, project_config, unit_options)
 }
 
-fn build_config_from_args(args: &Args) -> Result<(DiffObjConfig, MappingConfig)> {
+fn build_config_from_args(
+    args: &Args,
+    project_config: Option<&ProjectConfig>,
+    unit_options: Option<&ProjectOptions>,
+) -> Result<(DiffObjConfig, MappingConfig)> {
     let mut diff_config = DiffObjConfig::default();
-    apply_config_args(&mut diff_config, &args.config)?;
+    if let Some(options) = project_config.and_then(|config| config.options.as_ref()) {
-    let mut mapping_config = MappingConfig {
+        apply_project_options(&mut diff_config, options)?;
-        mappings: Default::default(),
-        selecting_left: args.selecting_left.clone(),
-        selecting_right: args.selecting_right.clone(),
-    };
-    for mapping in &args.mapping {
-        let (target, base) =
-            mapping.split_once('=').context("--mapping expects \"target=base\"")?;
-        mapping_config.mappings.insert(target.to_string(), base.to_string());
     }
-    Ok((diff_config, mapping_config))
+    if let Some(options) = unit_options {
+        apply_project_options(&mut diff_config, options)?;
+    }
+    apply_config_args(&mut diff_config, &args.config)?;
+    Ok((diff_config, MappingConfig::default()))
 }
 
 pub struct AppState {
@@ -316,11 +313,13 @@ fn run_interactive(
     target_path: Option<Utf8PlatformPathBuf>,
     base_path: Option<Utf8PlatformPathBuf>,
     project_config: Option<ProjectConfig>,
+    unit_options: Option<ProjectOptions>,
 ) -> Result<()> {
     let Some(symbol_name) = &args.symbol else { bail!("Interactive mode requires a symbol name") };
     let time_format = time::format_description::parse_borrowed::<2>("[hour]:[minute]:[second]")
         .context("Failed to parse time format")?;
-    let (diff_obj_config, mapping_config) = build_config_from_args(&args)?;
+    let (diff_obj_config, mapping_config) =
+        build_config_from_args(&args, project_config.as_ref(), unit_options.as_ref())?;
     let mut state = AppState {
         jobs: Default::default(),
         waker: Default::default(),

objdiff-cli/src/cmd/report.rs

@@ -7,7 +7,7 @@ use objdiff_core::{
         ChangeItem, ChangeItemInfo, ChangeUnit, Changes, ChangesInput, Measures, REPORT_VERSION,
         Report, ReportCategory, ReportItem, ReportItemMetadata, ReportUnit, ReportUnitMetadata,
     },
-    config::path::platform_path,
+    config::{ProjectObject, ProjectOptions, apply_project_options, path::platform_path},
     diff,
     obj::{self, SectionKind, SymbolFlag, SymbolKind},
 };
@@ -83,14 +83,13 @@ pub fn run(args: Args) -> Result<()> {
 }
 
 fn generate(args: GenerateArgs) -> Result<()> {
-    let mut diff_config = diff::DiffObjConfig {
+    let base_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)?;
 
     let output_format = OutputFormat::from_option(args.format.as_deref())?;
     let project_dir = args.project.as_deref().unwrap_or_else(|| Utf8PlatformPath::new("."));
@@ -101,31 +100,44 @@ fn generate(args: GenerateArgs) -> Result<()> {
         Some((Err(err), _)) => bail!("Failed to load project configuration: {}", err),
         None => bail!("No project configuration found"),
     };
-    info!(
-        "Generating report for {} units (using {} threads)",
-        project.units().len(),
-        if args.deduplicate { 1 } else { rayon::current_num_threads() }
-    );
-
     let target_obj_dir =
         project.target_dir.as_ref().map(|p| project_dir.join(p.with_platform_encoding()));
     let base_obj_dir =
         project.base_dir.as_ref().map(|p| project_dir.join(p.with_platform_encoding()));
-    let objects = project
+    let project_units = project.units.as_deref().unwrap_or_default();
-        .units
+    let objects = project_units
         .iter()
-        .flatten()
+        .enumerate()
-        .map(|o| {
+        .map(|(idx, o)| {
-            ObjectConfig::new(o, project_dir, target_obj_dir.as_deref(), base_obj_dir.as_deref())
+            (
+                ObjectConfig::new(
+                    o,
+                    project_dir,
+                    target_obj_dir.as_deref(),
+                    base_obj_dir.as_deref(),
+                ),
+                idx,
+            )
         })
         .collect::<Vec<_>>();
+    info!(
+        "Generating report for {} units (using {} threads)",
+        objects.len(),
+        if args.deduplicate { 1 } else { rayon::current_num_threads() }
+    );
 
