Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[gaglia]

I've finally decided to install amuled + amuleweb cvs from the vollstrecker repository. I run a low-power low-cost server with no graphical server, controlled via ssh. The other pc's in my home lan use amuled via the web interface. Or at least they should...

Since 2.1.3 there was a problem in doing so that made impossible for me to use amule this way, forcing me to resort to emule on windows. Everybody told me that a cvs version would have resolved all my issues, so after many months waiting for 2.2.0 i decided to try.

This morning i installed the last cvs and the problem still persists (altough in a somehow less "tragical" way). The problem is: after some time, if i open the amuleweb panel with the browser and go in the KAD nodes graph tab, amuleweb crashes: segmentation fault.

I've tried to manually start amuled without amuleweb, then started amuleweb via ssh to see detailed output. That is what happens when, after few minutes, i visit the kad panel with the browser:

Code: [Select]

WCThread: exited [WebSocket closed]
WCThread: exited [WebSocket closed]
WCThread: Started a new WCThread
Session ok, logged in
WCThread: exited [WebSocket closed]
WCThread: Started a new WCThread
Processing request [original]: amuleweb-main-kad.php
Session ok, logged in
Processing request [redirected]: amuleweb-main-kad.php
WCThread: exited [WebSocket closed]
WCThread: Started a new WCThread
Session ok, logged in
Segmentation fault

I'm not an expert but it seems to me like one of the two:
1) an overflow on the number of nodes
2) some sort of memory leak (not correctly deallocated memor)

EDIT: at the same moment KAD disconnects, i can see this using amulecmd. Only solution is to restart amuled first, then amuleweb.

Can someone confirm? Thanks

[RRT4]

Can someone confirm? Thanks

yes, from nearly two years

i have another issue with amuleweb...

http://forum.amule.org/index.php?topic=8361.msg46953#msg46953

[gaglia]

This is a very bad news.

How is it possible that no priority has been given to this problem? Don't misunderstand me: i think the developers' work is something wonderful (and free...). But i'd like to point at the following problem: the ed2k network is dieing (if not already died).

I suppose most of you are already aware of the massive ed2k server loss in the last weeks. And suppose those who know this, also know that this is because of a legal action by RIAA & co. True, ed2k is still superior to KAD in number of sources available... but how long is this going to last? And - above all - how much is ed2k still safe in respect to traffic analysis and data mining?

The things are going to worsen. Every day they do.

I live in Italy, the laws here are quite obsessive in this respect. Our ISPs are forced to keep track of every connection for many years, thus providing a powerful tool to be used by laws enforcement. But our legal system is strongly influenced by the majors too, and can force to reveal a IP's owner in many circumstances. This mix is lethal: some months ago, for example, Peppermint managed to obtain the name of the owners of some IPs who were reported to share illegal mp3s, with a large scale, massive scan of the main p2p networks (in partial violation of the european laws on privacy protection, in fact this scan was done by a swiss company). I'm not a law expert, and i don't know the details of this story... but i think that if a multibillionar music major manages to persuade the autorithies to persecute some user NOT for having done some big criminal/pedo/terroristic thing... but for having shared some mp3s... there is something wrong IMHO (in fact the sentence has recently been temporally blocked by the Costitutional Court).

However this is another matter entirely... I'm not justifying the use of p2p for unlawful purposes (none of us do this ). But i'm very worried about the netizen's privacy, the neutrality of the web and the future of the e-rights. In this matter you have to consider that KAD, altough offering no real anonimity, makes this "broad scale privacy violations" something more difficult, and - above all - it is resistant to large scale attacks to the network (i.e.: shutdowns of the servers), because it's totally decentralized.

What am i saying is that in the development of aMule, developers should be prevident, and give priority to KAD. Under this point of view, Windows users are somehow already protected, because eMule supports KAD perfectly. I myself use eMule on windows connecting ONLY to KAD, i stopped using ed2k by some years, and never felt the missing. I can share and find sources almost as good as with ed2k, and i encourage everyone to do the same. But since i'm passed to linux this is a big problem: my low-power server cannot afford to run aMule, i must resort to amuled+amuleweb - and therefore cannot use KAD efficently.

This is a major drawback for many linux users i've known. You know well that in the *nix world great importance is given to text-based shell, and the GUI is somewhat "accessory". We all would be VERY grateful if this issue will be totally solved.

That said, thank you very much for the already wonderful work

[lfroen]

First of all, thanks for reporting the problem. Now, can you please describe under which circumstances amuleweb crashing?

