Re: [PATCH v5] diff: ensure consistent diff behavior with ignore options

[Johannes Schindelin <Johannes.Schindelin@gmx.de>]

[-- Attachment #1: Type: text/plain, Size: 2120 bytes --]

Hi,

On Fri, 8 Aug 2025, Lidong Yan wrote:

> @@ -6778,8 +6816,15 @@ void diff_flush(struct diff_options *options)
>  			     DIFF_FORMAT_CHECKDIFF)) {
>  		for (i = 0; i < q->nr; i++) {
>  			struct diff_filepair *p = q->queue[i];
> -			if (check_pair_status(p))
> -				flush_one_pair(p, options);
> +
> +			if (!check_pair_status(p))
> +				continue;
> +
> +			if (options->flags.diff_from_contents &&
> +			    !diff_flush_patch_quietly(p, options))
> +				continue;
> +
> +			flush_one_pair(p, options);
>  		}
>  		separator++;
>  	}

I'll play Ἑρμῆς here and relay the bug report from
https://github.com/git/git/commit/77f8e1002be5c736e064ed0e656ba51c82d85b5d#r168025843:

fcharlie wrote:

	This code will cause git diff --patch --raw to produce garbled output
	under certain conditions (e.g., using --ignore-space-change and whitespace
	changes). flush_one_pair will output --raw format, but because
	diff_flush_patch_quietly has already output a patch, garbled output will
	be produced (actually, no diff_flush_patch_quietly should be called here).

	[...]

A few exchanges later, brandb97 (AKA Lidong Yan) suggested this patch:

	diff --git a/diff.c b/diff.c
	index 87fa16b730..4baf9b535e 100644
	--- a/diff.c
	+++ b/diff.c
	@@ -1351,6 +1351,9 @@ static void emit_diff_symbol_from_struct(struct diff_options *o,
		int len = eds->len;
		unsigned flags = eds->flags;
	 
	+       if (o->dry_run)
	+               return;
	+
		switch (s) {
		case DIFF_SYMBOL_NO_LF_EOF:
			context = diff_get_color_opt(o, DIFF_CONTEXT);
	@@ -4420,7 +4423,7 @@ static void run_external_diff(const struct external_diff *pgm,
	 {
		struct child_process cmd = CHILD_PROCESS_INIT;
		struct diff_queue_struct *q = &diff_queued_diff;
	-       int quiet = !(o->output_format & DIFF_FORMAT_PATCH);
	+       int quiet = !(o->output_format & DIFF_FORMAT_PATCH) || o->dry_run;
		int rc;
	 
		/*

fcharlie then responded by promising to test that patch tomorrow.

I will hold off from releasing Git for Windows v2.51.1 to await the result
of this test.

Ciao,
Johannes