     let start = Instant::now();
     let mut units = vec![];
     let mut existing_functions: HashSet<String> = HashSet::new();
     if args.deduplicate {
         // If deduplicating, we need to run single-threaded
-        for object in &objects {
+        for (object, unit_idx) in &objects {
+            let diff_config = build_unit_diff_config(
+                &base_diff_config,
+                project.options.as_ref(),
+                project_units.get(*unit_idx).and_then(ProjectObject::options),
+                &args.config,
+            )?;
             if let Some(unit) = report_object(object, &diff_config, Some(&mut existing_functions))?
             {
                 units.push(unit);
@@ -134,7 +146,15 @@ fn generate(args: GenerateArgs) -> Result<()> {
     } else {
         let vec = objects
             .par_iter()
-            .map(|object| report_object(object, &diff_config, None))
+            .map(|(object, unit_idx)| {
+                let diff_config = build_unit_diff_config(
+                    &base_diff_config,
+                    project.options.as_ref(),
+                    project_units.get(*unit_idx).and_then(ProjectObject::options),
+                    &args.config,
+                )?;
+                report_object(object, &diff_config, None)
+            })
             .collect::<Result<Vec<Option<ReportUnit>>>>()?;
         units = vec.into_iter().flatten().collect();
     }
@@ -156,6 +176,24 @@ fn generate(args: GenerateArgs) -> Result<()> {
     Ok(())
 }
 
+fn build_unit_diff_config(
+    base: &diff::DiffObjConfig,
+    project_options: Option<&ProjectOptions>,
+    unit_options: Option<&ProjectOptions>,
+    cli_args: &[String],
+) -> Result<diff::DiffObjConfig> {
+    let mut diff_config = base.clone();
+    if let Some(options) = project_options {
+        apply_project_options(&mut diff_config, options)?;
+    }
+    if let Some(options) = unit_options {
+        apply_project_options(&mut diff_config, options)?;
+    }
+    // CLI args override project and unit options
+    apply_config_args(&mut diff_config, cli_args)?;
+    Ok(diff_config)
+}
+
 fn report_object(
     object: &ObjectConfig,
     diff_config: &diff::DiffObjConfig,

objdiff-cli/src/views/function_diff.rs

@@ -378,7 +378,6 @@ impl UiView for FunctionDiffUi {
                     }
                     // Reload
                     KeyCode::Char('r') => {
-                        result.redraw = true;
                         return EventControlFlow::Reload;
                     }
                     // Scroll right
@@ -400,7 +399,6 @@ impl UiView for FunctionDiffUi {
                                 FunctionRelocDiffs::DataValue => FunctionRelocDiffs::All,
                                 FunctionRelocDiffs::All => FunctionRelocDiffs::None,
                             };
-                        result.redraw = true;
                         return EventControlFlow::Reload;
                     }
                     // Toggle three-way diff

objdiff-core/Cargo.toml

@@ -41,7 +41,8 @@ any-arch = [
     "dep:regex",
     "dep:similar",
     "dep:syn",
-    "dep:encoding_rs"
+    "dep:encoding_rs",
+    "demangler",
 ]
 bindings = [
     "dep:prost",
@@ -88,38 +89,37 @@ std = [
 ]
 mips = [
     "any-arch",
-    "dep:cpp_demangle",
-    "dep:cwdemangle",
     "dep:rabbitizer",
 ]
 ppc = [
     "any-arch",
-    "dep:cwdemangle",
     "dep:cwextab",
     "dep:powerpc",
     "dep:rlwinmdec",
 ]
 x86 = [
     "any-arch",
-    "dep:cpp_demangle",
     "dep:iced-x86",
-    "dep:msvc-demangler",
 ]
 arm = [
     "any-arch",
     "dep:arm-attr",
-    "dep:cpp_demangle",
     "dep:unarm",
 ]
 arm64 = [
     "any-arch",
-    "dep:cpp_demangle",
     "dep:yaxpeax-arch",
     "dep:yaxpeax-arm",
 ]
 superh = [
     "any-arch",
 ]
+demangler = [
+    "dep:cpp_demangle",
+    "dep:cwdemangle",
+    "dep:gnuv2_demangle",
+    "dep:msvc-demangler",
+]
 
 [package.metadata.docs.rs]
 features = ["all"]
@@ -135,10 +135,10 @@ num-traits = { version = "0.2", default-features = false, 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 }
+regex = { version = "1.12", default-features = false, features = [], optional = true }
 serde = { version = "1.0", default-features = false, features = ["derive"], 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 }
+typed-path = { version = "0.12", default-features = false, optional = true }
 