I suppose most of you are already aware of the massive ed2k server loss in the last weeks. And suppose those who know this, also know that this is because of a legal action by RIAA & co

Last time I checked, US laws stopped at US borders.

This is a major drawback for many linux users i've known. You know well that in the *nix world great importance is given to text-based shell

That's a myth.

Our ISPs are forced to keep track of every connection for many years

Every TCP connection?! Pleeeese, get real.

[gaglia]

First of all, thanks for reporting the problem. Now, can you please describe under which circumstances amuleweb crashing?

Let me try to be more complete:
1) amuled starts, with AUTOCONNECT=1  ACCEPT_EXT_CONNECTIONS=1  USE_ED2K=0  USE_KAD=1
2) amuleweb starts, connects to amuled and initializes the web server
3) amuled connects to KAD and begins to look for peers/sources
4) my browser connects to amuleweb, i authenticate myself and get access to the web interface
5) everything works fine for some minutes/hours, i can do searches, download/upload (ports are fowarded correctly on my router) etc
6) abruptly connection to KAD is lost, and the number of nodes drops to zero
7) download/upload already initialized continues normally (without finding new sources), no misfunction noticeable, save for the fact that there is no way of reconnecting to KAD
from the point 6) thereafter, if i try from the web interface on my browser to click on the "KAD" panel to view the graph of the number of nodes, amuleweb gets a sementation fault. Amuled continues to run as if nothing has happened, but still can't connect to KAD
9) only solution is to restart amuled. Me.Patience--
10) if (Me.Patience > 0) goto 1); else return 0

However i've read in this forum many users with the same problem, at least in the italian section. And dozens with very similar ones (amuled+amuleweb crashes / KAD disconnections).

Last time I checked, US laws stopped at US borders.

Besides the fact that this may be (unfortunately) opinable... last time i checked, US was not the only country to have some sort of copyright protection law
Moreover, most of the ed2k servers are (were) on US ground.

That's a myth.

That's not. It's just the right way things are done. I'm not saying linux is *all* text-based (in fact it's not), but it is born as such (just as MS-DOS is).
If you have just a little experience with *nix you should know that a GUI is somehow an "accessory" for most of the applications. A very useful one, but always an accessory.

Every TCP connection?! Pleeeese, get real.

Yeah. Sad but true. Not the CONTENT of the trasmissions obviously, but data such as ip_dest:ip:source:hour_and_date. For a period of time ranging from 6 to 24 months, depending on the single State. Not only in Italy, that's a UE standard. It's very expensive for our ISPs - and so it is for their customers

[lfroen]

Let me try to be more complete:
...
6) abruptly connection to KAD is lost, and the number of nodes drops to zero
However i've read in this forum many users with the same problem, at least in the italian section. And dozens with very similar ones (amuled+amuleweb crashes / KAD disconnections).

It's possible that number of nodes=0 causing amuleweb to segfault. I will check the issue. This have nothing to do with "disconnecting KAD" problem, which is up to Kry to investigate.

My comments to some off-topic ideas:

Besides the fact that this may be (unfortunately) opinable... last time i checked, US was not the only country to have some sort of copyright protection law

So what? Copyright protection law have nothing to do with it.

Moreover, most of the ed2k servers are (were) on US ground.

So american point of view. "Coast-to-coast" is not whole world. I'm not connected to a single US-based server. Does it surprise you?

That's a myth.

That's not. It's just the right way things are done. I'm not saying linux is *all* text-based (in fact it's not), but it is born as such (just as MS-DOS is).
If you have just a little experience with *nix you should know that a GUI is somehow an "accessory" for most of the applications. A very useful one, but always an accessory.

I don't really understand what does it have to do with anything. Is aMule an "accessory"? What does it mean? Don't know.

Every TCP connection?! Pleeeese, get real.

Yeah. Sad but true. Not the CONTENT of the trasmissions obviously, but data such as ip_dest:ip:source:hour_and_date. For a period of time ranging from 6 to 24 months, depending on the single State. Not only in Italy, that's a UE standard. It's very expensive for our ISPs - and so it is for their customers

You know what? Suppose you right. It's feasible to record sip/dip/date/time. So what? I can't see any point here.

[Starchild]

That's a myth.

That's not. It's just the right way things are done. I'm not saying linux is *all* text-based (in fact it's not), but it is born as such (just as MS-DOS is).
If you have just a little experience with *nix you should know that a GUI is somehow an "accessory" for most of the applications. A very useful one, but always an accessory.

