From: Stefan Sperling <stsp@stsp.name>
Subject: Re: Fix another edscript() segfault
To: Christian Weisgerber <naddy@mips.inka.de>
Cc: gameoftrees@openbsd.org
Date: Sat, 16 Oct 2021 15:58:05 +0200

On Sat, Oct 16, 2021 at 03:27:00PM +0200, Christian Weisgerber wrote:
> fix another segfault in edscript()
> 
> Same problem as fixed in 90de04bc7f1b, but a few lines further down.
> Found by a random segfault while running regress.
> 
> ==== update ====
> ./update.sh -q -r "/tmp"
> Segmentation fault (core dumped)
> 
> Core was generated by `got'.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0  0x000000d506bae148 in edscript (n=1, d3s=0xd747ea7380)
>     at ../lib/diff3.c:1051
> 1051                                line[linelen] == '\n' ? ":" : "",

Thanks, ok!

> diff 2e3a6c3be72d70d3e58780d82d7545cd0a3d9e82 /home/naddy/got
> blob - 9b918583a6db04e4fff29bcb04a7490e23db43cc
> file + lib/diff3.c
> --- lib/diff3.c
> +++ lib/diff3.c
> @@ -1048,7 +1048,7 @@ edscript(int n, struct diff3_state *d3s)
>  				goto done;
>  		} else {
>  			err = diff_output(d3s->diffbuf, "%s%s\n.\n",
> -			    line[linelen] == '\n' ? ":" : "",
> +			    linelen > 0 && line[linelen] == '\n' ? ":" : "",
>  			    d3s->f3mark);
>  			if (err)
>  				goto done;
> -- 
> Christian "naddy" Weisgerber                          naddy@mips.inka.de
> 
>