 # config
 globset = { version = "0.4", default-features = false, optional = true }
@@ -150,21 +150,18 @@ gimli = { git = "https://github.com/gimli-rs/gimli", rev = "7335f00e7c39fd501511
 typed-arena = { version = "2.0", default-features = false, optional = true }
 
 # ppc
-cwdemangle = { version = "1.0", 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.4", default-features = false, features = ["all_extensions"], optional = true }
+rabbitizer = { version = "2.0.0-alpha.7", default-features = false, features = ["all_extensions"], optional = true }
 
 # x86
-cpp_demangle = { version = "0.4", default-features = false, features = ["alloc"], optional = true }
 iced-x86 = { version = "1.21", default-features = false, features = ["decoder", "intel", "gas", "masm", "nasm", "exhaustive_enums", "no_std"], optional = true }
-msvc-demangler = { version = "0.11", optional = true }
 
 # arm
-unarm = { version = "1.9", optional = true }
+unarm = { version = "2.1", optional = true }
 arm-attr = { version = "0.2", optional = true }
 
 # arm64
@@ -172,13 +169,19 @@ yaxpeax-arch = { version = "0.3", default-features = false, optional = true }
 yaxpeax-arm = { version = "0.3", default-features = false, optional = true }
 
 # build
-notify = { version = "8.1.0", optional = true }
+notify = { version = "8.2.0", optional = true }
-notify-debouncer-full = { version = "0.5.0", optional = true }
+notify-debouncer-full = { version = "0.6.0", optional = true }
 shell-escape = { version = "0.1", optional = true }
-tempfile = { version = "3.20", optional = true }
+tempfile = { version = "3.23", optional = true }
 time = { version = "0.3", optional = true }
 encoding_rs = { version = "0.8.35", optional = true }
 
+# demangler
+cpp_demangle = { version = "0.5", optional = true, default-features = false, features = ["alloc"] }
+cwdemangle = { version = "1.0", optional = true }
+gnuv2_demangle = { version = "0.4", optional = true }
+msvc-demangler = { version = "0.11", optional = true }
+
 [target.'cfg(windows)'.dependencies]
 winapi = { version = "0.3", optional = true, features = ["winbase"] }
 

