user/sven/git.git/contrib/diff-highlight, branch master

diff-highlight: fetch all config with one process

2026-03-23T14:42:27Z

When diff-highlight was written, there was no way to fetch multiple config keys _and_ have them interpreted as colors. So we were stuck with either invoking git-config once for each config key, or fetching them all and converting human-readable color names into ANSI codes ourselves. I chose the former, but it means that diff-highlight kicks off 6 git-config processes (even if you haven't configured anything, it has to check each one). But since Git 2.18.0, we can do: git config --type=color --get-regexp=^color\.diff-highlight\. to get all of them in one shot. Note that any callers which pass in colors directly to the module via @OLD_HIGHLIGHT and @NEW_HIGHLIGHT (like diff-so-fancy plans to do) are unaffected; those colors suppress any config lookup we'd do ourselves. You can see the effect like: # diff-highlight suppresses git-config's stderr, so dump # trace through descriptor 3 git show d1f33c753d | GIT_TRACE=3 diff-highlight 3>&2 >/dev/null which drops from 6 lines down to 1. Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: allow module callers to pass in color config

2026-03-23T14:42:27Z

Users of the module may want to pass in their own color config for a few obvious reasons: - they are pulling the config from different variables than diff-highlight itself uses - they are loading the config in a more efficient way (say, by parsing git-config --list) and don't want to incur the six (!) git-config calls that DiffHighlight.pm runs to check all config Let's allow users of the module to pass in the color config, and lazy-load it when needed if they haven't. Signed-off-by: Scott Baker Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: test color config

2026-03-23T14:42:27Z

We added configurable colors long ago in bca45fbc1f (diff-highlight: allow configurable colors, 2014-11-20), but never actually tested it. Since we'll be touching the color code in a moment, this is a good time to beef up the tests. Note that we cover both the highlight/reset style used by the default colors, as well as the normal/highlight style added by that commit (which was previously totally untested). Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: use test_decode_color in tests

2026-03-23T14:42:27Z

The diff-highlight tests use raw color bytes when comparing expected and actual output. Let's use test_decode_color, which is our usual technique in other tests. It makes reading test output diffs a bit easier, since you're not relying on your terminal to interpret the result (or worse, interpreting characters yourself via "cat -A"). This will also make it easier to add tests with new colors/attributes, without having to pre-define the byte sequences ourselves. Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: check diff-highlight exit status in tests

2026-03-23T14:42:26Z

When testing diff-highlight, we pipe the output through a sanitizing function. This loses the exit status of diff-highlight itself, which could mean we are missing cases where it crashes or exits unexpectedly. Use an extra tempfile to avoid the pipe. Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: drop perl version dependency back to 5.8

2026-03-23T14:42:26Z

The diff-highlight code does not rely on any perl features beyond what perl 5.8 provides. We bumped it to v5.26 along with the rest of the project's perl scripts in 702d8c1f3b (Require Perl 5.26.0, 2024-10-23). There's some value in just having a uniform baseline for the project, but I think diff-highlight is special here: - it's in a contrib/ directory that is not frequently touched, so there is little risk of Git developers getting annoyed that modern perl features are not available - it provides a module used by other projects. In particular, diff-so-fancy relies on DiffHighlight.pm but does not otherwise require a perl version more modern than 5.8. Let's drop back to the more conservative requirement. Signed-off-by: Scott Baker Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

diff-highlight: mention build instructions

2026-03-23T14:42:26Z

Once upon a time, this was just a script in a directory that could be run directly. That changed in 0c977dbc81 (diff-highlight: split code into module, 2017-06-15). Let's update the README to make it more clear that you need to run make. Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

contrib/diff-highlight: do not highlight identical pairs

2026-03-20T05:29:14Z

We pair lines for highlighting based on their position in the hunk. So we should never see two identical lines paired, like: -one -two +one +something else which would pair -one/+one, because that implies that the diff could easily be shrunk by turning line "one" into context. But there is (at least) one exception: removing a newline at the end of a file will produce a diff like: -foo +foo \No newline at end of file And we will pair those two lines. As a result, we end up marking the whole line, including the newline, as the shared prefix. And there's an empty suffix. The most obvious bug here is that when we try to print the highlighted lines, we remove the trailing newline from the suffix, but do not bother with the prefix (under the assumption that there had to be a difference _somewhere_ in the line, and thus the prefix would not eat all the way up to the newline). And so you get an extra line like: -foo +foo \No newline at end of file This is obviously ugly, but also causes interactive.diffFilter to (rightly) complain that the input and output do not match their lines 1-to-1. This could easily be fixed by chomping the prefix, too, but I think the problem is deeper. For one, I suspect some of the other logic gets confused by forming an array with zero-indexed element "3" in a 3-element array. But more importantly, we try not to highlight whole lines, as there's nothing interesting to show there. So let's catch this early in is_pair_interesting() and bail to our usual passthrough strategy. Reported-by: Scott Baker Signed-off-by: Jeff King Signed-off-by: Junio C Hamano

Merge branch 'jc/diff-highlight-main-master-testfix'

2026-02-13T21:39:24Z

Test fix (in contrib/) * jc/diff-highlight-main-master-testfix: diff-highlight: allow testing with Git 3.0 breaking changes

diff-highlight: allow testing with Git 3.0 breaking changes

2026-02-03T21:27:35Z

The diff-highlight (in contrib/) comes with its own test script, which relies on the initial branch name being 'master'. This is not just encoded in the test logic, but in the illustration in the file that shows the topology of the history. Force the initial branch name to 'master' to allow it pass. Signed-off-by: Junio C Hamano