vendor `imara-diff` by Byron · Pull Request #2506 · GitoxideLabs/gitoxide

Sebastian Thiel (Byron) · 2026-04-07T04:26:42Z

Take imara-diff in and unify both versions of it, one that has sliders, the other one that doesn't.
See what LLMs can do as well if they see both versions at the same time easily.

Tasks

first vendor with separate crates for 0.1 and 0.2, used by all code and it compiles
refackiew fix-ci
cleanup (size adjustment, file removal)
enable fuzz-test to show that timeout is fixed.

Out of Scope

port gix-diff and its unified diff writer over to slider-aware imara-diff v0.2
let line-stats still use v0.1 and make 0.2 the default
merge everything into one crate, reuse shared types, etc.

github-advanced-security

zizmor found more than 20 potential problems in the proposed changes. Check the Files changed tab for more details.

This commit contains the intial code as documented, but also makes all adjustments needed to make it all pass CI. This essentially means that all backports have to be done by hand, but I am sure this can be automated. - Add imara-diff@d2930d174bd4469f4932b4658f4fa505e8e0b655 (0.1) for vendoring This is the baseline version, with a DoS issue fixed due to out-of-bound computation. Drop large JSON performance-test fixtures and the perf-only test that used them. - Add imara-diff@32d1e45d3df061e6ccba6db7fdce92db29e345d8 (0.2) for vendoring It contains a DoS fix for a runaway computation. Drop large JSON performance-test fixtures and the perf-only test that used them. - Make vendored, fixed imara-diff crates part of the workspace - remove big test-files to reduce noise - fix CI

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: fd49295c5e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "Codex (@codex) address that feedback".

gix-diff/Cargo.toml

justfile

The problem was detected by Google OSS-Fuzz, and occurs because of window-size problem in imara-diff's preprocessing. With `gix-imara-diff` the problem is fixed as it's now vendored.

Copilot

Pull request overview

This PR vendors imara-diff into the workspace by adding two internal crates (gix-imara-diff-01 for v0.1.8 and gix-imara-diff for v0.2.0) and rewires existing crates to depend on the vendored versions to avoid pulling multiple upstream variants.

Changes:

Add new workspace crates gix-imara-diff-01 (v0.1.8) and gix-imara-diff (v0.2.0), including tests and fuzzing harnesses.
Update gix-diff, gix-merge, and fuzz crates to use the vendored crates via package = ... + path = ....
Adjust justfile checks to validate dependency absence/presence using the new crate names.

