ci: Add id-token write permission

* Add data info for superh

* fmt

* Fetch symbol data dynamically from resolved.section

* Remove unused var

---------

Co-authored-by: Luke Street <luke@street.dev>

.github/workflows/build.yaml

@@ -8,6 +8,10 @@ on:
       - "LICENSE*"
   workflow_dispatch:
 
+permissions:
+  # For npm publish provenance
+  id-token: write
+
 env:
   BUILD_PROFILE: release-lto
   CARGO_INCREMENTAL: 0
@@ -72,7 +76,7 @@ jobs:
     name: Test
     strategy:
       matrix:
-        platform: [ubuntu-latest, windows-latest, macos-latest]
+        platform: [ ubuntu-latest, windows-latest, macos-latest ]
       fail-fast: false
     runs-on: ${{ matrix.platform }}
     steps:
@@ -249,22 +253,30 @@ jobs:
           components: rust-src
       - name: Cache Rust workspace
         uses: Swatinem/rust-cache@v2
+      - name: Setup Node
+        uses: actions/setup-node@v4
+        with:
+          node-version: lts/*
       - name: Install dependencies
-        run: npm -C objdiff-wasm install
+        run: npm -C objdiff-wasm ci
       - name: Build
         run: npm -C objdiff-wasm run build
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: wasm
+          path: objdiff-wasm/dist/
+          if-no-files-found: error
 
-  release:
-    name: Release
+  check-version:
+    name: Check package versions
     if: startsWith(github.ref, 'refs/tags/')
     runs-on: ubuntu-latest
-    needs: [build-cli, build-gui]
-    permissions:
-      contents: write
+    needs: [ build-cli, build-gui, build-wasm ]
     steps:
       - name: Checkout
         uses: actions/checkout@v4
-      - name: Check git tag against Cargo version
+      - name: Check git tag against package versions
         shell: bash
         run: |
           set -eou pipefail
@@ -276,9 +288,24 @@ jobs:
             echo "::error::Git tag doesn't match the Cargo version! ($tag != $version)"
             exit 1
           fi
+          version="v$(jq -r .version objdiff-wasm/package.json)"
+          if [ "$tag" != "$version" ]; then
+            echo "::error::Git tag doesn't match the npm version! ($tag != $version)"
+            exit 1
+          fi
+
+  release-github:
+    name: Release (GitHub)
+    if: startsWith(github.ref, 'refs/tags/')
+    runs-on: ubuntu-latest
+    needs: [ check-version ]
+    permissions:
+      contents: write
+    steps:
       - name: Download artifacts
         uses: actions/download-artifact@v4
         with:
+          pattern: objdiff-*
           path: artifacts
       - name: Rename artifacts
         working-directory: artifacts
@@ -308,3 +335,53 @@ jobs:
           files: out/*
           draft: true
           generate_release_notes: true
+
+  release-cargo:
+    name: Release (Cargo)
+    if: 'false' # TODO re-enable when all dependencies are published
+    runs-on: ubuntu-latest
+    needs: [ check-version ]
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Rust toolchain
+        uses: dtolnay/rust-toolchain@stable
+      - name: Publish
+        env:
+          CARGO_REGISTRY_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
+        run: cargo publish -p objdiff-core
+
+  release-npm:
+    name: Release (npm)
+    if: startsWith(github.ref, 'refs/tags/')
+    runs-on: ubuntu-latest
+    needs: [ check-version ]
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Node
+        uses: actions/setup-node@v4
+        with:
+          node-version: lts/*
+          registry-url: 'https://registry.npmjs.org'
+      - name: Download artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: wasm
+          path: objdiff-wasm/dist
+      - name: Publish
+        working-directory: objdiff-wasm
+        env:
+          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
+        run: |
+          set -euo pipefail
+          version=$(jq -r '.version' package.json)
+          tag="latest"
+          # Check for prerelease by looking for a dash
+          case "$version" in
+            *-*)
+              tag=$(echo "$version" | sed -e 's/^[^-]*-//' -e 's/\..*$//')
+              ;;
+          esac
+          echo "Publishing version $version with tag '$tag'..."
+          npm publish --provenance --access public --tag "$tag"

Cargo.toml

@@ -14,7 +14,7 @@ strip = "debuginfo"
 codegen-units = 1
 
 [workspace.package]
-version = "3.0.0-beta.5"
+version = "3.0.0-beta.6"
 authors = ["Luke Street <luke@street.dev>"]
 edition = "2024"
 license = "MIT OR Apache-2.0"

deny.toml

@@ -240,8 +240,6 @@ allow-git = []
 [sources.allow-org]
 # github.com organizations to allow git sources for
 github = [
-    "CelestialAmber", # cwextab, rlwinmdec
-    "Decompollaborate", # rabbitizer
     "enarx", # flagset
     "encounter",
 ]

