Hi apam,

I've started getting "Segmentation fault" error's with postie when it
tosses packets.

I have saved four of them for you to look at.

What's the best way to get them to you?



\/orlon
aka
Stephen


--- Talisman v0.46-dev (Linux/m68k)
* Origin: Vorlon Empire: Amiga 3000 powered in Sector 550 (21:1/195.1)

I've started getting "Segmentation fault" error's with postie when it
tosses packets.

I have saved four of them for you to look at.

What's the best way to get them to you?

Email

Andrew


--- Talisman v0.46-dev (Windows/x64)
* Origin: Smuggler's Cove - Private BBS (21:1/182)

Hi apam,

I've started getting "Segmentation fault" error's with postie when
it tosses packets.

I have saved four of them for you to look at.
What's the best way to get them to you?

Email

I can't seam to find it in my message base, so here is a link to it:

https://bester.activateit.net.au/index.php/s/J56Tr938AQKZ3N4

Postie leaves the pid file behind, and the only log info is that it's
started, and the console returns with "segment fault".



\/orlon
aka
Stephen


--- Talisman v0.46-dev (Linux/m68k)
* Origin: Vorlon Empire: Amiga 3000 powered in Sector 550 (21:1/195.1)

vorlon wrote (2023-01-26):

Hi apam,

I've started getting "Segmentation fault" error's with postie when
it tosses packets.

I have saved four of them for you to look at.
What's the best way to get them to you?

Email

I can't seam to find it in my message base, so here is a link to it:

https://bester.activateit.net.au/index.php/s/J56Tr938AQKZ3N4

Postie leaves the pid file behind, and the only log info is that it's started, and the console returns with "segment fault".

I checked first and second packet and they look fine as far as I can tell.

$ ./pktview /tmp/c434b101.pkt.toss
PKTView - Fidonet *.PKT viewer - last update 1999-07-27
Written by Tom Torfs, free for all use

/tmp/apam/c434b101.pkt.toss:
Detected Type 2+ packet
--- PACKET HEADER ---
Packet version : 2
Capability word : 0001
Capability validation : 0100
Product code : 16.255
Product revision : 1.9
Product specific info : 00000000
Date and time : 2023-01-16 04:13:50
Originating zone : 21
Originating net : 1
Originating node : 195
Originating point : 0
Destination zone : 21
Destination net : 1
Destination node : 195
Destination point : 1
Auxiliary net : 0
Baudrate : 0
Password :
--- PACKED MESSAGE #1 ---
Version : 2
Originating net : 1
Originating node : 195
Destination net : 1
Destination node : 195
Cost : 0
Attributes :
Date and time : 15 Jan 23 11:21:56
From : XXXXXXXXX
To : XXXXXXXXXXXXXXXXXX
Subject : Re: Any Reccomendations?
Control information:
AREA:FSX_VIDEO
[...]



⁂

---
* Origin: War is Peace. Freedom is Slavery. Ignora

Postie leaves the pid file behind, and the only log info is that it's started, and the console returns with "segment fault".

I've commited a fix, I don't know if it will fix the issue or not.

My guess is it's something to do with reply linking, because at first
glance it seems ratsoft puts dodgy reply ids (7 hex chars, the 8th char
is missing) and my guess is it's triggering a bug in postie.

I don't have my m68k vm set up at present, if this quick fix doesn't do
it i'll look at it more in depth later.

Andrew


--- Talisman v0.46-dev (Windows/x64)
* Origin: Smuggler's Cove - Private BBS (21:1/182)

I checked first and second packet and they look fine as far as I can
tell.

Yeah, if it's a segfault it's a bug in postie. It shouldn't crash even if
the packets are bad which as you say they are not.

Andrew


--- Talisman v0.46-dev (Windows/x64)
* Origin: Smuggler's Cove - Private BBS (21:1/182)

apam wrote (2023-01-26):

My guess is it's something to do with reply linking, because at first glance it seems ratsoft puts dodgy reply ids (7 hex chars, the 8th char
is missing) and my guess is it's triggering a bug in postie.

Good catch. I looked at the MSGID and REPLY, but I didn't thought about counting the hex chars. The REPLY is indeed missing the last char from the original MSGID. I guess it's impossible to fix that 28 year old Atari BBS / Fido software. Maybe Ratsoft 2.1c (1/9/96) has a fix?
https://milan.kovac.cc/atari/software/?folder=/COMM/BBS#RATSOFT.FTP

The buggy one says:
=== RATSoft / ST ===
by Steve Hughey, (c)1991-1995 R0dent Labs Software
version 2.1c [10/21/95 - 12:16pm]

I'll ask the sysop, if he is aware of the bug ...

⁂

---
* Origin: War is Peace. Freedom is Slavery. Ignorance is Strength. (21:3/102)

Hi apam,

I've commited a fix, I don't know if it will fix the issue or not.

Updated and compiled. Just need to wait....



\/orlon
aka
Stephen


--- Talisman v0.46-dev (Linux/m68k)
* Origin: Vorlon Empire: Amiga 3000 powered in Sector 550 (21:1/195.1)

Hi apam,

I've commited a fix, I don't know if it will fix the issue or not.

I've now seen a few messages come in from "Commodore Clifford", and
postie hasn't crashed.

Now the only anoyying thing is that messages get tossed out of order.
I see responses to posts before the post regularly.

This point system only generally polls once, sometimes twice a day. It receives/sends only raw packets, so it would be great if postie could
toss the messages in the correct order.



\/orlon
aka
Stephen


--- Talisman v0.46-dev (Linux/m68k)
* Origin: Vorlon Empire: Amiga 3000 powered in Sector 550 (21:1/195.1)