rdmenezes / synecdoche

Automatically exported from code.google.com/p/synecdoche
GNU Lesser General Public License v3.0
0 stars 0 forks source link

Manager crashes when loosing the connection to the core client while project tab is selected. #45

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. Connect the manager to a client and select the projects tab.
2. Shut down the client (directly, not via the the manager).
3. The manager will receive a SIGPIPE when trying to do send the get_state 
RPC.

Here is the stacktrace:
Program received signal SIGPIPE, Broken pipe.
[Switching to Thread 0xb65ac9c0 (LWP 22606)]
0xb7f67424 in __kernel_vsyscall ()
(gdb) bt
#0  0xb7f67424 in __kernel_vsyscall ()
#1  0xb7322ac1 in send () from /lib/libpthread.so.0
#2  0x0815aeb9 in RPC_CLIENT::send_request (this=0x9847d1c, 
    p=0x8301ad7 "<get_state/>\n") at gui_rpc_client.C:277
#3  0x0815af7b in RPC::do_rpc (this=0xbfa80b14, req=0x8301ad7 "<get_state/>
\n")
    at gui_rpc_client.C:327
#4  0x0816d752 in RPC_CLIENT::get_state (this=0x9847d1c, 
state=@0x9847d44)
    at gui_rpc_client_ops.C:1059
#5  0x080d6902 in CMainDocument::CachedStateUpdate (this=0x9847d00)
    at MainDocument.cpp:436
#6  0x080d6a11 in CMainDocument::ForceCacheUpdate (this=0x9847d00)
    at MainDocument.cpp:613
#7  0x080d6e00 in CMainDocument::CachedProjectStatusUpdate (this=0x9847d00)
    at MainDocument.cpp:694
#8  0x080d6e32 in CMainDocument::GetProjectCount (this=0x9847d00)
    at MainDocument.cpp:737
#9  0x08081076 in CBOINCBaseView::OnListRender (this=0x98aa738, 
    event=@0xbfa80d54) at BOINCBaseView.cpp:139
#10 0x0806fc55 in CAdvancedFrame::OnRefreshView (this=0x9846ee8)
    at AdvancedFrame.cpp:1579
#11 0xb76a9173 in wxAppConsole::HandleEvent ()
   from /usr/lib/libwx_baseu-2.8.so.0

This happens on Linux, Windows seems to be immune to this crash.

Original issue reported on code.google.com by DerMeister667@googlemail.com on 4 Apr 2009 at 10:58

GoogleCodeExporter commented 9 years ago
Writing to a closed socket produces SIGPIPE.

Original comment by nicolas....@gmail.com on 4 Apr 2009 at 6:11

GoogleCodeExporter commented 9 years ago
I can't reproduce this any more. The manager handles SIGPIPE and even prints a 
message when getting this signal. I was able to reproduce the situation when 
this 
signal is emitted but the manager just printed the message and didn't crash.

Anyone else able to reproduce this crash?

Original comment by DerMeister667@googlemail.com on 8 Apr 2009 at 5:10

GoogleCodeExporter commented 9 years ago
I'm reducing the priority and removing the milestone tag unless anyone can 
reproduce 
this crash. I can't any more.

Original comment by DerMeister667@googlemail.com on 23 Apr 2009 at 5:45

GoogleCodeExporter commented 9 years ago
Changing back to "new" (the definition of "accepted" says the problem is 
reproduced).

Original comment by nicolas....@gmail.com on 14 May 2009 at 12:27