objdiff-core/config-schema.json

@@ -25,6 +25,42 @@
         }
       ]
     },
+    {
+      "id": "demangler",
+      "type": "choice",
+      "default": "auto",
+      "name": "Demangler",
+      "description": "Which demangler should be used to demangle each symbol.",
+      "items": [
+        {
+          "value": "auto",
+          "name": "Auto",
+          "description": "Try to automatically guess the mangling format."
+        },
+        {
+          "value": "none",
+          "name": "None",
+          "description": "Disable demangling."
+        },
+        {
+          "value": "codewarrior",
+          "name": "CodeWarrior"
+        },
+        {
+          "value": "itanium",
+          "name": "Itanium"
+        },
+        {
+          "value": "msvc",
+          "name": "MSVC"
+        },
+        {
+          "value": "gnu_legacy",
+          "name": "GNU g++ (Legacy)",
+          "description": "Use the old GNU mangling ABI. Used up to g++ 2.9.x"
+        }
+      ]
+    },
     {
       "id": "analyzeDataFlow",
       "type": "boolean",
@@ -46,6 +82,27 @@
       "name": "Space between args",
       "description": "Adds a space between arguments in the diff output."
     },
+    {
+      "id": "showSymbolSizes",
+      "type": "choice",
+      "default": "off",
+      "name": "Show symbol sizes",
+      "description": "Shows symbol sizes in the symbol view.",
+      "items": [
+        {
+          "value": "off",
+          "name": "Off"
+        },
+        {
+          "value": "hex",
+          "name": "Hex"
+        },
+        {
+          "value": "decimal",
+          "name": "Decimal"
+        }
+      ]
+    },
     {
       "id": "combineDataSections",
       "type": "boolean",
@@ -71,20 +128,43 @@
           "value": "auto",
           "name": "Auto"
         },
+        {
+          "value": "v4",
+          "name": "ARMv4"
+        },
         {
           "value": "v4t",
           "name": "ARMv4T (GBA)"
         },
+        {
+          "value": "v5t",
+          "name": "ARMv5T"
+        },
         {
           "value": "v5te",
           "name": "ARMv5TE (DS)"
         },
+        {
+          "value": "v5tej",
+          "name": "ARMv5TEJ"
+        },
+        {
+          "value": "v6",
+          "name": "ARMv6"
+        },
         {
           "value": "v6k",
           "name": "ARMv6K (3DS)"
         }
       ]
     },
+    {
+      "id": "arm.vfpV2",
+      "type": "boolean",
+      "default": true,
+      "name": "VFPv2 instructions",
+      "description": "Adds floating-point instructions from VFPv2."
+    },
     {
       "id": "arm.unifiedSyntax",
       "type": "boolean",
@@ -158,6 +238,10 @@
           "value": "o32",
           "name": "O32"
         },
+        {
+          "value": "o64",
+          "name": "O64"
+        },
         {
           "value": "n32",
           "name": "N32"
@@ -165,6 +249,14 @@
         {
           "value": "n64",
           "name": "N64"
+        },
+        {
+          "value": "eabi32",
+          "name": "eabi32"
+        },
+        {
+          "value": "eabi64",
+          "name": "eabi64"
         }
       ]
     },
@@ -247,6 +339,8 @@
       "name": "General",
       "properties": [
         "functionRelocDiffs",
+        "demangler",
+        "showSymbolSizes",
         "spaceBetweenArgs",
         "combineDataSections",
         "combineTextSections"
@@ -257,6 +351,7 @@
       "name": "ARM",
       "properties": [
         "arm.archVersion",
+        "arm.vfpV2",
         "arm.unifiedSyntax",
         "arm.avRegisters",
         "arm.r9Usage",
@@ -268,26 +363,17 @@
     {
       "id": "mips",
       "name": "MIPS",
-      "properties": [
+      "properties": ["mips.abi", "mips.instrCategory", "mips.registerPrefix"]
-        "mips.abi",
-        "mips.instrCategory",
-        "mips.registerPrefix"
-      ]
     },
     {
       "id": "ppc",
       "name": "PowerPC",
-      "properties": [
+      "properties": ["ppc.calculatePoolRelocations", "analyzeDataFlow"]
-        "ppc.calculatePoolRelocations",
-        "analyzeDataFlow"
-      ]
     },
     {
       "id": "x86",
       "name": "x86",
-      "properties": [
+      "properties": ["x86.formatter"]
-        "x86.formatter"
-      ]
     }
   ]
 }

objdiff-core/src/arch/arm64.rs

@@ -1,8 +1,4 @@
-use alloc::{
+use alloc::{format, string::ToString, vec::Vec};
-    format,
-    string::{String, ToString},
-    vec::Vec,
-};
 use core::cmp::Ordering;
 
 use anyhow::Result;
@@ -14,7 +10,7 @@ use yaxpeax_arm::armv8::a64::{
 };
 
 use crate::{
-    arch::Arch,
+    arch::{Arch, OPCODE_INVALID},
     diff::{DiffObjConfig, display::InstructionPart},
     obj::{
         InstructionRef, Relocation, RelocationFlags, ResolvedInstructionRef, ResolvedRelocation,
@@ -60,7 +56,7 @@ impl Arch for ArchArm64 {
                         ops.push(InstructionRef {
                             address,
                             size: 4,
-                            opcode: u16::MAX,
+                            opcode: OPCODE_INVALID,
                             branch_dest: None,
                         });
                         continue;
@@ -87,7 +83,7 @@ impl Arch for ArchArm64 {
         let decoder = InstDecoder::default();
         let mut ins = Instruction::default();
         if decoder.decode_into(&mut ins, &mut reader).is_err() {
-            cb(InstructionPart::opcode("<invalid>", u16::MAX))?;
+            cb(InstructionPart::opcode("<invalid>", OPCODE_INVALID))?;
             return Ok(());
         }
 
@@ -108,12 +104,6 @@ impl Arch for ArchArm64 {
         Ok(())
     }
 
-    fn demangle(&self, name: &str) -> Option<String> {
-        cpp_demangle::Symbol::new(name)
-            .ok()
-            .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
-    }
-
     fn reloc_name(&self, flags: RelocationFlags) -> Option<&'static str> {
         match flags {
             RelocationFlags::Elf(r_type) => match r_type {
@@ -2295,7 +2285,7 @@ where Cb: FnMut(InstructionPart<'static>) {
 // Opcode is #[repr(u16)], but the tuple variants negate that, so we have to do this instead.
 const fn opcode_to_u16(opcode: Opcode) -> u16 {
     match opcode {
-        Opcode::Invalid => u16::MAX,
+        Opcode::Invalid => OPCODE_INVALID,
         Opcode::UDF => 0,
         Opcode::MOVN => 1,
         Opcode::MOVK => 2,

objdiff-core/src/arch/mips.rs

@@ -1,6 +1,6 @@
 use alloc::{
     collections::{BTreeMap, BTreeSet},
-    string::{String, ToString},
+    string::ToString,
     vec::Vec,
 };
 