I don't really understand what does it have to do with anything. Is aMule an "accessory"? What does it mean? Don't know.

No. The statement was that the GUI was somehow an accessory. The GUI is only a part of the application. (Slightly OT: It's common practice, or at least it is commonly considered a good practice to separate the code for the user interface and the application logic.)

Every TCP connection?! Pleeeese, get real.

Yeah. Sad but true. Not the CONTENT of the trasmissions obviously, but data such as ip_dest:ip:source:hour_and_date. For a period of time ranging from 6 to 24 months, depending on the single State. Not only in Italy, that's a UE standard. It's very expensive for our ISPs - and so it is for their customers

You know what? Suppose you right. It's feasible to record sip/dip/date/time. So what? I can't see any point here.

"Hey $ISP, $IP, belonging to you,  downloaded/made available $INFRINGING_MATERIAL, $DATE, $TIME. Give us the poor saps name and adress so we can sue!"
...Which, ofc, is only one example. This one is only legally viable in jurisdictions with legislation similar to the DMCA though.

[lfroen]

You know what? Suppose you right. It's feasible to record sip/dip/date/time. So what? I can't see any point here.

"Hey $ISP, $IP, belonging to you,  downloaded/made available $INFRINGING_MATERIAL, $DATE, $TIME. Give us the poor saps name and adress so we can sue!"
...Which, ofc, is only one example. This one is only legally viable in jurisdictions with legislation similar to the DMCA though.
[/quote]

Still no point. I personally don't give a flying fuck whatever information my ISP want to record. If I want to hide something - there's lot of encryption applications available.
You don't like DMCA-style laws? Feel free to participate in political campaign, given that you live in democratic party.

[Starchild]

Still, not everyone actually uses encryption, even though they should;)
I was under the impression that encrytion in p2p apps was mainly to get around packet shaping?

Besides, if you want to know who is up/downloading/"making available" something, you generally don't  have to do much more than connect to the network in question.
I'm sure you know that the RIAA have prosecuted people in the US, and I'm pretty certain they've couldn't have gotten the names of the victims unless the ISP:s kept track of who had what ip when.

And it might help catching ignorant script-kiddies. If they didn't keep track, I'm not certain they'd even have a chance.

...And if you don't live in a democratic country, you probably have bigger problems;)

[lfroen]

Back to the topic: I was unable to re-create bug on my machine. If you still experiencing the issue, please enable core dumps, and create backtrace with gdb using "bt" command. If you don't know what I'm talking about - feel free to ask and/or read wiki page.

I was under the impression that encryption in p2p apps was mainly to get around packet shaping?

You're correct. p2p doesn't feature encryption, it features obfuscation. I believe, however, that the best way to go around traffic shaping is to switch provider.

Regarding your rand about "ISP keeps track on ...": it's completely off-topic and have nothing to do neither with KAD nor with p2p in general. Your phone company track all number you call or calling you - does it prevent you from saying something on the phone?!

...And if you don't live in a democratic country, you probably have bigger problems;)

That's blatantly incorrect. Whether I have problems or not, depend on my social status which, I believe, you have no idea about.

[RRT4]

amuleweb cvs 31.10.2007 compiled with
./configure --prefix=/usr --enable-debug --disable-optimize --enable-amule-daemon --enable-amulecmd --enable-webserver --enable-geoip
wxGTK 2.8.6 compiled with:
./configure --prefix=/usr --with-gtk --enable-unicode --disable-compat24 --enable-optimise

i have no experience so i don't know if I am giving the correct information, anyway:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb62feb90 (LWP 5471)]
0x0805429f in ?? ()
(gdb) bt
#0  0x0805429f in ?? ()
#1  0x00000006 in ?? ()
#2  0x00000000 in ?? ()
(gdb) bt full
#0  0x0805429f in ?? ()
No symbol table info available.
#1  0x00000006 in ?? ()
No symbol table info available.
#2  0x00000000 in ?? ()
No symbol table info available.
(gdb) thread apply all bt

Thread 48 (Thread 0xb62feb90 (LWP 5471)):
#0  0x0805429f in ?? ()
#1  0x00000006 in ?? ()
#2  0x00000000 in ?? ()