objdiff-cli/src/cmd/diff.rs

@@ -1,7 +1,6 @@
 use std::{
     io::stdout,
     mem,
-    str::FromStr,
     sync::{
         Arc,
         atomic::{AtomicBool, Ordering},
@@ -29,10 +28,7 @@ use objdiff_core::{
         ProjectConfig, ProjectObject, ProjectObjectMetadata, build_globset,
         path::{check_path_buf, platform_path, platform_path_serde_option},
     },
-    diff::{
-        self, ConfigEnum, ConfigPropertyId, ConfigPropertyKind, DiffObjConfig, MappingConfig,
-        ObjectDiff,
-    },
+    diff::{self, DiffObjConfig, MappingConfig, ObjectDiff},
     jobs::{
         Job, JobQueue, JobResult,
         objdiff::{ObjDiffConfig, start_build},
@@ -43,6 +39,7 @@ use ratatui::prelude::*;
 use typed_path::{Utf8PlatformPath, Utf8PlatformPathBuf};
 
 use crate::{
+    cmd::apply_config_args,
     util::{
         output::{OutputFormat, write_output},
         term::crossterm_panic_handler,
@@ -183,28 +180,7 @@ pub fn run(args: Args) -> Result<()> {
 
 fn build_config_from_args(args: &Args) -> Result<(DiffObjConfig, MappingConfig)> {
     let mut diff_config = DiffObjConfig::default();
-    for config in &args.config {
-        let (key, value) = config.split_once('=').context("--config expects \"key=value\"")?;
-        let property_id = ConfigPropertyId::from_str(key)
-            .map_err(|()| anyhow!("Invalid configuration property: {}", key))?;
-        diff_config.set_property_value_str(property_id, value).map_err(|()| {
-            let mut options = String::new();
-            match property_id.kind() {
-                ConfigPropertyKind::Boolean => {
-                    options = "true, false".to_string();
-                }
-                ConfigPropertyKind::Choice(variants) => {
-                    for (i, variant) in variants.iter().enumerate() {
-                        if i > 0 {
-                            options.push_str(", ");
-                        }
-                        options.push_str(variant.value);
-                    }
-                }
-            }
-            anyhow!("Invalid value for {}. Expected one of: {}", property_id.name(), options)
-        })?;
-    }
+    apply_config_args(&mut diff_config, &args.config)?;
     let mut mapping_config = MappingConfig {
         mappings: Default::default(),
         selecting_left: args.selecting_left.clone(),

objdiff-cli/src/cmd/mod.rs

@@ -1,2 +1,33 @@
 pub mod diff;
 pub mod report;
+
+use std::str::FromStr;
+
+use anyhow::{Context, Result, anyhow};
+use objdiff_core::diff::{ConfigEnum, ConfigPropertyId, ConfigPropertyKind, DiffObjConfig};
+
+pub fn apply_config_args(diff_config: &mut DiffObjConfig, args: &[String]) -> Result<()> {
+    for config in args {
+        let (key, value) = config.split_once('=').context("--config expects \"key=value\"")?;
+        let property_id = ConfigPropertyId::from_str(key)
+            .map_err(|()| anyhow!("Invalid configuration property: {}", key))?;
+        diff_config.set_property_value_str(property_id, value).map_err(|()| {
+            let mut options = String::new();
+            match property_id.kind() {
+                ConfigPropertyKind::Boolean => {
+                    options = "true, false".to_string();
+                }
+                ConfigPropertyKind::Choice(variants) => {
+                    for (i, variant) in variants.iter().enumerate() {
+                        if i > 0 {
+                            options.push_str(", ");
+                        }
+                        options.push_str(variant.value);
+                    }
+                }
+            }
+            anyhow!("Invalid value for {}. Expected one of: {}", property_id.name(), options)
+        })?;
+    }
+    Ok(())
+}

objdiff-cli/src/cmd/report.rs

@@ -17,7 +17,7 @@ use tracing::{info, warn};
 use typed_path::{Utf8PlatformPath, Utf8PlatformPathBuf};
 
 use crate::{
-    cmd::diff::ObjectConfig,
+    cmd::{apply_config_args, diff::ObjectConfig},
     util::output::{OutputFormat, write_output},
 };
 
@@ -52,6 +52,9 @@ pub struct GenerateArgs {
     #[argp(option, short = 'f')]
     /// Output format (json, json-pretty, proto) (default: json)
     format: Option<String>,
+    #[argp(option, short = 'c')]
+    /// Configuration property (key=value)
+    config: Vec<String>,
 }
 
 #[derive(FromArgs, PartialEq, Debug)]
@@ -80,6 +83,12 @@ pub fn run(args: Args) -> Result<()> {
 }
 
 fn generate(args: GenerateArgs) -> Result<()> {
+    let mut diff_config = diff::DiffObjConfig {
+        function_reloc_diffs: diff::FunctionRelocDiffs::None,
+        ..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("."));
     info!("Loading project {}", project_dir);
@@ -114,14 +123,15 @@ fn generate(args: GenerateArgs) -> Result<()> {
     if args.deduplicate {
         // If deduplicating, we need to run single-threaded
         for object in &objects {
-            if let Some(unit) = report_object(object, Some(&mut existing_functions))? {
+            if let Some(unit) = report_object(object, &diff_config, Some(&mut existing_functions))?
+            {
                 units.push(unit);
             }
         }
     } else {
         let vec = objects
             .par_iter()
-            .map(|object| report_object(object, None))
+            .map(|object| report_object(object, &diff_config, None))
             .collect::<Result<Vec<Option<ReportUnit>>>>()?;
         units = vec.into_iter().flatten().collect();
     }
@@ -145,6 +155,7 @@ fn generate(args: GenerateArgs) -> Result<()> {
 
