gaoesa

Oh. That changes it all. Yes, it is statically linked then. For some reason I assumed a Linux server. EDIT: But still the issue itself is the same, regardless how the standard library is included.

The standard library (also know as the C POSIX library) is not statically linked. It is provided by the OS. This is also why there are compatibility issues with different Linux versions and some server providers would not be able to run the mod if we would link against too new version of libc. I don't think there is any issue with that library itself, it would manifest all around if there were. But if the OS for some reason would not want the file to be written, the calls to the standard library would then fail. Thanks. I will keep that in mind. I wouldn't mind deleting the whole profile. I have been suspecting it is something to do with flash. At least I finally cleaned up the left overs of the duckduckgo, the damn thing requires javascript, or it doesn't work at all, and I'm not going to enable it for a search page that I don't think needs it for anything. That was another surprisingly hard task as it hides itself well.

Already lost twice what I was writing because mozilla keeps crashing on me. Anyway again: The file should always have a database header. This is static data that is written and flushed right after the file has been opened. (Consecutive lines) I'm looking at this and I don't see how the file could be empty because of programmatic error. EDIT: Now that I have something out, bit more chat. The program in this case does not output possible writing errors to the log. By writing errors I mean if the underlying standard library is refusing to co-operate for any reason. This thing probably needs fixing now. Possible reasons would include another process locking the file. EDIT: Found out there was an update to 33 version of Firefox. Less surprisingly, it managed to crash instantly after the update. I wonder how they can manage making it less and less stable on every update.

The X in the name refers to extra data. It contains needed data for admin level protection, personal greetings and similar things. There is an extremely small window of chance the server could be killed right between the file has been truncated and before it gets written. It is that way a special file that it is rewritten on every shutdown. To find the moment, you would need to go to the last shutdown before the first initialization where it said it is corrupted. Looking if the shutdown does not complete. If the shutdown completes, the issue is external. EDIT: By killed I mean a signal from console or similar event and that the engine would not wait for the shutdown it has previously issued, to complete.

It shouldn't be empty. Did you lose data or is this fresh install?

Too many airstrikes requested? EDIT: Did you mean airstrike can exploding and not starting or artillery starting the shelling instantly without smoke?

Actually I didn't find these cvars from ETPub. However, added cg_numPopups, cg_popupStayTime and cd_popupFadeTime. I don't know why the popups should be delayed from popping up, so I left cg_popupTime out.

Ok. So you indeed meant the client crashed. When the memory runs out, the ui and cgame can crash because there is practically no handing for that in the dozens of cases where allocation is attempted. This is a an issue starting all the way from the etmain. We will fix that in the next version of silEnT.

I doubt that. The ui library can run out of memory if you add too much menus. And it should have some spare room because the memory is allocated during map too.

I have just learned that with ET Legacy, on top of the crash bug they have, it also generates server timed out kicks that are automatic and instant when connecting. I would like to know if there is a higher frequency for the timeout issue with players using ET Legacy? I still haven't been able to reproduce anything of the sort.

I got some additional information from clan DIABOLIK and remembered that this is the kind of issue that came from a failed attempt to fix the "warmup kick" in the engine. I know that at least in some point it (the bad fix) was added in to the ET 3.00. I wonder if that is the reason for some of the cases here. I take blame since I posted the non working thing to these forums. I can't find it anymore.

I would really like to be able to reproduce the issue on myself. This could be the DDOS protection in server engine or server firewall, and for some cases it probably is, but with this frequency, it could also be some sort of client side problem. Combined with the other thread. Any information on how to reproduce would be appreciated. Client engine versions etc.

Oops, my responses in that thread make no sense actually in relation to the presented problems. So it is my fault. I don't know how that has happened.

This is a known issue and indeed a bug.

@clan DIABOLIK I have never heard of a bug that causes it. Though I am not disputing that this has been happening and is a valid reason for this thread. Also, like I already explained, it couldn't be a mod bug if such bug exists.

You took a screenshot. So when you have that dialog, open console and scroll upwards.

@clan DIABOLIC That is incorrect, Getting stuck in Awaiting gamestate is not silEnT mod related. Your mixing it with the getting disconnected for unknwn reason kick. Which can happen to new connections at the end of the warmup only, and which is a known bug. But completely different and easily resolved by the user by clicking reconnect. Do note, the mod ony kicks people, the engine handles all connection related things.

Not from the server. From the lient console after you have dropped from server. The string to find is "UI_Alloc: Failure. Out of memory!".

I have recently found out that too much menus can cause the UI library to run out of memory. Either because of too much menus to load or too much strings to handle. Please scroll console upwards to see if there is UI_Alloc out of memory line there. The emptyness could be caused by not having crash handler for the UI library and this specific error is not very well handled by the code. It's still not right to show missing pacs in such a dialog.

It is the directory named "silent" under Wolfenstein. Under the path where the ET executable is or if you're on Linux, it could be in a hidden directory under you home. However, if you don't have it, then there has not been a crash.

I think the menus are a bit messed. That is not real crash information, even though it says so. But you can check the silent_crash.txt from silent directory to be sure.

I already answered the question in your topic. Add the wanted flags together into a single value. I didn't know if she had found the wiki setting yet because the original question did not reference it. Seemed the most likely reason for confusion.

http://mygamingtalk.com/wiki/index.php/Silent_Mod_Server_Cvar#g_dmg

You add all the flags you want together into one value.

g_realHead does not exist anymore. ETPro head boxes are the only option. There is also g_dmg for using advanced damage settings. I.e. less damage from arm and leg hits.