@@ -46,8 +46,11 @@ impl ArchMips {
             object::FileFlags::None => {}
             object::FileFlags::Elf { e_flags, .. } => {
                 abi = match e_flags & EF_MIPS_ABI {
-                    elf::EF_MIPS_ABI_O32 | elf::EF_MIPS_ABI_O64 => Abi::O32,
+                    elf::EF_MIPS_ABI_O32 => Abi::O32,
-                    elf::EF_MIPS_ABI_EABI32 | elf::EF_MIPS_ABI_EABI64 => Abi::N32,
+                    elf::EF_MIPS_ABI_O64 if e_flags & elf::EF_MIPS_ABI2 != 0 => Abi::N64,
+                    elf::EF_MIPS_ABI_O64 => Abi::O64,
+                    elf::EF_MIPS_ABI_EABI32 => Abi::EABI32,
+                    elf::EF_MIPS_ABI_EABI64 => Abi::EABI64,
                     _ => {
                         if e_flags & elf::EF_MIPS_ABI2 != 0 {
                             Abi::N32
@@ -170,8 +173,11 @@ impl ArchMips {
         .with_abi(match diff_config.mips_abi {
             MipsAbi::Auto => self.abi,
             MipsAbi::O32 => Abi::O32,
+            MipsAbi::O64 => Abi::O64,
             MipsAbi::N32 => Abi::N32,
             MipsAbi::N64 => Abi::N64,
+            MipsAbi::Eabi32 => Abi::EABI32,
+            MipsAbi::Eabi64 => Abi::EABI64,
         })
     }
 
@@ -287,6 +293,7 @@ impl Arch for ArchMips {
                         }
                         elf::R_MIPS_PC16 => 0, // PC-relative relocation
                         R_MIPS15_S3 => ((code & 0x001FFFC0) >> 3) as i64,
+                        elf::R_MIPS_GPREL32 => (code as i32 as i64) + self.ri_gp_value as i64,
                         flags => bail!("Unsupported MIPS implicit relocation {flags:?}"),
                     };
                     Ok(Some(RelocationOverride { target: RelocationOverrideTarget::Keep, addend }))
@@ -298,13 +305,6 @@ impl Arch for ArchMips {
         }
     }
 
-    fn demangle(&self, name: &str) -> Option<String> {
-        cpp_demangle::Symbol::new(name)
-            .ok()
-            .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
-            .or_else(|| cwdemangle::demangle(name, &cwdemangle::DemangleOptions::default()))
-    }
-
     fn reloc_name(&self, flags: RelocationFlags) -> Option<&'static str> {
         match flags {
             RelocationFlags::Elf(r_type) => match r_type {
@@ -319,6 +319,7 @@ impl Arch for ArchMips {
                 elf::R_MIPS_GOT16 => Some("R_MIPS_GOT16"),
                 elf::R_MIPS_PC16 => Some("R_MIPS_PC16"),
                 elf::R_MIPS_CALL16 => Some("R_MIPS_CALL16"),
+                elf::R_MIPS_GPREL32 => Some("R_MIPS_GPREL32"),
                 R_MIPS15_S3 => Some("R_MIPS15_S3"),
                 _ => None,
             },
@@ -356,6 +357,7 @@ impl Arch for ArchMips {
         while new_address >= symbol.address + 4
             && let Some(data) = section.data_range(new_address - 4, 4)
             && data == [0u8; 4]
+            && section.relocation_at(next_address - 4, 4).is_none()
         {
             new_address -= 4;
         }

objdiff-core/src/arch/mod.rs

@@ -7,12 +7,10 @@ use alloc::{
 };
 use core::{
     any::Any,
-    ffi::CStr,
     fmt::{self, Debug},
 };
 
 use anyhow::{Result, bail};
-use encoding_rs::SHIFT_JIS;
 use object::Endian as _;
 
 use crate::{
@@ -41,6 +39,19 @@ pub mod superh;
 #[cfg(feature = "x86")]
 pub mod x86;
 