 fn report_object(
     object: &ObjectConfig,
+    diff_config: &diff::DiffObjConfig,
     mut existing_functions: Option<&mut HashSet<String>>,
 ) -> Result<Option<ReportUnit>> {
     match (&object.target_path, &object.base_path) {
@@ -158,16 +169,12 @@ fn report_object(
         }
         _ => {}
     }
-    let diff_config = diff::DiffObjConfig {
-        function_reloc_diffs: diff::FunctionRelocDiffs::None,
-        ..Default::default()
-    };
     let mapping_config = diff::MappingConfig::default();
     let target = object
         .target_path
         .as_ref()
         .map(|p| {
-            obj::read::read(p.as_ref(), &diff_config)
+            obj::read::read(p.as_ref(), diff_config)
                 .with_context(|| format!("Failed to open {}", p))
         })
         .transpose()?;
@@ -175,12 +182,12 @@ fn report_object(
         .base_path
         .as_ref()
         .map(|p| {
-            obj::read::read(p.as_ref(), &diff_config)
+            obj::read::read(p.as_ref(), diff_config)
                 .with_context(|| format!("Failed to open {}", p))
         })
         .transpose()?;
     let result =
-        diff::diff_objs(target.as_ref(), base.as_ref(), None, &diff_config, &mapping_config)?;
+        diff::diff_objs(target.as_ref(), base.as_ref(), None, diff_config, &mapping_config)?;
 
     let metadata = ReportUnitMetadata {
         complete: object.metadata.complete,

objdiff-core/Cargo.toml

@@ -27,6 +27,7 @@ all = [
     "mips",
     "ppc",
     "x86",
+    "superh"
 ]
 # Implicit, used to check if any arch is enabled
 any-arch = [
@@ -75,6 +76,7 @@ std = [
     "object/std",
     "prost?/std",
     "serde?/std",
+    "similar?/std",
     "typed-path?/std",
     "dep:filetime",
     "dep:memmap2",
@@ -110,6 +112,9 @@ arm64 = [
     "dep:yaxpeax-arch",
     "dep:yaxpeax-arm",
 ]
+superh = [
+    "any-arch",
+]
 
 [package.metadata.docs.rs]
 features = ["all"]
@@ -122,12 +127,12 @@ 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 = { version = "0.36", default-features = false, features = ["read_core", "elf", "pe"] }
+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 }
 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, optional = true, git = "https://github.com/encounter/similar.git", branch = "no_std" }
+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 }
 
 # config
@@ -140,12 +145,12 @@ gimli = { version = "0.31", default-features = false, features = ["read"], optio
 
 # ppc
 cwdemangle = { version = "1.0", optional = true }
-cwextab = { version = "1.0", optional = true, git = "https://github.com/CelestialAmber/cwextab.git" }
+cwextab = { version = "1.0", optional = true }
 ppc750cl = { version = "0.3", optional = true }
-rlwinmdec = { version = "1.1", optional = true, git = "https://github.com/CelestialAmber/rlwinmdec.git" }
+rlwinmdec = { version = "1.1", optional = true }
 
 # mips
-rabbitizer = { git = "https://github.com/Decompollaborate/rabbitizer.git", branch = "🦀", default-features = false, features = ["all_extensions"], optional = true }
+rabbitizer = { version = "2.0.0-alpha.1", default-features = false, features = ["all_extensions"], optional = true }
 
 # x86
 cpp_demangle = { version = "0.4", default-features = false, features = ["alloc"], optional = true }
@@ -164,8 +169,9 @@ yaxpeax-arm = { version = "0.3", default-features = false, optional = true }
 notify = { version = "8.0.0", optional = true }
 notify-debouncer-full = { version = "0.5.0", optional = true }
 shell-escape = { version = "0.1", optional = true }
-tempfile = { version = "3.17", optional = true }
+tempfile = { version = "3.19", optional = true }
 time = { version = "0.3", optional = true }
+encoding_rs = "0.8.35"
 
 [target.'cfg(windows)'.dependencies]
 winapi = { version = "0.3", optional = true }

objdiff-core/src/arch/arm.rs

@@ -15,7 +15,7 @@ use crate::{
     diff::{ArmArchVersion, ArmR9Usage, DiffObjConfig, display::InstructionPart},
     obj::{
         InstructionRef, Relocation, RelocationFlags, ResolvedInstructionRef, ResolvedRelocation,
-        ScannedInstruction, SymbolFlag, SymbolFlagSet, SymbolKind,
+        ScannedInstruction, Section, SectionKind, Symbol, SymbolFlag, SymbolFlagSet, SymbolKind,
     },
 };
 