Reviewed changes

Copilot reviewed 55 out of 57 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
justfile	Updates workspace check logic to look for vendored `gix-imara-diff*` crates in the dependency tree.
gix-merge/fuzz/Cargo.toml	Switches fuzz crate to vendored `gix-imara-diff-01`.
gix-merge/Cargo.toml	Switches merge crate to vendored `gix-imara-diff-01`.
gix-imara-diff/tests/sliders.rs	Adds a placeholder/commented test file for slider-related experimentation.
gix-imara-diff/tests/integration/main.rs	Adds integration tests for tokenization, unified diff formatting, and word-diff behavior.
gix-imara-diff/tests/fixtures/make_git_diffs.sh	Adds a helper script intended to generate fixture diffs from real repos.
gix-imara-diff/src/util.rs	Adds shared utilities for common-prefix/suffix, sqrt approximation, and hunk navigation.
gix-imara-diff/src/unified_diff.rs	Adds unified diff formatting support (v0.2-style API).
gix-imara-diff/src/tests.rs	Adds unit tests targeting Myers split parity and unified diff output expectations.
gix-imara-diff/src/sources.rs	Adds token sources (lines/words/byte lines) used for interning/tokenization.
gix-imara-diff/src/slider_heuristic.rs	Adds slider heuristics (including indentation-based heuristic) for postprocessing.
gix-imara-diff/src/postprocess.rs	Adds postprocessing to slide/merge hunks using heuristics.
gix-imara-diff/src/myers/slice.rs	Adds Myers file-slice abstraction for divide-and-conquer.
gix-imara-diff/src/myers/preprocess.rs	Adds Myers preprocessing to prune tokens and improve performance.
gix-imara-diff/src/myers/middle_snake.rs	Adds middle-snake search implementation used by Myers split.
gix-imara-diff/src/myers.rs	Adds the Myers diff implementation (with heuristics/minimal mode).
gix-imara-diff/src/intern.rs	Adds token interning + `TokenSource`/`InternedInput` infrastructure.
gix-imara-diff/src/histogram/list_pool.rs	Adds pooled list storage used by histogram diff bookkeeping.
gix-imara-diff/src/histogram/lcs.rs	Adds histogram LCS search logic.
gix-imara-diff/src/histogram.rs	Adds histogram diff algorithm and fallback to Myers on repetitive inputs.
gix-imara-diff/README.md	Vendors upstream README (benchmarks, stability policy, fuzzing notes).
gix-imara-diff/LICENSE	Vendors upstream license text.
gix-imara-diff/fuzz/README.md	Adds documentation for running fuzz targets locally.
gix-imara-diff/fuzz/fuzz_targets/unified_diff_printer.rs	Adds fuzz target for unified diff formatting invariants.
gix-imara-diff/fuzz/fuzz_targets/postprocess_heuristics.rs	Adds fuzz target for postprocessing under different heuristics.
gix-imara-diff/fuzz/fuzz_targets/diff_compute_with.rs	Adds fuzz target for the `compute_with` API over raw token sequences.
gix-imara-diff/fuzz/fuzz_targets/comprehensive_diff.rs	Adds fuzz target combining diffing, postprocessing, and word-diff calls.
gix-imara-diff/fuzz/Cargo.toml	Defines the vendored fuzz package wiring to `gix-imara-diff`.
gix-imara-diff/fuzz/.gitignore	Ignores fuzz artifacts/corpus/coverage directories.
gix-imara-diff/Cargo.toml	Adds vendored crate manifest for v0.2.0 (`gix-imara-diff`).
gix-imara-diff/.gitignore	Ignores vendored crate build + bench data directories.
gix-imara-diff/.gitattributes	Normalizes EOL for fixture files.
gix-imara-diff-01/src/util.rs	Vendors v0.1 utility helpers used by Myers/histogram.
gix-imara-diff-01/src/unified_diff.rs	Vendors v0.1 unified diff sink/builder.
gix-imara-diff-01/src/tests.rs	Vendors v0.1 test suite for unified diff output.
gix-imara-diff-01/src/sources.rs	Vendors v0.1 token sources (including with/without line terminators).
gix-imara-diff-01/src/sink.rs	Vendors v0.1 sink abstraction and token insertion/removal counter.
gix-imara-diff-01/src/myers/slice.rs	Vendors v0.1 Myers file-slice abstraction.
gix-imara-diff-01/src/myers/preprocess.rs	Vendors v0.1 Myers preprocessing (including prefix/postfix stripping).
gix-imara-diff-01/src/myers/middle_snake.rs	Vendors v0.1 middle-snake search implementation.
gix-imara-diff-01/src/myers.rs	Vendors v0.1 Myers algorithm wiring + sink processing.
gix-imara-diff-01/src/lib.rs	Vendors v0.1 public API and algorithm selection.
gix-imara-diff-01/src/intern.rs	Vendors v0.1 interning infrastructure.
gix-imara-diff-01/src/histogram/list_pool.rs	Vendors v0.1 histogram list pool implementation.
gix-imara-diff-01/src/histogram/lcs.rs	Vendors v0.1 histogram LCS search.
gix-imara-diff-01/src/histogram.rs	Vendors v0.1 histogram diff wiring + fallback behavior.
gix-imara-diff-01/README.md	Vendors upstream README for v0.1 crate.
gix-imara-diff-01/LICENSE	Vendors upstream license text for v0.1 crate.
gix-imara-diff-01/Cargo.toml	Adds vendored crate manifest for v0.1.8 (`gix-imara-diff-01`).
gix-imara-diff-01/.gitignore	Ignores vendored v0.1 crate build + lock + bench data.
gix-imara-diff-01/.gitattributes	Normalizes EOL for v0.1 fixture files.
gix-diff/Cargo.toml	Repoints `imara-diff` and `imara-diff-v2` dependencies to vendored workspace crates.
Cargo.toml	Adds vendored crates to workspace members.