+pub const OPCODE_INVALID: u16 = u16::MAX;
+pub const OPCODE_DATA: u16 = u16::MAX - 1;
+
+const SUPPORTED_ENCODINGS: [(&encoding_rs::Encoding, &str); 7] = [
+    (encoding_rs::UTF_8, "UTF-8"),
+    (encoding_rs::SHIFT_JIS, "Shift JIS"),
+    (encoding_rs::UTF_16BE, "UTF-16BE"),
+    (encoding_rs::UTF_16LE, "UTF-16LE"),
+    (encoding_rs::WINDOWS_1252, "Windows-1252"),
+    (encoding_rs::EUC_JP, "EUC-JP"),
+    (encoding_rs::BIG5, "Big5"),
+];
+
 /// Represents the type of data associated with an instruction
 #[derive(PartialEq)]
 pub enum DataType {
@@ -74,7 +85,7 @@ impl DataType {
         let mut strs = Vec::new();
         for (literal, label_override) in self.display_literals(endian, bytes) {
             let label = label_override.unwrap_or_else(|| self.to_string());
-            strs.push(format!("{label}: {literal}"))
+            strs.push(format!("{label}: {literal:?}"))
         }
         strs
     }
@@ -161,16 +172,18 @@ impl DataType {
                 strs.push((format!("{bytes:#?}"), None));
             }
             DataType::String => {
-                if let Ok(cstr) = CStr::from_bytes_until_nul(bytes) {
-                    strs.push((format!("{cstr:?}"), None));
-                }
                 if let Some(nul_idx) = bytes.iter().position(|&c| c == b'\0') {
-                    let (cow, _, had_errors) = SHIFT_JIS.decode(&bytes[..nul_idx]);
+                    let str_bytes = &bytes[..nul_idx];
-                    if !had_errors {
+                    // Special case to display (ASCII) as the label for ASCII-only strings.
-                        let str = format!("{cow:?}");
+                    let (cow, _, had_errors) = encoding_rs::UTF_8.decode(str_bytes);
-                        // Only add the Shift JIS string if it's different from the ASCII string.
+                    if !had_errors && cow.is_ascii() {
-                        if !strs.iter().any(|x| x.0 == str) {
+                        strs.push((format!("{cow}"), Some("ASCII".into())));
-                            strs.push((str, Some("Shift JIS".into())));
+                    }
+                    for (encoding, encoding_name) in SUPPORTED_ENCODINGS {
+                        let (cow, _, had_errors) = encoding.decode(str_bytes);
+                        // Avoid showing ASCII-only strings more than once if the encoding is ASCII-compatible.
+                        if !had_errors && (!encoding.is_ascii_compatible() || !cow.is_ascii()) {
+                            strs.push((format!("{cow}"), Some(encoding_name.into())));
                         }
                     }
                 }
@@ -368,8 +381,6 @@ pub trait Arch: Any + Debug + Send + Sync {
         Ok(None)
     }
 
-    fn demangle(&self, _name: &str) -> Option<String> { None }
-
     fn reloc_name(&self, _flags: RelocationFlags) -> Option<&'static str> { None }
 
     fn data_reloc_size(&self, flags: RelocationFlags) -> usize;

objdiff-core/src/arch/ppc/mod.rs

@@ -308,18 +308,6 @@ impl Arch for ArchPpc {
         }
     }
 
-    fn demangle(&self, mut name: &str) -> Option<String> {
-        if name.starts_with('?') {
-            msvc_demangler::demangle(name, msvc_demangler::DemangleFlags::llvm()).ok()
-        } else {
-            name = name.trim_start_matches('.');
-            cpp_demangle::Symbol::new(name)
-                .ok()
-                .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
-                .or_else(|| cwdemangle::demangle(name, &cwdemangle::DemangleOptions::default()))
-        }
-    }
-
     fn reloc_name(&self, flags: RelocationFlags) -> Option<&'static str> {
         match flags {
             RelocationFlags::Elf(r_type) => match r_type {
@@ -469,6 +457,7 @@ impl Arch for ArchPpc {
         while next_address >= symbol.address + 4
             && let Some(data) = section.data_range(next_address - 4, 4)
             && data == [0u8; 4]
+            && section.relocation_at(next_address - 4, 4).is_none()
         {
             next_address -= 4;
         }

objdiff-core/src/arch/superh/mod.rs

@@ -1,4 +1,4 @@
-use alloc::{collections::BTreeMap, format, string::String, vec, vec::Vec};
+use alloc::{collections::BTreeMap, format, vec, vec::Vec};
 
 use anyhow::Result;
 use object::elf;
@@ -132,12 +132,6 @@ impl Arch for ArchSuperH {
         Ok(())
     }
 