@@ -32,7 +32,8 @@ impl ArchArm {
         let endianness = file.endianness();
         match file {
             object::File::Elf32(_) => {
-                let disasm_modes = Self::elf_get_mapping_symbols(file);
+                // The disasm_modes mapping is populated later in the post_init step so that we have access to merged sections.
+                let disasm_modes = BTreeMap::new();
                 let detected_version = Self::elf_detect_arm_version(file)?;
                 Ok(Self { disasm_modes, detected_version, endianness })
             }
@@ -73,18 +74,22 @@ impl ArchArm {
         Ok(None)
     }
 
-    fn elf_get_mapping_symbols(file: &object::File) -> BTreeMap<usize, Vec<DisasmMode>> {
-        file.sections()
-            .filter(|s| s.kind() == object::SectionKind::Text)
-            .map(|s| {
-                let index = s.index();
-                let mut mapping_symbols: Vec<_> = file
-                    .symbols()
-                    .filter(|s| s.section_index().map(|i| i == index).unwrap_or(false))
-                    .filter_map(|s| DisasmMode::from_symbol(&s))
+    fn get_mapping_symbols(
+        sections: &[Section],
+        symbols: &[Symbol],
+    ) -> BTreeMap<usize, Vec<DisasmMode>> {
+        sections
+            .iter()
+            .enumerate()
+            .filter(|(_, section)| section.kind == SectionKind::Code)
+            .map(|(index, _)| {
+                let mut mapping_symbols: Vec<_> = symbols
+                    .iter()
+                    .filter(|s| s.section.map(|i| i == index).unwrap_or(false))
+                    .filter_map(DisasmMode::from_symbol)
                     .collect();
                 mapping_symbols.sort_unstable_by_key(|x| x.address);
-                (s.index().0 - 1, mapping_symbols)
+                (index, mapping_symbols)
             })
             .collect()
     }
@@ -178,6 +183,10 @@ impl ArchArm {
 }
 
 impl Arch for ArchArm {
+    fn post_init(&mut self, sections: &[Section], symbols: &[Symbol]) {
+        self.disasm_modes = Self::get_mapping_symbols(sections, symbols);
+    }
+
     fn scan_instructions(
         &self,
         address: u64,
@@ -338,6 +347,7 @@ impl Arch for ArchArm {
             cb(InstructionPart::reloc())?;
         } else {
             push_args(
+                ins,
                 &parsed_ins,
                 resolved.relocation,
                 resolved.ins_ref.address as u32,
@@ -441,7 +451,7 @@ impl Arch for ArchArm {
 
     fn extra_symbol_flags(&self, symbol: &object::Symbol) -> SymbolFlagSet {
         let mut flags = SymbolFlagSet::default();
-        if DisasmMode::from_symbol(symbol).is_some() {
+        if DisasmMode::from_object_symbol(symbol).is_some() {
             flags |= SymbolFlag::Hidden;
         }
         flags
@@ -455,15 +465,21 @@ struct DisasmMode {
 }
 
 impl DisasmMode {
-    fn from_symbol<'a>(sym: &object::Symbol<'a, '_, &'a [u8]>) -> Option<Self> {
+    fn from_object_symbol<'a>(sym: &object::Symbol<'a, '_, &'a [u8]>) -> Option<Self> {
         sym.name()
             .ok()
             .and_then(unarm::ParseMode::from_mapping_symbol)
             .map(|mapping| DisasmMode { address: sym.address() as u32, mapping })
     }
+
+    fn from_symbol(sym: &Symbol) -> Option<Self> {
+        unarm::ParseMode::from_mapping_symbol(&sym.name)
+            .map(|mapping| DisasmMode { address: sym.address as u32, mapping })
+    }
 }
 
 fn push_args(
+    ins: unarm::Ins,
     parsed_ins: &unarm::ParsedIns,
     relocation: Option<ResolvedRelocation>,
     cur_addr: u32,
@@ -609,6 +625,14 @@ fn push_args(
             arg_cb(InstructionPart::opaque("!"))?;
         }
     }
+
+    let branch_dest = get_pc_relative_load_address(ins, cur_addr);
+    if let Some(branch_dest) = branch_dest {
+        arg_cb(InstructionPart::basic(" (->"))?;
+        arg_cb(InstructionPart::branch_dest(branch_dest))?;
+        arg_cb(InstructionPart::basic(")"))?;
+    }
+
     Ok(())
 }
 
