ID:1868923
 
Resolved
Projects compiled in previous versions didn't communicate with clients properly.
BYOND Version:508
Operating System:Windows 7 Pro 64-bit
Web Browser:Chrome 43.0.2357.65
Applies to:Dream Seeker
Status: Resolved (508.1288)

This issue has been resolved.
Descriptive Problem Summary: After I compiled and hosted my project via DD in version 508, I tried to connect to it locally. Every time, the client crashes, or runs extremely slow to the point that it crashes when I try to press anything. The server itself runs fine until I attempt to log in.

Here is the information I get from Windows:
Problem signature:
Problem Event Name: APPCRASH
Application Name: dreamseeker.exe
Application Version: 5.0.508.1287
Application Timestamp: 55762121
Fault Module Name: byondcore.dll
Fault Module Version: 5.0.508.1287
Fault Module Timestamp: 55762044
Exception Code: c0000005
Exception Offset: 000313b0
OS Version: 6.1.7601.2.1.0.256.48
Locale ID: 1033
Additional Information 1: 0a9e
Additional Information 2: 0a9e372d3b4ad19135b953a78882e789
Additional Information 3: 0a9e
Additional Information 4: 0a9e372d3b4ad19135b953a78882e789


Ask away if more information or a source is needed. I don't have any code to post though since it doesn't seem to be tied to any code.

Expected Results: Connecting to the game without my client freezing.

Actual Results: My client freezes and crashes any time I try to log in.

Does the problem occur:
Every time? Or how often? Every time.
In another game? Happens in two projects I tested, both using completely different source code. I cannot log in to them without my client freezing.

When does the problem NOT occur? When I use version 507.1285.

Did the problem NOT occur in any earlier versions? If so, what was the last version that worked? Version 507.1285.

Workarounds: None apparently; I can't seem to locate the cause.
Do you have a project I can test this with? I haven't seen this myself.
You know, I think I found this. If you recompile in 508, does the problem disappear? That would be consistent with the bug I found.
Lummox JR resolved issue with message:
Projects compiled in previous versions didn't communicate with clients properly.
The problem does not fix when I recompile in 508. In fact, it doesn't matter what version I use now, the problem occurs. Compiling a project in 508 seems to break it.

The project involved doesn't matter from what I've tested. Create a new environment, compile it in 508, and host a local DD session. Attempt to join the server and it should freeze if reproduceable (the server stays up sometimes while other times it freezes; the client freezes every time).

Edit: Here's a different error message when trying to load a project broken by 508, in 507.

Description:
A problem caused this program to stop interacting with Windows.

Problem signature:
Problem Event Name: AppHangB1
Application Name: dreamseeker.exe
Application Version: 5.0.507.1286
Application Timestamp: 55694a6e
Hang Signature: 16ef
Hang Type: 0
OS Version: 6.1.7601.2.1.0.256.48
Locale ID: 1033
Additional Hang Signature 1: 16ef9fc0d17a366518f893502181b95a
Additional Hang Signature 2: 05f1
Additional Hang Signature 3: 05f15f30684004289f78b682188c440e
Additional Hang Signature 4: 16ef
Additional Hang Signature 5: 16ef9fc0d17a366518f893502181b95a
Additional Hang Signature 6: 05f1
Additional Hang Signature 7: 05f15f30684004289f78b682188c440e
If you an send me the project you compiled in 508 that had an issue in 507, I can take a look.
In response to Lummox JR
Have an email I could send it to by chance?

It also only seems to be this project acting like this. The other two projects I used, both a lot smaller, work fine in 507.
You can email me at [email protected], or if you have a link to the file somewhere you can just send me a private message.
aol

2015

I thought you were cool man.
I thought you were cool man.

Old school cool. I heard Lummox still programs BYOND on a Trash 80. I also heard he uses a trackball.
In response to Ter13
Ter13 wrote:
I thought you were cool man.

Old school cool. I heard Lummox still programs BYOND on a Trash 80. I also heard he uses a trackball.

I use a trackball.
I haven't received any file yet. Please get that to me as soon as you can. I can't reproduce this bug but I suspect it's serious.

[edit]
Also, it'd probably be good to post a brand new bug report for it. The crash I fixed earlier in this report I was able to identify as only happening with worlds compiled before 508, when the server and client were both 508.
In response to Lummox JR
I sent the source in full, I wasn't sure if you only wanted the host files. It worked fine in 507 until I compiled it and tried to run it in 508. At first, I could go back to 507 and it would fix itself, but after a couple times, it broke completely.

I'll get another report made when I get a chance, going to be out for a bit. Hope you can find whatever this is.
I can confirm that it eventually crashes for me once the interface comes up. Looking into it now. Thanks for sending the source.

[edit]
Argh, no, it's not crashing. It's that guest check screwing me over.

The startup behavior of this game is really, really terrible. There's a New() proc on turfs that's causing slow startup, and mob/human/Move() is being called way too much before it seems like it's even needed.
Are you certain the problem isn't something in your source being outright broken?

What I've found is that mob/human/Move() is failing and returning null on Combat.dm:89 in the source you provided. It's being called by the default, built-in Login() which is attempting to move my mob into an area. Because your Move() returns null (false), the default Login() tries the next turf, and the next, and the next, for the whole area which presumably is the whole map or close to it. Since the Move() code also contains some really heinous loops and logic that needs cleaning up, it's also happening fairly slowly. Eventually, the interface does come up (Windows may simply be interpreting the extremely slow load as a hang).

This is looking like you simply made a recent change in your source and hadn't recompiled (therefore hadn't noticed the issue) yet. I tested this in a 507-compiled version, running in 507, and got identical results.
Ah yeah, good catch. I still get the crashing client though.

I'm sure it's not the game (I'm well aware of the lack of TLC, that's being rectified). The crashes happen on multiple projects, not just this one. I had no issues until 508 came into play.
Yes, if a 508 server and client handle a game compiled from before 507, there is an issue; that's the crash I addressed above.
I've had the hanging FKI is describing in a completely blank project. No code whatsoever, no map, no interface, nothing.
You're likely seeing the aforementioned issue.
This is a fresh project made in 508, compiled in 508, hosted in 508, and connected to on 508. Client eats all available memory, hangs, and eventually crashes.
Page: 1 2