On Thu, Dec 04, 2014 at 02:39:34PM +0100, Axel Beckert wrote:
>
> So there are indeed a small number of code locations which look
> suspiciously similar.
There are other similarities between the initial commit of abduco and
dtach, such as the code declaring abduco's client to server protocol:
[abduco/abduco.c lines 61-67] |[dtach-0.8/dtach.h lines 89-96]
enum PacketType { |enum
MSG_CONTENT = 0, |{
MSG_ATTACH = 1, | MSG_PUSH = 0,
MSG_DETACH = 2, | MSG_ATTACH = 1,
MSG_RESIZE = 3, | MSG_DETACH = 2,
MSG_REDRAW = 4, | MSG_WINCH = 3,
}; | MSG_REDRAW = 4,
|};
[abduco/abduco.c lines 69-78] |[dtach-0.8/dtach.h lines 106-116]
/* packet sent from client to server */ |/* The client to master protocol. */
typedef struct { |struct packet
unsigned char type; |{
unsigned char len; | unsigned char type;
union { | unsigned char len;
char msg[sizeof(struct w| union
insize)]; | {
struct winsize ws; | unsigned char buf[sizeo
int i; |f(struct winsize)];
} u; | struct winsize ws;
} ClientPacket; | } u;
|};
and the same escape sequence trick that dtach uses to move to the bottom
of the screen along with similar code for handling the suspend key:
[abduco/abduco.c line 161] |[dtach-0.8/dtach.h line 128]
fprintf(stdout, "\e[999H\r\n"); |#define EOS "\033[999H"
[abduco/client.c lines 126-133] |[dtach-0.8/attach.c lines 105-115]
} else i| if (!no_suspend && (pkt->u.buf[
f (pkt.u.msg[0] == cur_term.c_cc[VSUSP])|0] == cur_term.c_cc[VSUSP]))
{ | {
| /* Tell the master that
pkt.type = MSG_DETACH; | we are suspending. */
| pkt->type = MSG_DETACH;
client_send_packet(&pkt); | write(s, pkt, sizeof(st
|ruct packet));
tcsetattr(0, TCSADRAIN, &orig_term); |
| /* And suspend... */
client_show_cursor(); | tcsetattr(0, TCSADRAIN,
| &orig_term);
info(NULL); | printf(EOS "\r\n");
| kill(getpid(), SIGTSTP)
kill(getpid(), SIGTSTP); |;
| tcsetattr(0, TCSADRAIN,
tcsetattr(0, TCSADRAIN, &cur_term); | &cur_term);
Note that in abduco, client_send_packet(&pkt) will write the raw
contents of pkt to a socket, and info(NULL) will write the "\e[999H\r\n"
escape sequence to stdout.
> It should also be mentioned that the initial reporter of this issue,
> Ned, is the author of dtach, see http://dtach.sourceforge.net/.
>
> I neither claim that this is a bad sign nor that this is a good sign.
>
> I though consider it suspicious that this was not mentioned in the
> original report, hence I think it's important to mention it myself.
>
I'm not sure why this would be suspicious, however I do confirm that I
am the author of dtach.
--
Ned T. Crigler
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
