10 replies to this topic

[Deathlike]

There are times that when I close the patcher after patching, Windows (XP) tells me that that the patcher application has crashed.

I honestly wonder if the patcher is causing some of the problems... since I can't see any good that comes from a crashing patcher.

This is strictly with closing the patcher after having patched or installed the game (I just want the game patched, but not run it immediately)... I use the "X" button at the upper right hand corner to close the app (I'm not sure if there's any other possible way, outside of task manager).

I'm not even sure why the patcher crashes itself.

The patcher doesn't do it frequently.. but occasionally it crashes if I'm not patching either...

Edited by Deathlike, 05 April 2013 - 09:44 PM.

[Krist Smith]

Quick answer: an idle thread was being released incorrectly in some situations. But, there's no side-effect, since, like I say, the thread was idle (this bug is actually more likely to occur if that thread never had anything to do in the first place).

The next patcher update will have a fix for this particular bad bit of behaviour.

[Deathlike]

Hi Krist,

Have you fixed the hanging patcher bug? The one were you try to patch the game, and the patch tells you out of the blue "you're trying to close the patcher" and when I look at the logs, it's waiting on some file to patch even though it should be a non-issue...

[Krist Smith]

Well, it depends on what's causing the hang. (Sorry for stating the obvious )

In some situations, there's a genuine hang while attempting to download (sometimes resolves itself, sometimes the downloads time out). This is related to both the CDN and your ISP (or other local net configuration issues, like firewalls and proxies).

There was also some improper error handling (particularly when dealing with running out of disk space) that made the patcher either crash or become unresponsive. These cases have also been fixed, with the patcher doing as much looking ahead for space as possible, as well as a more graceful handling of running out of disk space during all operations.

The last case is when it looks like a stall but it's actually doing work. I'm aware of some cases (such as "Finalizing Patch") where the patcher is doing a rather large operation that can't be broken down for smaller progress updates (for example, while it's patching or copying Objects.pak, the largest file in the install base). These operations can take anywhere from 30 seconds to 3 or 4 minutes (depending on whether or not the computer is being used for other things at the time, as well as not having enough contiguous space for the patching process, so be sure to defrag those non-SSD drives and keep plenty of overhead - general usage suggests to keep at least 10% of any hard drive free at all times for best OS and program performance).

If you want a more specific explanation, I'd probably have to see the log line in question, as well as confirmation that you had plenty of space free on the drive that contains your game and also the one that contains the temporary working folder (if they're on different drives).

[Deathlike]

I don't think it's a space issue, but it simply "hangs" in the sense that it thinks i want to "quit the patching process", so it stays stuck there for no particular reason. It gets stuck @ like some DX11 dll even though I have plenty of space available on the system.

I'll see if this is properly addressed in the next patch with the patcher.

I will get logs for you if this is still the case.

There has been cases of this before... and I cannot get backlogs far enough in the forums to find it.

There were mega-threads on how people couldn't straight upgrade months ago, and it still causes problems today. Not everyone was affected, but it still happens.

Sample steps:
1) It's patch day.
2) Load up patcher/launcher.
3) Click on "patch".
4) Download patch.
5) See patcher trying to finalize.
6) Patcher "prompts" you that you shouldn't quit this process. So I click on whatever the response is to continue the patching.
7) Patcher "hangs" there indefinitely.

Patcher logs say "it can't access the resource to file CryRenderD3D11.dll" or something along those lines.

That's the problem with the current patcher.

Edited by Deathlike, 10 April 2013 - 03:02 PM.

[Krist Smith]

Hmmm... Ok, the only situation I've seen where the patcher tries to close itself is when there's an exception thrown (this was being mistakenly handled as "User closing patcher" due to the code path). From reports I've seen, the most frequent cause of exceptions is running out of disk space, so I'm not sure what's causing it for you. Bear in mind that even if you've installed the game to a different drive, the default temp folder is created on the C: (or default Windows) drive. If you've already taken this into account (or you installed the game to the default drive), then I'd need to look through some of the logs to get more info.

[Deathlike]

Found the thread in question:
https://mwomercs.com...patching-stage/

This happened enough times for me.. so much I just did the frontend_version.dll deletion straight up. Eventually that didn't work out, so I always end up doing a full reinstall (uninstall game, the install).

I installed the game on the main drive (C) which is "1TB" in capacity.

Edited by Deathlike, 10 April 2013 - 03:10 PM.

[Krist Smith]

Ok, that's more weirdness. Deleting the frontend_version.dll file is essentially the same as reinstalling, it just doesn't wipe out the old files. If the patcher can't determine the version (which comes from that file), then it assumes you're at version 0 and applies the "zero-to-now" patch that downloads that 1.9+ GB patch file and does the complete overwrite of any files you already have. This is much more work than you'd need, though. If you're only having troubles with some files in the Bin32 folder, there's no need to sit there and wait for like 2 GB to download. The Bin32 folder in its entirety is just over 31 MB, so that's a whole lot of downloading overkill.

Maybe we should move this into PMs so I can provide more direct support (and do fewer distracting "try it now" tactics that would mislead anyone else reading this).

[Deathlike]

Well, the weirdness is more about the patcher reacting poorly... the logs in the thread are "accurate" to my experience with the current patch system.

We can move it into PMs if you wish. I just want these poor patcher experiences to go away in the future.

[Krist Smith]

Deathlike, on 10 April 2013 - 03:54 PM, said:

I just want these poor patcher experiences to go away in the future.

As do I, my friend. It would be a dream come true for me (since this is my primary task on this project ).

What we can do is play the waiting game for now. I'm pretty sure that the patcher update will be out next week (unless we find a major bug in my latest iteration that I can't fix in time). Once it's out, see if you still have the same problems. If you do, let me know right away - I'll be waiting. We can then run through the logs and try out some alternative methods to get everything all fixed up and, if needed, fix up the patcher itself.

How does that sound? If you're more interested in squashing all your problems ASAP, then we can probably run through some tests earlier. It's totally up to you, I'll work within your time allotment.

[Deathlike]

Sure, we do all that on patch doooomsday.