@@ -636,3 +660,21 @@ fn find_reloc_arg(
         None
     }
 }
+
+fn get_pc_relative_load_address(ins: unarm::Ins, address: u32) -> Option<u32> {
+    match ins {
+        unarm::Ins::Arm(ins)
+            if ins.op == arm::Opcode::Ldr
+                && ins.modifier_addr_ldr_str() == arm::AddrLdrStr::Imm
+                && ins.field_rn_deref().reg == args::Register::Pc =>
+        {
+            let offset = ins.field_offset_12().value;
+            Some(address.wrapping_add_signed(offset + 8))
+        }
+        unarm::Ins::Thumb(ins) if ins.op == thumb::Opcode::LdrPc => {
+            let offset = ins.field_rel_immed_8().value;
+            Some((address & !3).wrapping_add_signed(offset + 4))
+        }
+        _ => None,
+    }
+}

objdiff-core/src/arch/mod.rs

@@ -2,6 +2,7 @@ use alloc::{borrow::Cow, boxed::Box, format, string::String, vec::Vec};
 use core::{ffi::CStr, fmt, fmt::Debug};
 
 use anyhow::{Result, bail};
+use encoding_rs::SHIFT_JIS;
 use object::Endian as _;
 
 use crate::{
@@ -11,7 +12,7 @@ use crate::{
     },
     obj::{
         InstructionArg, Object, ParsedInstruction, Relocation, RelocationFlags,
-        ResolvedInstructionRef, ScannedInstruction, Symbol, SymbolFlagSet, SymbolKind,
+        ResolvedInstructionRef, ScannedInstruction, Section, Symbol, SymbolFlagSet, SymbolKind,
     },
     util::ReallySigned,
 };
@@ -24,6 +25,8 @@ pub mod arm64;
 pub mod mips;
 #[cfg(feature = "ppc")]
 pub mod ppc;
+#[cfg(feature = "superh")]
+pub mod superh;
 #[cfg(feature = "x86")]
 pub mod x86;
 
@@ -57,13 +60,18 @@ impl fmt::Display for DataType {
 impl DataType {
     pub fn display_labels(&self, endian: object::Endianness, bytes: &[u8]) -> Vec<String> {
         let mut strs = Vec::new();
-        for literal in self.display_literals(endian, bytes) {
-            strs.push(format!("{}: {}", self, literal))
+        for (literal, label_override) in self.display_literals(endian, bytes) {
+            let label = label_override.unwrap_or_else(|| format!("{}", self));
+            strs.push(format!("{}: {}", label, literal))
         }
         strs
     }
 
-    pub fn display_literals(&self, endian: object::Endianness, bytes: &[u8]) -> Vec<String> {
+    pub fn display_literals(
+        &self,
+        endian: object::Endianness,
+        bytes: &[u8],
+    ) -> Vec<(String, Option<String>)> {
         let mut strs = Vec::new();
         if self.required_len().is_some_and(|l| bytes.len() < l) {
             log::warn!(
@@ -87,56 +95,72 @@ impl DataType {
         match self {
             DataType::Int8 => {
                 let i = i8::from_ne_bytes(bytes.try_into().unwrap());
-                strs.push(format!("{:#x}", i));
+                strs.push((format!("{:#x}", i), None));
 
                 if i < 0 {
-                    strs.push(format!("{:#x}", ReallySigned(i)));
+                    strs.push((format!("{:#x}", ReallySigned(i)), None));
                 }
             }
             DataType::Int16 => {
                 let i = endian.read_i16_bytes(bytes.try_into().unwrap());
-                strs.push(format!("{:#x}", i));
+                strs.push((format!("{:#x}", i), None));
 
                 if i < 0 {
-                    strs.push(format!("{:#x}", ReallySigned(i)));
+                    strs.push((format!("{:#x}", ReallySigned(i)), None));
                 }
             }
             DataType::Int32 => {
                 let i = endian.read_i32_bytes(bytes.try_into().unwrap());
-                strs.push(format!("{:#x}", i));
+                strs.push((format!("{:#x}", i), None));
 
                 if i < 0 {
-                    strs.push(format!("{:#x}", ReallySigned(i)));
+                    strs.push((format!("{:#x}", ReallySigned(i)), None));
                 }
             }
             DataType::Int64 => {
                 let i = endian.read_i64_bytes(bytes.try_into().unwrap());
-                strs.push(format!("{:#x}", i));
+                strs.push((format!("{:#x}", i), None));
 
                 if i < 0 {
-                    strs.push(format!("{:#x}", ReallySigned(i)));
+                    strs.push((format!("{:#x}", ReallySigned(i)), None));
                 }
             }
             DataType::Float => {
                 let bytes: [u8; 4] = bytes.try_into().unwrap();
-                strs.push(format!("{:?}f", match endian {
-                    object::Endianness::Little => f32::from_le_bytes(bytes),
-                    object::Endianness::Big => f32::from_be_bytes(bytes),
-                }));
+                strs.push((
+                    format!("{:?}f", match endian {
+                        object::Endianness::Little => f32::from_le_bytes(bytes),
+                        object::Endianness::Big => f32::from_be_bytes(bytes),
+                    }),
+                    None,
+                ));
             }
             DataType::Double => {
                 let bytes: [u8; 8] = bytes.try_into().unwrap();
-                strs.push(format!("{:?}", match endian {
-                    object::Endianness::Little => f64::from_le_bytes(bytes),
-                    object::Endianness::Big => f64::from_be_bytes(bytes),
-                }));
+                strs.push((
+                    format!("{:?}", match endian {
+                        object::Endianness::Little => f64::from_le_bytes(bytes),
+                        object::Endianness::Big => f64::from_be_bytes(bytes),
+                    }),
+                    None,
+                ));
             }
             DataType::Bytes => {
-                strs.push(format!("{:#?}", bytes));
+                strs.push((format!("{:#?}", bytes), None));
             }
             DataType::String => {
                 if let Ok(cstr) = CStr::from_bytes_until_nul(bytes) {
-                    strs.push(format!("{:?}", cstr));
+                    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]);
+                    if !had_errors {
+                        let str = format!("{:?}", cow);
+                        // Only add the Shift JIS string if it's different from the ASCII string.
+                        if !strs.iter().any(|x| x.0 == str) {
+                            strs.push((str, Some("Shift JIS".into())));
+                        }
+                    }
                 }
             }
         }
@@ -159,6 +183,9 @@ impl DataType {
 }
 
