Description of problem: As discussed, it is prolly best to submit a bug to gain wider exposure. This symptom has been noticed a while ago and confirmed during QA Event, where a number of users has noticed this bug on numerous occassion. Version-Release number of selected component (if applicable): before scim-1.4.4-34.fc6 scim-bridge-0.4.2-1.fc6 How reproducible: Not 100% Steps to Reproduce: 1.normal usage of IM over a period of time on Gtk Apps 2. 3. Actual results: Connection between scim and apps gets cut off, as a result, no IME available. Select IME from scim icon on systray will re-establish the connection again. Expected results: Connection should be persistant Additional info: The frequency is much higher on vnc, Jens had seen it once. ;-)
That time, is the agent down? If it's down, what error is shown in dsmg? Please move into a virtual terminal and check it.
Thanks for the tips Ryo-san. I will watch out for these error message.
Btw, you can get the corefile if the agent is crashed. It will help us to solve this problem. :) 1. Make a shell script "scim-bridge" as follows: #!/usr/bin/bash ulimic -c unlimited exec /usr/bin/scim-bridge 2. Install the script into /usr/local/bin/ 3. Patch "scim-bridge-agent.cpp" as follows and make install it: retval_t ScimBridgeAgentImpl::launch () { scim_bridge_pdebugln (5, "launch ()"); ... } else { // This is the child process. - if (chdir ("/")) { + if (chdir ("/tmp")) { scim_bridge_perrorln ("Cannot change the working directory: %s", strerror (errno)); return RETVAL_FAILED; } close (0); close (1); close (2); } scim_bridge_pdebugln (5, "Daemonize done"); ... } 4. Logout once, and relogin. 5. Wait until the agent crashes.
Is this any better with scim-bridge-0.4.5?
Closing for now. Please reopen if you still see this.