Re: Bug#146964: diff: Argument --horizon-lines documented, handled, not used (fwd)

[Paul Eggert]

> From: Santiago Vila <address@hidden>
> Date: Sun, 26 May 2002 01:44:58 +0200 (CEST)
> 
> I received this report from the Debian bug system. By looking at the
> code I believe this is already fixed in diffutils-2.8, but I didn't
> find yet a test case. It's fixed?

It's not clear to me, because the original bug report seems to be
flawed.  The bug report says:

> Judging both by the output and by diff.c (snippet below),
> --horizon-lines is ignored.
> 
>         case 140:
>           if (ck_atoi (optarg, &horizon_lines) || horizon_lines < 0)
>             fatal ("horizon must be a nonnegative integer");
>           break;
> 
> Note that nothing seems to be actually done with the argument.

But if memory served, ck_atoi stored into horizon_lines, so it
actually did do something with its argument.

As you mention, the code is different in diffutils 2.8.1, and it now
clearly modifies horizon_lines, but I'm not sure that this fixes the
bug, whatever it was.  The original bug report doesn't provide a test
case that illustrates the bug, so it's not clear to me what the bug
was.