“Sabiyhan” System Search Crash Mystery

Danger Investigates

Danger Investigates
Jmanis said:
That's what I was alluding to here. TEST2 is ostensibly a nonexistent system (probably one of several) used for parking factions (in this case, League of Reparation).
View attachment 204917
Searching TEST2 does nothing on the galaxy map, because presumably the system doesn't exist, or there's no corresponding entry in whatever index/lookup gets used to search for these things.

This sort of crashing behaviour usually arises when there is an entry which can be searched, but it has otherwise invalid/missing data that isn't validated before use due to an assumption that all entries are valid and that there shouldn't be invalid entries.

These things happen 🤷‍♀️
Click to expand...
I never knew about the League of Reparation TEST2 system! That info greatly helps this discussion Thankyou :D
 

Jmanis

Jmanis
Filthymick420 said:
I dont understand coding or software programs and how they do their thing, let me just say that in advance.

That said, what if there is some kind of program or subroutine that creates this issue internally to interfere with whatever is done to data mine or any other forms of otherwise hacking/cheating to prevent information regarding this word to cause to be aquire through means other than specifically intended. And the resulting crash is because it's trying to do something it's not yet capable of even though it is but...... and that's where I kinda run out of theory crafting atm.

I just wanted to put this out before I end up forgetting and perhaps some of you might gain inspiration towards an idea that actually makes sense
Click to expand...
Tbh... unlikely, and if so, it'd be pretty amateur.

Crashing is bad. Someone in a previous comment mentioned something about an audio traceback. There's every chance that's a red herring of sorts, and that's just the first thing to fail as a result of the issue, because proper crashes are almost always unexpected.
Since you don't have that background, think of it like an aircraft. If an engine fails or some other system doesn't work, there's usually lots of fallbacks which allow the aircraft ditch in a controlled and safe manner. The flight is still terminated, but it's done in a controlled manner. But an actual crash is catastrophic and the result of all fallbacks failing, and the outcomes usually entirely unexpected.

A similar concept exists for software. If there's a chance something could fail, you want to catch that and handle it in a controlled manner. That's why we've got all these orange sidewinders, scarlet kraits and stuff. They're circumstances where failure is forseen and handled cleanly. A crash to desktop outside the software's organic error handling is generally a catastrophic failure. In the best case you just look a bit unprofessional. In the worst case you could cause potentially irreversible client damage. There's a great case where an EVE Online update could brick people's OS.

With that in mind, you never want to rely on a crash to desktop as your fallback.

So yeah... relying on crashes in predictable circumstances is very bad.
 
Last edited:

[PS4] Filthymick420

[PS4] Filthymick420
It's the audio trace back that makes me think this because of reasons I've been asked to keep to myself and cannot yet discuss. what if it crashes trying to play an audio track in relation to the word but can because it's locked by such and such for (Raxxla) reasons?


Also I forgot to mention this is a full tinfoil theory
 
Last edited:

Kelster87

Kelster87
Hmm, thinking on this further.. If the crash is triggered by something in the audio system, what sound is causing the crash? Can't be the regular map move/zoom sound.

Anyone know where Elite hides it sound files?
 
Last edited:

VR CMDR Novindus

VR CMDR Novindus
Jmanis said:
Tbh... unlikely, and if so, it'd be pretty amateur.

Crashing is bad. Someone in a previous comment mentioned something about an audio traceback. There's every chance that's a red herring of sorts, and that's just the first thing to fail as a result of the issue, because proper crashes are almost always unexpected.
Since you don't have that background, think of it like an aircraft. If an engine fails or some other system doesn't work, there's usually lots of fallbacks which allow the aircraft ditch in a controlled and safe manner. The flight is still terminated, but it's done in a controlled manner. But an actual crash is catastrophic and the result of all fallbacks failing, and the outcomes usually entirely unexpected.

A similar concept exists for software. If there's a chance something could fail, you want to catch that and handle it in a controlled manner. That's why we've got all these orange sidewinders, scarlet kraits and stuff. They're circumstances where failure is forseen and handled cleanly. A crash to desktop outside the software's organic error handling is generally a catastrophic failure. In the best case you just look a bit unprofessional. In the worst case you could cause potentially irreversible client damage. There's a great case where an EVE Online update could brick people's OS.