 pub trait Arch: Send + Sync + Debug {
+    // Finishes arch-specific initialization that must be done after sections have been combined.
+    fn post_init(&mut self, _sections: &[Section], _symbols: &[Symbol]) {}
+
     /// Generate a list of instructions references (offset, size, opcode) from the given code.
     ///
     /// The opcode IDs are used to generate the initial diff. Implementations should do as little
@@ -247,7 +274,13 @@ pub trait Arch: Send + Sync + Debug {
         SymbolFlagSet::default()
     }
 
-    fn guess_data_type(&self, _resolved: ResolvedInstructionRef) -> Option<DataType> { None }
+    fn guess_data_type(
+        &self,
+        _resolved: ResolvedInstructionRef,
+        _bytes: &[u8],
+    ) -> Option<DataType> {
+        None
+    }
 
     fn symbol_hover(&self, _obj: &Object, _symbol_index: usize) -> Vec<HoverItem> { Vec::new() }
 
@@ -285,6 +318,8 @@ pub fn new_arch(object: &object::File) -> Result<Box<dyn Arch>> {
         object::Architecture::Arm => Box::new(arm::ArchArm::new(object)?),
         #[cfg(feature = "arm64")]
         object::Architecture::Aarch64 => Box::new(arm64::ArchArm64::new(object)?),
+        #[cfg(feature = "superh")]
+        object::Architecture::SuperH => Box::new(superh::ArchSuperH::new(object)?),
         arch => bail!("Unsupported architecture: {arch:?}"),
     })
 }

objdiff-core/src/arch/ppc.rs

@@ -221,12 +221,21 @@ impl Arch for ArchPpc {
         }
     }
 
-    fn guess_data_type(&self, resolved: ResolvedInstructionRef) -> Option<DataType> {
+    fn guess_data_type(&self, resolved: ResolvedInstructionRef, bytes: &[u8]) -> Option<DataType> {
         if resolved.relocation.is_some_and(|r| r.symbol.name.starts_with("@stringBase")) {
+            // Pooled string.
             return Some(DataType::String);
         }
         let opcode = ppc750cl::Opcode::from(resolved.ins_ref.opcode as u8);
-        guess_data_type_from_load_store_inst_op(opcode)
+        if let Some(ty) = guess_data_type_from_load_store_inst_op(opcode) {
+            // Numeric type.
+            return Some(ty);
+        }
+        if bytes.len() >= 2 && bytes.iter().position(|&c| c == b'\0') == Some(bytes.len() - 1) {
+            // It may be an unpooled string if the symbol contains exactly one null byte at the end of the symbol.
+            return Some(DataType::String);
+        }
+        None
     }
 
     fn symbol_hover(&self, _obj: &Object, symbol_index: usize) -> Vec<HoverItem> {
@@ -449,7 +458,7 @@ fn decode_exception_info(
                 log::warn!(
                     "Exception table decoding failed for function {}, reason: {}",
                     extab_func_name,
-                    e.to_string()
+                    e
                 );
                 return Ok(None);
             }

objdiff-core/src/diff/code.rs

@@ -330,6 +330,7 @@ fn reloc_eq(
                     || address_eq(left_reloc, right_reloc))
                 && (diff_config.function_reloc_diffs == FunctionRelocDiffs::NameAddress
                     || left_reloc.symbol.kind != SymbolKind::Object
+                    || right_reloc.symbol.size == 0 // Likely a pool symbol like ...data, don't treat this as a diff
                     || display_ins_data_literals(left_obj, left_ins)
                         == display_ins_data_literals(right_obj, right_ins))
         }