gix-imara-diff/src/unified_diff.rs

gix-imara-diff/tests/fixtures/make_git_diffs.sh

gix-diff/Cargo.toml

gix-imara-diff/README.md

gix-imara-diff-01/README.md

gix-imara-diff/src/histogram/list_pool.rs

Co-authored-by: Sebastian Thiel <sebastian.thiel@icloud.com>

Sebastian Thiel (Byron) · 2026-04-13T10:29:32Z

Whatever happened here, it's good and shows that 0.2 is the only version we'd want, particularly for the linecount. No need for extra complexity, apparently.

> cargo bench -p gix-diff --features blob-experimental
     Running benches/line_count.rs (target/release/deps/line_count-59627e52fae8c28f)
Gnuplot not found, using plotters backend
imara-diff 0.1          time:   [3.3171 ms 3.3266 ms 3.3363 ms]

imara-diff 0.2          time:   [2.2665 ms 2.2753 ms 2.2853 ms]
Found 9 outliers among 100 measurements (9.00%)

github-advanced-security bot found potential problems Apr 7, 2026

View reviewed changes

Sebastian Thiel (Byron) force-pushed the vendor-imara-diff branch 5 times, most recently from d22622d to 6d0fd32 Compare April 13, 2026 08:10

Sebastian Thiel (Byron) force-pushed the vendor-imara-diff branch from 6d0fd32 to fd49295 Compare April 13, 2026 08:18

Sebastian Thiel (Byron) marked this pull request as ready for review April 13, 2026 08:22

Sebastian Thiel (Byron) requested a review from Copilot April 13, 2026 08:24

Copilot started reviewing on behalf of Sebastian Thiel (Byron) April 13, 2026 08:25 View session

Sebastian Thiel (Byron) mentioned this pull request Apr 13, 2026

Add a test that reproduces a timeout issue in gix-merge. #2482

Closed

5 tasks

chatgpt-codex-connector bot reviewed Apr 13, 2026

View reviewed changes

gix-diff/Cargo.toml Show resolved Hide resolved

justfile Outdated Show resolved Hide resolved

Add a test that reproduces a timeout issue in gix-merge.

3bf4bdc

The problem was detected by Google OSS-Fuzz, and occurs because of window-size problem in imara-diff's preprocessing. With `gix-imara-diff` the problem is fixed as it's now vendored.

Copilot AI reviewed Apr 13, 2026

View reviewed changes

address auto-review

ecf0c17

Co-authored-by: Sebastian Thiel <sebastian.thiel@icloud.com>

Sebastian Thiel (Byron) force-pushed the vendor-imara-diff branch from 268ead4 to ecf0c17 Compare April 13, 2026 10:26

Sebastian Thiel (Byron) enabled auto-merge April 13, 2026 10:33

Sebastian Thiel (Byron) merged commit 8f091d1 into main Apr 13, 2026
32 checks passed

Sebastian Thiel (Byron) deleted the vendor-imara-diff branch April 13, 2026 10:44

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

vendor `imara-diff`#2506

vendor `imara-diff`#2506
Sebastian Thiel (Byron) merged 3 commits intomainfrom
vendor-imara-diff

Sebastian Thiel (Byron) commented Apr 7, 2026 •

edited

Loading

Uh oh!

github-advanced-security bot left a comment

Uh oh!

chatgpt-codex-connector bot left a comment

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Sebastian Thiel (Byron) commented Apr 13, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

Sebastian Thiel (Byron) commented Apr 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Tasks

Out of Scope

Uh oh!

github-advanced-security bot left a comment

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Sebastian Thiel (Byron) commented Apr 13, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Sebastian Thiel (Byron) commented Apr 7, 2026 •

edited

Loading