Thread 2 (Thread 0xb74d8b90 (LWP 5425)):
#0  0xb7f43410 in __kernel_vsyscall ()
#1  0xb7b64521 in select () from /lib/libc.so.6
#2  0xb7ebac81 in GSocket::Select () from /usr/lib/libwx_baseu_net-2.8.so.0
#3  0xb7eb5975 in wxSocketBase::_Wait () from /usr/lib/libwx_baseu_net-2.8.so.0
#4  0xb7eb5c53 in wxSocketServer::WaitForAccept () from /usr/lib/libwx_baseu_net-2.8.so.0
#5  0x08082353 in ?? ()
#6  0xb7e2d6e8 in wxThreadInternal::PthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#7  0xb7e2d75d in wxPthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#8  0xb7eee162 in start_thread () from /lib/libpthread.so.0
#9  0xb7b6afee in clone () from /lib/libc.so.6

Thread 1 (Thread 0xb7a60aa0 (LWP 5422)):
#0  0xb7f43410 in __kernel_vsyscall ()
#1  0xb7ef4f2b in read () from /lib/libpthread.so.0
#2  0xb7d2bc73 in rl_getc () from /lib/libreadline.so.5
#3  0xb7d2c0e0 in rl_read_key () from /lib/libreadline.so.5
#4  0xb7d18e8e in readline_internal_char () from /lib/libreadline.so.5
#5  0xb7d19327 in readline () from /lib/libreadline.so.5
#6  0x080881ac in ?? ()
#7  0x08089ea1 in ?? ()
#8  0x08088b03 in ?? ()
#9  0x08080328 in ?? ()
#10 0xb7dc3e2a in wxEntry () from /usr/lib/libwx_baseu-2.8.so.0
#11 0xb7dc3ed7 in wxEntry () from /usr/lib/libwx_baseu-2.8.so.0
#12 0x080803b0 in ?? ()
#13 0xb7abef90 in __libc_start_main () from /lib/libc.so.6
#14 0x08053f31 in ?? ()
#0  0x0805429f in ?? ()
(gdb) 

[lfroen]

Debug information from amuleweb is missing despite --enable-debug. Maybe you using wrong binary?

[RRT4]

amuleweb CVS using wxGTK2 v2.8.6 (Snapshot: Thu Nov  1 07:01:55 CET 2007)

