diff: EOL/header normalization - empty time_format (git-style header), ignore_missing_final_newline - #234
Merged
Merged
Conversation
An empty DiffOptions::time_format now yields a git-style file header (`--- name` / `+++ name`, name only, no per-file mtime) instead of `name ` with a trailing space and empty timestamp. This makes unified/context diff output reproducible across machines and time zones, which matters for golden tests and content-addressed comparisons. A non-empty time_format is unchanged; time_format stays library-only. Test: DiffTest.EmptyTimeFormatOmitsHeaderTimestamp.
DiffOptions::ignore_missing_final_newline (the --ignore_missing_final_newline CLI flag and the diff_test `ignore_missing_final_newline` rule attribute) makes a file with and one without a trailing newline compare equal: the `\ No newline at end of file` marker is suppressed in Data's shared line preprocessing, so all three algorithms honor it. It only ignores the terminator - the zero-length-input guard still fires on the real got_nl, so an empty file stays distinct from a single empty line. This is the one whitespace/EOL normalization mbo::diff was missing; leading / consecutive / trailing space and CRLF-vs-LF (a trailing '\r' is trailing whitespace) are already covered by ignore_all_space / ignore_consecutive_space / ignore_trailing_space. Tests: DiffTest.IgnoreMissingFinalNewline (equal with, differ without, empty != "\n") and an AlgorithmFeatureMatrix case across naive/myers/direct.
Fab-Cat
approved these changes
Jul 5, 2026
helly25
added a commit
to helly25/xff
that referenced
this pull request
Jul 5, 2026
Re-pins the mbo git_override to the main commit merging helly25/mbo#234 (which added the empty-time_format git-style header and ignore_missing_final_newline), and wires the two xff-side bits: - -diff now sets time_format="" so the unified/context header omits the per-file mtime (`--- a/one.txt`), making the output reproducible. The diff golden tests drop the mtime-strip sed (normalize) and the fixture drops its `touch` - the git-style header is verbatim, so the committed goldens match directly. - --diff-ignore gains an `eofnl` token (-> mbo ignore_missing_final_newline): a file with and one without a final newline compare equal. No `lead`/`eol` token - leading whitespace is subsumed by `change`/`ws`, and CRLF-vs-LF by `trail` (a `\r` is trailing whitespace). Tests: run_test + diff_test.sh gain eofnl cases (equal with the token, differ without); the diff goldens pass with no normalization. `bazel test //xff/...` green (66 tests).
helly25
added a commit
to helly25/xff
that referenced
this pull request
Jul 12, 2026
mbo 0.13.0 is now in the Bazel Central Registry, so drop the temporary git_override (pinned at the main commit merging helly25/mbo#234) for a plain bazel_dep version bump. 0.13.0 carries mbo/diff (the rich diff API behind -diff, #88) and mbo/digest (the -hash algorithms, #105), which the override provided; it also keeps the macOS clang+asan verification (#188) and StatusIs code-matcher support (#190) our test conventions rely on. Comment refreshed to match. Test: bazel test //xff/cli:hash_test //xff/cli:diff_test //xff/examples:cookbook_test (mbo/digest + mbo/diff + smoke) pass against the BCR 0.13.0; pre-commit green.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Two
mbo::diffnormalization features that make diff output and comparison reproducible:DiffOptions::time_formatomits the header timestamp: unified/context file headers become git-style--- name/+++ name(name only, no per-file mtime) instead ofnamewith a trailing space and empty timestamp. Output is reproducible across machines and time zones (golden tests, content-addressed comparisons). Non-emptytime_formatunchanged; stays library-only.ignore_missing_final_newline(DiffOptionsfield,--ignore_missing_final_newlineCLI flag,diff_testrule attribute): a file with and one without a trailing newline compare equal; the\ No newline at end of filemarker is suppressed inData's shared line preprocessing so all three algorithms (naive/myers/direct) honor it identically. Only the terminator is ignored - an empty file stays distinct from a single empty line.This completes the whitespace/EOL normalization set alongside
ignore_all_space/ignore_consecutive_space/ignore_trailing_space.Testing
DiffTest.EmptyTimeFormatOmitsHeaderTimestampDiffTest.IgnoreMissingFinalNewline(equal with option, differ without, empty != "\n")AlgorithmFeatureMatrixcase across naive/myers/direct