Commit Briefs

Stefan Sperling

print trailing 'No newline at EOF' message for final chunk only (main)

This makes test 126 pass.


Stefan Sperling

add a test which triggers redundant 'No newline at end of file' messages

The message appears after each hunk, but should only appear at the end of the entire diff. Regression introduced in commit 4125b1eff8d9dfcd961dc70ebbd5d54c97bdc3f6 "diff: handle missing new lines in trailing context" problem noticed by op@


Stefan Sperling

back out "improve accuracy of function prototype in hunk headers"

which was commit 8993f42562696079866fc2dec7191877b6cc1f18 This commit introduced a regression where hunk headers would contain lines taken from below or within a change. Reported by tb@ and sthen@. There is some confusion around how the -p feature was documented: -p With unified and context diffs, show with each change the first 40 characters of the last line before the context beginning with a letter, an underscore or a dollar sign. For C source code following standard layout conventions, this will show the prototype of the function the change applies to. The original behaviour matches "the last line before the context beginning with a letter ...". The new (reverted) behaviour sometimes matches "will show the prototype of the function the change applies to". This did always happen as intended, as seen when some .c files in the OpenBSD src tree were changed. The new behaviour was also worse for Makefiles in the OpenBSD ports tree. Let's revert to the traditional -p behaviour until a better fix is found. Details: https://marc.gameoftrees.org/mail/1692696001.69872_0.html https://marc.info/?l=gameoftrees&m=169269624516330&w=2 ok tb@


Christian Weisgerber

do not treat \r\n line endings as special

Previously, we would go out of our way to handle both \n and \r\n line endings as equivalent. However, for files with \r\n line endings, this results in diffs that omit the final \r and therefore cannot be applied with patch(1) on Unix platforms. ok stsp


Mark Jamsek

make diff_chunk_type() public and clarify comment

As discussed with stsp, reword an easily misunderstood comment, and move diff_chunk_type() into the public diff API to improve caller efficiency. ok stsp@


Branches

Tags

This repository contains no tags

Tree

.gitignorecommits | blame
LICENCEcommits | blame
READMEcommits | blame
compat/
diff/
diff-version.mkcommits | blame
include/
lib/
man/
test/

README

This is a collection of diff algorithms, to test various combinations.

The initial aim was to provide a faster diff implementation for got
(gameoftrees.org) with a BSD license, at the u2k20 OpenBSD hackathon.
A side effect could be improving OpenBSD's /usr/bin/diff utility.

At the time of writing, this is little more than a playground / benchmark basis
/ diff algorithm analysis platform. What could be done:
- add profiling and test series to rate diff algorithm combinations.
- interface with / merge into got.

The Myers and Patience Diff algorithm implementations found here are based on
the explanations found in these blog post series:
  https://blog.jcoglan.com/2017/02/12/the-myers-diff-algorithm-part-1/ ff.
and
  https://blog.jcoglan.com/2017/09/19/the-patience-diff-algorithm/ ff.
-- possibly the single most comprehensive explanations of these algorithms.
Many thanks for this valuable door opener!
The source code itself is not based on the code found in those blogs, but
written from scratch with the knowledge gained.

Compile:
  make -C diff

Test:
  make -C test/