Stefan Sperling:

> The patch below was generated with the new diff code :)

Which already reveals a problem.  The patch doesn't apply, because
this hunk...

> --- lib/diff.c
> +++ lib/diff.c
[...]
> @@ -796,8 +834,20 @@ got_diff_objects_as_trees(struct got_object_id *id1, s
>  	arg.diff_context = diff_context;
>  	arg.ignore_whitespace = ignore_whitespace;
>  	arg.outfile = outfile;
> +	if (want_lineoffsets) {
> +		arg.line_offsets = *line_offsets;
> +		arg.nlines = *nlines;
> +	} else {
> +		arg.line_offsets = NULL;
> +		arg.nlines = 0;
> +	}
>  	err = got_diff_tree(tree1, tree2, label1, label2, repo,
> -	    got_diff_blob_output_unidiff, &arg, 1);
> +	    got_diff_blob_output_unidiff, &arg, 1);
> +
> +	if (want_lineoffsets) {
> +		*line_offsets = arg.line_offsets; /* was likely re-allocated */
> +		*nlines = arg.nlines;
> +	}
>  	    got_diff_blob_output_unidiff, &arg, 1);
>  done:
>  	if (tree1)
>  		got_object_tree_close(tree1);

... ends with an extraneous line and patch(1) becomes very confused:

|blob - 7e5ee06994a5158bc937ce8307bd19f51d7e0ef5
|blob + 9ed331be5c5b709d71129cb883b6c47a9f9c6a96
|--- lib/diff.c
|+++ lib/diff.c
--------------------------
Patching file lib/diff.c using Plan A...
Hunk #1 succeeded at 17.
Hunk #2 succeeded at 40.
Hunk #3 succeeded at 89.
Hunk #4 succeeded at 111.
Hunk #5 succeeded at 136.
Hunk #6 succeeded at 202.
Hunk #7 succeeded at 243.
Hunk #8 succeeded at 288.
Hunk #9 succeeded at 776.
Hunk #10 succeeded at 797.
Hunk #11 succeeded at 808.
Hunk #12 failed at 834.
1 out of 12 hunks failed
Hmm...  The next patch looks like a unified diff to me...
The text leading up to this was:
--------------------------
|               got_object_tree_close(tree1);
--------------------------
File to patch: No file found--skip this patch? [n] patch: **** can't find got_object_tree_close(tree1);

-- 
Christian "naddy" Weisgerber                          naddy@mips.inka.de