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb0bf9b90 (LWP 28737)]
0x08054491 in CNumImageMask::Apply (this=0xb7579008, image=0x81b1858, offx=6, offy=2) at WebServer.cpp:1549
1549                    png_bytep num_row = m_row_mask_ptrs;
(gdb) bt
#0  0x08054491 in CNumImageMask::Apply (this=0xb7579008, image=0x81b1858, offx=6, offy=2) at WebServer.cpp:1549
#1  0x08057965 in CDynStatisticImage::DrawImage (this=0x81b17d8) at WebServer.cpp:1448
#2  0x0805a451 in CDynStatisticImage::RequestData (this=0x81b17d8, size=@0xb0bf90a8) at WebServer.cpp:1499
#3  0x080603cf in CScriptWebServer::ProcessImgFileReq (this=0x81a8598, Data=@0xb0bf91d4) at WebServer.cpp:292
#4  0x08089b26 in CWebSocket::OnRequestReceived (this=0xb3c237a8,
    pHeader=0xb3c23991 "HTTP/1.1\r\nHost: 192.168.0.10:4711\r\nUser-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1. Gecko/20071020 BonEcho/2.0.0.8\r\nAccept: image/png,*/*;q=0.5\r\nAccept-Language: en-us,en;q=0.5\r\nAccept-"..., pData=0x0, dwDataLen=0) at WebSocket.cpp:373
#5  0x0808a380 in CWCThread::Entry (this=0xb3c23798) at WebSocket.cpp:240
#6  0xb7ecd6e8 in wxThreadInternal::PthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#7  0xb7ecd75d in wxPthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#8  0xb7f8e162 in start_thread () from /lib/libpthread.so.0
#9  0xb7c0afee in clone () from /lib/libc.so.6
(gdb) bt full
#0  0x08054491 in CNumImageMask::Apply (this=0xb7579008, image=0x81b1858, offx=6, offy=2) at WebServer.cpp:1549
        img_row = (
    png_bytep) 0xb75c3bc0 "\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001@\001\001"...
        num_row = (png_bytep) 0x81b20e8 ""
        i = 0
#1  0x08057965 in CDynStatisticImage::DrawImage (this=0x81b17d8) at WebServer.cpp:1448
        maxval = 1039
        m_scale_up = 1
        m_scale_down = 6
        img_delta = 2
        y_axis_max = 1104
        prev_data = 134657959
        graph_color = 255
#2  0x0805a451 in CDynStatisticImage::RequestData (this=0x81b17d8, size=@0xb0bf90a8) at WebServer.cpp:1499
No locals.
#3  0x080603cf in CScriptWebServer::ProcessImgFileReq (this=0x81a8598, Data=@0xb0bf91d4) at WebServer.cpp:292
        img_size = 0
        img_data = (unsigned char *) 0xb0bf9200 "dÄ)\bgE\213kš7³\230Oö·\230Oö·h À³H\222¿°\224ªõ·h À³"
        lock = {m_isOk = true, m_mutex = @0x81a85a0}
        session = (const CSession *) 0x825c64c
        imgName = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0x829f4e4}, <No data fields>}
        img = (CAnyImage *) 0x81b1834
#4  0x08089b26 in CWebSocket::OnRequestReceived (this=0xb3c237a8,
    pHeader=0xb3c23991 "HTTP/1.1\r\nHost: 192.168.0.10:4711\r\nUser-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1. Gecko/20071020 BonEcho/2.0.0.8\r\nAccept: image/png,*/*;q=0.5\r\nAccept-Language: en-us,en;q=0.5\r\nAccept-"..., pData=0x0, dwDataLen=0) at WebSocket.cpp:373
        url_ext = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0x829929c}, <No data fields>}
        is_post = false
        path = 0xb3c2397c "/amule_stats_kad.png"
        sURL = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0x829c464}, <No data fields>}
        sessid = 1804289383
        current_cookie = 0xb3c23b0b "amuleweb_session_id=1804289383\r\n\r\n"
        Data = {parsedURL = {m_path = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0xb7ef0cb8}, <No data fields>},
    m_file = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0x829cfa4}, <No data fields>},
    m_params = {<std::__norm::map<wxString,wxString,std::less<wxString>,std::allocator<std::pair<const wxString, wxString> > >> = {_M_t = {
---Type <return> to continue, or q <return> to quit---
          _M_impl = {<std::allocator<std::_Rb_tree_node<std::pair<const wxString, wxString> > >> = {<__gnu_cxx::new_allocator<std::_Rb_tree_node<std::pair<const wxString, wxString> > >> = {<No data fields>}, <No data fields>},
            _M_key_compare = {<std::binary_function<wxString,wxString,bool>> = {<No data fields>}, <No data fields>}, _M_header = {
              _M_color = std::_S_red, _M_parent = 0x0, _M_left = 0xb0bf9184, _M_right = 0xb0bf9184},
            _M_node_count = 0}}}, <__gnu_debug::_Safe_sequence<std::__debug::map<wxString, wxString, std::less<wxString>, std::allocator<std::pair<const wxString, wxString> > > >> = {<__gnu_debug::_Safe_sequence_base> = {_M_iterators = 0x0, _M_const_iterators = 0x0,
          _M_version = 1}, <No data fields>}, <No data fields>}}, sURL = {<wxStringBase> = {static npos = 4294967295,
      m_pchData = 0x829c464}, <No data fields>}, SessionID = 1804289383, pSocket = 0xb3c237a8}
        sFile = {<wxStringBase> = {static npos = 4294967295, m_pchData = 0x829cfa4}, <No data fields>}
#5  0x0808a380 in CWCThread::Entry (this=0xb3c23798) at WebSocket.cpp:240
        IsGet = true
        IsPost = false
        lock = {m_isOk = false, m_mutex = @0x0}
#6  0xb7ecd6e8 in wxThreadInternal::PthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
No symbol table info available.
#7  0xb7ecd75d in wxPthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
No symbol table info available.
#8  0xb7f8e162 in start_thread () from /lib/libpthread.so.0
No symbol table info available.
#9  0xb7c0afee in clone () from /lib/libc.so.6
No symbol table info available.
(gdb) thread apply all bt

Thread 47 (Thread 0xb0bf9b90 (LWP 28737)):
#0  0x08054491 in CNumImageMask::Apply (this=0xb7579008, image=0x81b1858, offx=6, offy=2) at WebServer.cpp:1549
#1  0x08057965 in CDynStatisticImage::DrawImage (this=0x81b17d8) at WebServer.cpp:1448
#2  0x0805a451 in CDynStatisticImage::RequestData (this=0x81b17d8, size=@0xb0bf90a8) at WebServer.cpp:1499
#3  0x080603cf in CScriptWebServer::ProcessImgFileReq (this=0x81a8598, Data=@0xb0bf91d4) at WebServer.cpp:292
#4  0x08089b26 in CWebSocket::OnRequestReceived (this=0xb3c237a8,
    pHeader=0xb3c23991 "HTTP/1.1\r\nHost: 192.168.0.10:4711\r\nUser-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1. Gecko/20071020 BonEcho/2.0.0.8\r\nAccept: image/png,*/*;q=0.5\r\nAccept-Language: en-us,en;q=0.5\r\nAccept-"..., pData=0x0, dwDataLen=0) at WebSocket.cpp:373
#5  0x0808a380 in CWCThread::Entry (this=0xb3c23798) at WebSocket.cpp:240
#6  0xb7ecd6e8 in wxThreadInternal::PthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#7  0xb7ecd75d in wxPthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#8  0xb7f8e162 in start_thread () from /lib/libpthread.so.0
#9  0xb7c0afee in clone () from /lib/libc.so.6

