Adch++ doesnt work on ARM architecture

Bug #930960 reported by oopepe
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
ADCH++
New
Undecided
Unassigned

Bug Description

Hi,

i compiled ADCH++ 2.7.0 on Debian Squeeze on an ARM machine (Seagate Dockstar). It starts, but when i try to connect via adc or adcs i get this error message "This hub requires BASE support". The TCP connection then closes.
I compiled the same version on an x86_64 machine and used the same config files and there it worked perfectly.
I compiled via "scons mode=release".
I watched at the TCP traffic via wireshark and this is what i see:
client --> hub: HSUP ADBAS0 ADBASE ADTIGR ADUCM0 ADBLO0
hub --> client: ISTA 240 This\shub\srequires\sBASE\ssupport
What can i do to make adch++ work on my ARM machine?

EDIT: I just tried version 2.8.0 in debug mode, but its the same problem. Heres the output:

Starting.Logging: 2012-02-12 14:52:15: core: Core initialized
.Processing HubName
Processing Description
Processing Log
Processing LogFile
Processing MaxSendSize
Processing MaxBufferSize
Processing BufferSize
Processing MaxCommandSize
Processing OverflowTimeout
Processing KeepSlowUsers
Processing BufferSize
Processing DisconnectTimeout
Processing LoginTimeout
.
2.8.0 (r0) Debug running, press ctrl-c to exit...
Logging: 2012-02-12 14:52:15: ScriptManager: Starting
Logging: 2012-02-12 14:52:15: PluginManager: Script.so loaded
Logging: 2012-02-12 14:52:15: SocketManager: Starting
Logging: 2012-02-12 14:52:15: SocketManager: Listening on 0.0.0.0:3002 (Encrypted: No)
Logging: 2012-02-12 14:52:15: SocketManager: Listening on [::]:3002 (Encrypted: No)
V4S7 entering IDENTIFY
V4S7 entering NORMAL
BNGJ connected
BNGJ ready
BNGJ disconnecting because 12
Removing BNGJ - AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
ManagedSocket deleted
SocketStream deleted

Regards
oopepe

description: updated
Revision history for this message
eMTee (realprogger) wrote :

What is the output in the debug console if you add the following patch http://pastebin.com/KUU1WBne and compile in debug mode?

Revision history for this message
oopepe (wurstwasserflocke) wrote :

Hi, i get this:

V4S7 entering IDENTIFY
V4S7 entering NORMAL
BNGJ connected
BNGJ ready
BASE FourCC convert: BASE
BNGJ disconnecting because 12
Removing BNGJ - AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
ManagedSocket deleted
SocketStream deleted

Revision history for this message
eMTee (realprogger) wrote :

At least it isn't an issue with some types and casting in supports... must be something else

Revision history for this message
oopepe (wurstwasserflocke) wrote :

Yes, maybe the message parsing? What can i do to find the problem?

Revision history for this message
Jacek Sieka (arnetheduck) wrote : Re: [Bug 930960] Re: Adch++ doesnt work on ARM architecture

I would guess it is an endianness bug - afair the parser does some
casting which is not safe. Not much to do except fix the bug..

> Yes, maybe the message parsing? What can i do to find the problem?

Revision history for this message
oopepe (wurstwasserflocke) wrote :

Hi, when i replace toFourCC with this

static uint32_t toFourCC(const char* x) { return (uint32_t)((x[3]<<24)|(x[2]<<16)|(x[1]<<8)|(x[0])); }

then it works.
But now i have another problem. Passwords with length >= 64 are always wrong. Passwords < 64 do work and i can successfully login to the hub.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.