objdiff-core/src/diff/display.rs

@@ -444,8 +444,8 @@ pub fn relocation_context(
         let literals = display_ins_data_literals(obj, ins);
         if !literals.is_empty() {
             out.push(ContextItem::Separator);
-            for literal in literals {
-                out.push(ContextItem::Copy { value: literal, label: None });
+            for (literal, label_override) in literals {
+                out.push(ContextItem::Copy { value: literal, label: label_override });
             }
         }
     }
@@ -564,13 +564,14 @@ pub fn instruction_hover(
     if let Some(reloc) = resolved.relocation {
         out.push(HoverItem::Separator);
         out.append(&mut relocation_hover(obj, reloc, None));
-        if let Some(ty) = obj.arch.guess_data_type(resolved) {
+        let bytes = obj.symbol_data(reloc.relocation.target_symbol).unwrap_or(&[]);
+        if let Some(ty) = obj.arch.guess_data_type(resolved, bytes) {
             let literals = display_ins_data_literals(obj, resolved);
             if !literals.is_empty() {
                 out.push(HoverItem::Separator);
-                for literal in literals {
+                for (literal, label_override) in literals {
                     out.push(HoverItem::Text {
-                        label: format!("{}", ty),
+                        label: label_override.unwrap_or_else(|| format!("{}", ty)),
                         value: literal,
                         color: HoverItemColor::Normal,
                     });
@@ -758,12 +759,15 @@ pub fn display_ins_data_labels(obj: &Object, resolved: ResolvedInstructionRef) -
     };
     let bytes = &data[reloc.relocation.addend as usize..];
     obj.arch
-        .guess_data_type(resolved)
+        .guess_data_type(resolved, bytes)
         .map(|ty| ty.display_labels(obj.endianness, bytes))
         .unwrap_or_default()
 }
 
-pub fn display_ins_data_literals(obj: &Object, resolved: ResolvedInstructionRef) -> Vec<String> {
+pub fn display_ins_data_literals(
+    obj: &Object,
+    resolved: ResolvedInstructionRef,
+) -> Vec<(String, Option<String>)> {
     let Some(reloc) = resolved.relocation else {
         return Vec::new();
     };
@@ -775,7 +779,7 @@ pub fn display_ins_data_literals(obj: &Object, resolved: ResolvedInstructionRef)
     };
     let bytes = &data[reloc.relocation.addend as usize..];
     obj.arch
-        .guess_data_type(resolved)
+        .guess_data_type(resolved, bytes)
         .map(|ty| ty.display_literals(obj.endianness, bytes))
         .unwrap_or_default()
 }

objdiff-core/src/obj/mod.rs

@@ -9,7 +9,10 @@ use alloc::{
     vec,
     vec::Vec,
 };
-use core::{fmt, num::NonZeroU32};
+use core::{
+    fmt,
+    num::{NonZeroU32, NonZeroU64},
+};
 
 use flagset::{FlagSet, flags};
 
@@ -70,6 +73,7 @@ pub struct Section {
     pub kind: SectionKind,
     pub data: SectionData,
     pub flags: SectionFlagSet,
+    pub align: Option<NonZeroU64>,
     pub relocations: Vec<Relocation>,
     /// Line number info (.line or .debug_line section)
     pub line_info: BTreeMap<u64, u32>,
@@ -105,6 +109,17 @@ impl Section {
         }
     }
 
+    pub fn symbol_data(&self, symbol: &Symbol) -> Option<&[u8]> {
+        let offset = symbol.address.checked_sub(self.address)?;
+        self.data.get(offset as usize..offset as usize + symbol.size as usize)
+    }
+
+    // The alignment to use when "Combine data/text sections" is enabled.
+    pub fn combined_alignment(&self) -> u64 {
+        const MIN_ALIGNMENT: u64 = 4;
+        self.align.map(|align| align.get().max(MIN_ALIGNMENT)).unwrap_or(MIN_ALIGNMENT)
+    }
+
     pub fn relocation_at<'obj>(
         &'obj self,
         obj: &'obj Object,
@@ -363,6 +378,7 @@ static DUMMY_SECTION: Section = Section {
     kind: SectionKind::Unknown,
     data: SectionData(Vec::new()),
     flags: SectionFlagSet::empty(),
+    align: None,
     relocations: Vec::new(),
     line_info: BTreeMap::new(),
     virtual_address: None,

objdiff-core/src/obj/read.rs

@@ -4,7 +4,7 @@ use alloc::{
     string::{String, ToString},
     vec::Vec,
 };
-use core::cmp::Ordering;
+use core::{cmp::Ordering, num::NonZeroU64};
 
 use anyhow::{Context, Result, anyhow, bail, ensure};
 use object::{Object as _, ObjectSection as _, ObjectSymbol as _};
@@ -17,7 +17,7 @@ use crate::{
         Symbol, SymbolFlag, SymbolKind,
         split_meta::{SPLITMETA_SECTION, SplitMeta},
     },
-    util::{read_u16, read_u32},
+    util::{align_data_slice_to, align_u64_to, read_u16, read_u32},
 };
 
 fn map_section_kind(section: &object::Section) -> SectionKind {
@@ -257,6 +257,7 @@ fn map_sections(
             kind,
             data: SectionData(data),
             flags: Default::default(),
+            align: NonZeroU64::new(section.align()),
             relocations: Default::default(),
             virtual_address,
             line_info: Default::default(),
@@ -739,7 +740,10 @@ fn do_combine_sections(
         }
         offsets.push(current_offset);
         current_offset += section.size;
+        let align = section.combined_alignment();
+        current_offset = align_u64_to(current_offset, align);
         data_size += section.data.len();
+        data_size = align_u64_to(data_size as u64, align) as usize;
         num_relocations += section.relocations.len();
     }
     if data_size > 0 {
@@ -754,6 +758,7 @@ fn do_combine_sections(
         let section = &mut sections[i];
         section.size = 0;
         data.append(&mut section.data.0);
+        align_data_slice_to(&mut data, section.combined_alignment());
         section.relocations.iter_mut().for_each(|r| r.address += offset);
         relocations.append(&mut section.relocations);
         line_info.append(&mut section.line_info.iter().map(|(&a, &l)| (a + offset, l)).collect());
@@ -843,7 +848,7 @@ pub fn read(obj_path: &std::path::Path, config: &DiffObjConfig) -> Result<Object
 
