Am 03.09.21 um 20:49 schrieb Junio C Hamano: > It does not seem to trigger with an empty range, e.g. > > $ git range-diff -Ix HEAD... > $ git range-diff -Ix HEAD~... > -: ---------- > 1: 0a0bc7d03a scalar: accept -C and -c opt... > > but when it needs real comparison, the -I seems to kill the command > quite easily. Reverting c45dc9cf30 (diff: plug memory leak from regcomp() on {log,diff} -I, 2021-02-11) fixes the segfault. diff.c::diff_free() frees the ignore_regex member of a struct diff_options, but doesn't reset the ignore_regex_nr nor clears the pointer. It is called from diff.c::diff_flush(). range-diff.c::output() calls diff.c::diff_flush() in a loop with the same struct diff_options (via range-diff.c::patch_diff()). So the second iteration of that loop tries to use the already freed ignore regexes. Here's a patch for that: --- >8 --- Subject: [PATCH] range-diff: avoid segfault with -I output() reuses the same struct diff_options for multiple calls of diff_flush(). Set the option no_free to instruct it to keep the ignore regexes between calls and release them explicitly at the end. Signed-off-by: René Scharfe <l.s.r@xxxxxx> --- Test missing because I couldn't see any effect of -I on range-diff. range-diff.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/range-diff.c b/range-diff.c index e731525e66..cac89a2f4f 100644 --- a/range-diff.c +++ b/range-diff.c @@ -482,6 +482,7 @@ static void output(struct string_list *a, struct string_list *b, else diff_setup(&opts); + opts.no_free = 1; if (!opts.output_format) opts.output_format = DIFF_FORMAT_PATCH; opts.flags.suppress_diff_headers = 1; @@ -542,6 +543,8 @@ static void output(struct string_list *a, struct string_list *b, strbuf_release(&buf); strbuf_release(&dashes); strbuf_release(&indent); + opts.no_free = 0; + diff_free(&opts); } int show_range_diff(const char *range1, const char *range2, -- 2.33.0