-    fn demangle(&self, name: &str) -> Option<String> {
-        cpp_demangle::Symbol::new(name)
-            .ok()
-            .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
-    }
-
     fn reloc_name(&self, flags: RelocationFlags) -> Option<&'static str> {
         match flags {
             RelocationFlags::Elf(r_type) => match r_type {

objdiff-core/src/arch/x86.rs

@@ -1,4 +1,4 @@
-use alloc::{boxed::Box, format, string::String, vec::Vec};
+use alloc::{boxed::Box, format, vec::Vec};
 use core::cmp::Ordering;
 
 use anyhow::{Context, Result, anyhow, bail};
@@ -9,7 +9,7 @@ use iced_x86::{
 use object::{Endian as _, Object as _, ObjectSection as _, elf, pe};
 
 use crate::{
-    arch::{Arch, RelocationOverride, RelocationOverrideTarget},
+    arch::{Arch, OPCODE_DATA, RelocationOverride, RelocationOverrideTarget},
     diff::{DiffObjConfig, X86Formatter, display::InstructionPart},
     obj::{InstructionRef, Relocation, RelocationFlags, ResolvedInstructionRef, Section, Symbol},
 };
@@ -89,8 +89,6 @@ impl ArchX86 {
     }
 }
 
-const DATA_OPCODE: u16 = u16::MAX - 1;
-
 impl Arch for ArchX86 {
     fn scan_instructions_internal(
         &self,
@@ -121,7 +119,7 @@ impl Arch for ArchX86 {
                             out.push(InstructionRef {
                                 address,
                                 size: size as u8,
-                                opcode: DATA_OPCODE,
+                                opcode: OPCODE_DATA,
                                 branch_dest: None,
                             });
 
@@ -148,7 +146,7 @@ impl Arch for ArchX86 {
                                     out.push(InstructionRef {
                                         address: indirect_array_address + i as u64,
                                         size: 1,
-                                        opcode: DATA_OPCODE,
+                                        opcode: OPCODE_DATA,
                                         branch_dest: None,
                                     });
                                 }
@@ -187,14 +185,14 @@ impl Arch for ArchX86 {
         diff_config: &DiffObjConfig,
         cb: &mut dyn FnMut(InstructionPart) -> Result<()>,
     ) -> Result<()> {
-        if resolved.ins_ref.opcode == DATA_OPCODE {
+        if resolved.ins_ref.opcode == OPCODE_DATA {
             let (mnemonic, imm) = match resolved.ins_ref.size {
                 1 => (".byte", resolved.code[0] as u64),
                 2 => (".word", self.endianness.read_u16_bytes(resolved.code.try_into()?) as u64),
                 4 => (".dword", self.endianness.read_u32_bytes(resolved.code.try_into()?) as u64),
                 _ => bail!("Unsupported x86 inline data size {}", resolved.ins_ref.size),
             };
-            cb(InstructionPart::opcode(mnemonic, DATA_OPCODE))?;
+            cb(InstructionPart::opcode(mnemonic, OPCODE_DATA))?;
             if resolved.relocation.is_some() {
                 cb(InstructionPart::reloc())?;
             } else {
@@ -302,16 +300,6 @@ impl Arch for ArchX86 {
         Ok(Some(RelocationOverride { target: RelocationOverrideTarget::Keep, addend }))
     }
 
-    fn demangle(&self, name: &str) -> Option<String> {
-        if name.starts_with('?') {
-            msvc_demangler::demangle(name, msvc_demangler::DemangleFlags::llvm()).ok()
-        } else {
-            cpp_demangle::Symbol::new(name)
-                .ok()
-                .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
-        }
-    }
-
     fn reloc_name(&self, flags: RelocationFlags) -> Option<&'static str> {
         match self.arch {
             Architecture::X86 => match flags {
@@ -836,7 +824,7 @@ mod test {
                 ins_ref: InstructionRef {
                     address: 0x1234,
                     size: 1,
-                    opcode: DATA_OPCODE,
+                    opcode: OPCODE_DATA,
                     branch_dest: None,
                 },
                 code: &code,
@@ -850,7 +838,7 @@ mod test {
         )
         .unwrap();
         assert_eq!(parts, &[
-            InstructionPart::opcode(".byte", DATA_OPCODE),
+            InstructionPart::opcode(".byte", OPCODE_DATA),
             InstructionPart::unsigned(0xABu64),
         ]);
     }

objdiff-core/src/config/mod.rs

@@ -1,6 +1,7 @@
 pub mod path;
 
