On Sun, Jan 19, 2020 at 02:33:34PM +0100, Martin Pieuchot wrote:
> To my understanding got_object_id_by_path() always return an error code
> if `obj_id' is NULL.  That means the checks below are redundant.
> 
> ok?

Yes, that's right. And tests are still passing with this. ok.

> diff --git got/got.c got/got.c
> index 5490ed3..f52a257 100644
> --- got/got.c
> +++ got/got.c
> @@ -2726,10 +2726,6 @@ cmd_blame(int argc, char *argv[])
>  	error = got_object_id_by_path(&obj_id, repo, commit_id, in_repo_path);
>  	if (error)
>  		goto done;
> -	if (obj_id == NULL) {
> -		error = got_error(GOT_ERR_NO_OBJ);
> -		goto done;
> -	}
>  
>  	error = got_object_get_type(&obj_type, repo, obj_id);
>  	if (error)
> diff --git tog/tog.c tog/tog.c
> index fc01887..8892f4b 100644
> --- tog/tog.c
> +++ tog/tog.c
> @@ -3616,8 +3616,6 @@ run_blame(struct tog_blame *blame, struct tog_view *view, int *blame_complete,
>  	err = got_object_id_by_path(&obj_id, repo, commit_id, path);
>  	if (err)
>  		return err;
> -	if (obj_id == NULL)
> -		return got_error(GOT_ERR_NO_OBJ);
>  
>  	err = got_object_get_type(&obj_type, repo, obj_id);
>  	if (err)
> 
>