Thread 2 (Thread 0xb7578b90 (LWP 28692)):
#0  0xb7fbb410 in __kernel_vsyscall ()
#1  0xb7c04521 in select () from /lib/libc.so.6
#2  0xb7f5ac81 in GSocket::Select () from /usr/lib/libwx_baseu_net-2.8.so.0
#3  0xb7f55975 in wxSocketBase::_Wait () from /usr/lib/libwx_baseu_net-2.8.so.0
#4  0xb7f55c53 in wxSocketServer::WaitForAccept () from /usr/lib/libwx_baseu_net-2.8.so.0
#5  0x0808b251 in CWSThread::Entry (this=0x81b3d28) at WebSocket.cpp:97
#6  0xb7ecd6e8 in wxThreadInternal::PthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#7  0xb7ecd75d in wxPthreadStart () from /usr/lib/libwx_baseu-2.8.so.0
#8  0xb7f8e162 in start_thread () from /lib/libpthread.so.0
#9  0xb7c0afee in clone () from /lib/libc.so.6

Thread 1 (Thread 0xb7b00970 (LWP 28689)):
#0  0xb7fbb410 in __kernel_vsyscall ()
#1  0xb7f94f2b in read () from /lib/libpthread.so.0
#2  0xb7dcbc73 in rl_getc () from /lib/libreadline.so.5
#3  0xb7dcc0e0 in rl_read_key () from /lib/libreadline.so.5
#4  0xb7db8e8e in readline_internal_char () from /lib/libreadline.so.5
#5  0xb7db9327 in readline () from /lib/libreadline.so.5
#6  0x080923e6 in CaMuleExternalConnector::GetCommand (this=0x81851f0, prompt=@0xbfeb5870, buffer=0xbfeb5688 "H<\033\b ", buffer_size=256)
    at ../../../src/ExternalConnector.cpp:304
#7  0x0809496f in CaMuleExternalConnector::TextShell (this=0x81851f0, prompt=@0xbfeb5870) at ../../../src/ExternalConnector.cpp:339
#8  0x08092fe3 in CaMuleExternalConnector::ConnectAndRun (this=0x81851f0, ProgName=@0xbfeb5870, ProgVersion=@0xbfeb586c)
---Type <return> to continue, or q <return> to quit---
    at ../../../src/ExternalConnector.cpp:423
#9  0x0808866c in CamulewebApp::OnRun (this=0x81851f0) at WebInterface.cpp:62
#10 0xb7e63e2a in wxEntry () from /usr/lib/libwx_baseu-2.8.so.0
#11 0xb7e63ed7 in wxEntry () from /usr/lib/libwx_baseu-2.8.so.0
#12 0x08088700 in main (argc=-579090249, argv=0xb7) at WebInterface.cpp:48
0x08054491      1549                    png_bytep num_row = m_row_mask_ptrs;
(gdb)

[lfroen]

Now we're talking. I seems to see the problem.
Thank you for report

[gaglia]

I'd like to point out however that the amuleweb segfault error is somewhat "less critical" than the amule-daemon KAD disconnections. With the bug in amuleweb, you can still "use" amuleweb, all you have to do is avoid clicking on the KAD tab panel (not essential for tipical usage).

I'm trying the 1_10_07 CVS version from a couple of days, it seems that amule-daemon stays connected to KAD for a longer period before disconnecting, but it still hangs after tipically less than 20 hours (so i cannot really say if this is due to code improvement or because of other factors). Please note that all these considerations apply without even executing amuleweb: i'm tracking amule-daemon status via amulecmd, so i don't think this issue is somewhat bound to the amuleweb bug.

Now i'm trying to see if the connection time can be improved by lowering max number of connections. I'm going to post the results.

Thx everyone for the attention