 use alloc::{
+    borrow::Cow,
     collections::BTreeMap,
     string::{String, ToString},
     vec::Vec,
@@ -45,6 +46,8 @@ pub struct ProjectConfig {
     pub units: Option<Vec<ProjectObject>>,
     #[cfg_attr(feature = "serde", serde(skip_serializing_if = "Option::is_none"))]
     pub progress_categories: Option<Vec<ProjectProgressCategory>>,
+    #[cfg_attr(feature = "serde", serde(skip_serializing_if = "Option::is_none"))]
+    pub options: Option<ProjectOptions>,
 }
 
 impl ProjectConfig {
@@ -116,6 +119,8 @@ pub struct ProjectObject {
     pub metadata: Option<ProjectObjectMetadata>,
     #[cfg_attr(feature = "serde", serde(skip_serializing_if = "Option::is_none"))]
     pub symbol_mappings: Option<BTreeMap<String, String>>,
+    #[cfg_attr(feature = "serde", serde(skip_serializing_if = "Option::is_none"))]
+    pub options: Option<ProjectOptions>,
 }
 
 #[derive(Default, Clone)]
@@ -143,6 +148,15 @@ pub struct ProjectProgressCategory {
     pub name: String,
 }
 
+pub type ProjectOptions = BTreeMap<String, ProjectOptionValue>;
+
+#[derive(Clone, Debug)]
+#[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize), serde(untagged))]
+pub enum ProjectOptionValue {
+    Bool(bool),
+    String(String),
+}
+
 impl ProjectObject {
     pub fn name(&self) -> &str {
         if let Some(name) = &self.name {
@@ -179,6 +193,8 @@ impl ProjectObject {
     pub fn auto_generated(&self) -> Option<bool> {
         self.metadata.as_ref().and_then(|m| m.auto_generated)
     }
+
+    pub fn options(&self) -> Option<&ProjectOptions> { self.options.as_ref() }
 }
 
 #[derive(Default, Clone, Eq, PartialEq)]
@@ -310,3 +326,47 @@ pub fn build_globset(vec: &[Glob]) -> Result<GlobSet, globset::Error> {
     }
     builder.build()
 }
+
+#[cfg(feature = "any-arch")]
+pub fn apply_project_options(
+    diff_config: &mut crate::diff::DiffObjConfig,
+    options: &ProjectOptions,
+) -> Result<()> {
+    use core::str::FromStr;
+
+    use crate::diff::{ConfigEnum, ConfigPropertyId, ConfigPropertyKind};
+
+    let mut result = Ok(());
+    for (key, value) in options.iter() {
+        let property_id = ConfigPropertyId::from_str(key)
+            .map_err(|()| anyhow!("Invalid configuration property: {key}"))?;
+        let value = match value {
+            ProjectOptionValue::Bool(value) => Cow::Borrowed(if *value { "true" } else { "false" }),
+            ProjectOptionValue::String(value) => Cow::Borrowed(value.as_str()),
+        };
+        if diff_config.set_property_value_str(property_id, &value).is_err() {
+            if result.is_err() {
+                // Already returning an error, skip further errors
+                continue;
+            }
+            let mut expected = String::new();
+            match property_id.kind() {
+                ConfigPropertyKind::Boolean => expected.push_str("true, false"),
+                ConfigPropertyKind::Choice(variants) => {
+                    for (idx, variant) in variants.iter().enumerate() {
+                        if idx > 0 {
+                            expected.push_str(", ");
+                        }
+                        expected.push_str(variant.value);
+                    }
+                }
+            }
+            result = Err(anyhow!(
+                "Invalid value for {}. Expected one of: {}",
+                property_id.name(),
+                expected
+            ));
+        }
+    }
+    result
+}

objdiff-core/src/diff/code.rs

@@ -14,7 +14,7 @@ use super::{
 };
 use crate::obj::{
     InstructionArg, InstructionArgValue, InstructionRef, Object, ResolvedInstructionRef,
-    ResolvedRelocation, ResolvedSymbol, SymbolFlag, SymbolKind,
+    ResolvedRelocation, ResolvedSymbol, SymbolKind,
 };
 
 pub fn no_diff_code(
@@ -333,11 +333,7 @@ fn reloc_eq(
                     || display_ins_data_literals(left_obj, left_ins)
                         == display_ins_data_literals(right_obj, right_ins))
         }
-        (None, Some(_)) => {
+        (Some(_), None) | (None, Some(_)) | (None, None) => symbol_name_addend_matches,
-            // Match if possibly stripped weak symbol
-            symbol_name_addend_matches && right_reloc.symbol.flags.contains(SymbolFlag::Weak)
-        }
-        (Some(_), None) | (None, None) => symbol_name_addend_matches,
     }
 }