 pub fn parse(data: &[u8], config: &DiffObjConfig) -> Result<Object> {
     let obj_file = object::File::parse(data)?;
-    let arch = new_arch(&obj_file)?;
+    let mut arch = new_arch(&obj_file)?;
     let split_meta = parse_split_meta(&obj_file)?;
     let (mut sections, section_indices) =
         map_sections(arch.as_ref(), &obj_file, split_meta.as_ref())?;
@@ -857,6 +862,7 @@ pub fn parse(data: &[u8], config: &DiffObjConfig) -> Result<Object> {
     if config.combine_data_sections || config.combine_text_sections {
         combine_sections(&mut sections, &mut symbols, config)?;
     }
+    arch.post_init(&sections, &symbols);
     Ok(Object {
         arch,
         endianness: obj_file.endianness(),

objdiff-core/src/obj/snapshots/objdiff_core__obj__read__test__combine_sections.snap

@@ -14,6 +14,7 @@ expression: "(sections, symbols)"
                 8,
             ),
             flags: FlagSet(),
+            align: None,
             relocations: [
                 Relocation {
                     flags: Elf(
@@ -53,6 +54,7 @@ expression: "(sections, symbols)"
                 12,
             ),
             flags: FlagSet(Combined),
+            align: None,
             relocations: [
                 Relocation {
                     flags: Elf(
@@ -87,6 +89,7 @@ expression: "(sections, symbols)"
                 0,
             ),
             flags: FlagSet(Hidden),
+            align: None,
             relocations: [],
             line_info: {},
             virtual_address: None,
@@ -101,6 +104,7 @@ expression: "(sections, symbols)"
                 0,
             ),
             flags: FlagSet(Hidden),
+            align: None,
             relocations: [],
             line_info: {},
             virtual_address: None,

objdiff-core/src/obj/split_meta.rs

@@ -3,6 +3,10 @@ use alloc::{string::String, vec, vec::Vec};
 use anyhow::{Result, anyhow};
 use object::{Endian, ObjectSection, elf::SHT_NOTE};
 
+#[cfg(feature = "std")]
+use crate::util::align_data_to_4;
+use crate::util::align_size_to_4;
+
 pub const SPLITMETA_SECTION: &str = ".note.split";
 pub const SHT_SPLITMETA: u32 = SHT_NOTE;
 pub const ELF_NOTE_SPLIT: &[u8] = b"Split";
@@ -190,17 +194,6 @@ where E: Endian
     }
 }
 
-fn align_size_to_4(size: usize) -> usize { (size + 3) & !3 }
-
-#[cfg(feature = "std")]
-fn align_data_to_4<W: std::io::Write + ?Sized>(writer: &mut W, len: usize) -> std::io::Result<()> {
-    const ALIGN_BYTES: &[u8] = &[0; 4];
-    if len % 4 != 0 {
-        writer.write_all(&ALIGN_BYTES[..4 - len % 4])?;
-    }
-    Ok(())
-}
-
 // ELF note format:
 // Name Size | 4 bytes (integer)
 // Desc Size | 4 bytes (integer)

objdiff-core/src/util.rs

@@ -1,4 +1,4 @@
-use alloc::format;
+use alloc::{format, vec::Vec};
 use core::fmt;
 
 use anyhow::{Result, ensure};
@@ -39,3 +39,24 @@ pub fn read_u16(obj_file: &object::File, reader: &mut &[u8]) -> Result<u16> {
     *reader = &reader[2..];
     Ok(obj_file.endianness().read_u16(value))
 }
+
+pub fn align_size_to_4(size: usize) -> usize { (size + 3) & !3 }
+
+#[cfg(feature = "std")]
+pub fn align_data_to_4<W: std::io::Write + ?Sized>(
+    writer: &mut W,
+    len: usize,
+) -> std::io::Result<()> {
+    const ALIGN_BYTES: &[u8] = &[0; 4];
+    if len % 4 != 0 {
+        writer.write_all(&ALIGN_BYTES[..4 - len % 4])?;
+    }
+    Ok(())
+}
+
+pub fn align_u64_to(len: u64, align: u64) -> u64 { len + ((align - (len % align)) % align) }
+
+pub fn align_data_slice_to(data: &mut Vec<u8>, align: u64) {
+    const ALIGN_BYTE: u8 = 0;
+    data.resize(align_u64_to(data.len() as u64, align) as usize, ALIGN_BYTE);
+}