With that in mind, you never want to rely on a crash to desktop as your fallback.

So yeah... relying on crashes in predictable circumstances is very bad.
Click to expand...

Agreed. Were it a deliberate attempt to track data mining, it would be quietly captured by an event handler, and the user wouldn’t be alerted to it. From the Galaxy Map, the result would have simply been the same as if someone searched for a system name that doesn’t exist.
 

Jmanis

Jmanis
Filthymick420 said:
It's the audio trace back that makes me think this because of reasons I've been asked to keep to myself and cannot yet discuss. what if it crashes trying to play an audio track in relation to the word but can because it's locked by such and such for (Raxxla) reasons?


Also I forgot to mention this is a full tinfoil theory
Click to expand...
Kelster87 said:
Hmm, thinking on this further.. If the crash is triggered by something in the audio system, what sound is causing the crash? Can't be the regular map move/zoom sound.
Click to expand...
That's why i keep suggesting it's probably just a corrupt/ bad index entry.

A common problem with searching exists around existence of data. It's a reasonable assumption that your database indexes are correct... if they aren't, that's a problem that needs to be fixed in your database (corollary: it's a bit of a rabbit hole to code assuming that nothing in the project is correct and functional... it can result in major inefficiencies).

Pseudocoding, the galaxy map lookup probably looks like:
  • searchForFeature(name)
  • if no results, do nothing
  • if results:
1. Get first result
2. Play audio for result (as the audio starts before anything happens)[1]
3. Move to result location
4. Etc

... but if we're dealing with a corrupt index, the result probably points to null data. 2. Is then trying to play audio on a feature which doesn't exist which would crash.

It'd be totally fair to code assuming the feature exists; why would it be in your index if it didn't exist, on the assumption your index is functioning correctly.... and so an audio failure would just be the first thing of potentially many that goes wrong.

Play audio could check if audio exists and fail cleanly, but if the object you're trying to call that in doesn't exist, your hot out of luck... and again, why would you not assume it exists... the search you just did said there was a system by that name... but that system is probably just on someone's dev environment.

[1] Two reasons I make this assumption. One is that in my experience, if there's latency in the client or other stuff going on, the audio plays, but the screen doesn't move straight away... suggesting #2... it'd be a good design practice to do audio/other "sensory" things to entertain, while the number crunching happens.
 
Last edited:

[PS4] Filthymick420

[PS4] Filthymick420
Jmanis said:
That's why i keep suggesting it's probably just a corrupt/ bad index entry.

A common problem with searching exists around existence of data. It's a reasonable assumption that your database indexes are correct... if they aren't, that's a problem that needs to be fixed in your database (corollary: it's a bit of a rabbit hole to code assuming that nothing in the project is correct and functional... it can result in major inefficiencies).

Pseudocoding, the galaxy map lookup probably looks like:
  • searchForFeature(name)
  • if no results, do nothing
  • if results:
1. Get first result
2. Play audio for result (as the audio starts before anything happens)
3. Move to result location
4. Etc

... but if we're dealing with a corrupt index, the result probably points to null data. 2. Is then trying to play audio on a feature which doesn't exist which would crash.

It'd be totally fair to code assuming the feature exists; why would it be in your index if it didn't exist, on the assumption your index is functioning correctly.... and so an audio failure would just be the first thing of potentially many that goes wrong.

Play audio could check if audio exists and fail cleanly, but if the object you're trying to call that in doesn't exist, your hot out of luck... and again, why would you not assume it exists... the search you just did said there was a system by that name... but that system is probably just on someone's dev environment.
Click to expand...
Do you think there is a way to find what sound file is attached to it? I'm suspicious of a specific one...
 

Jmanis

Jmanis
Christopher Roberto said:
I don't think so. AudioKinetic gets a lot of blame for crashes in many games because of people knowing just enough to run tools they don't understand the output of.
Click to expand...
But my car is allergic to vanilla icecream!

Kelster87 said:
Because it would be a hard sell for FDev to ban me for using MS tools..
Click to expand...
Would depend entirely on what you achieved with it.

EDIT: Also, quite convinced now this is just a botched index entry, because <reasons>
 
Last edited:

Disemboweled Ego

D
To me the quick crash to desktop seems similar to when you edit the Custom.misc galmap settings file and put a random/non existent/invalid system id64 number into RouteDestinationSystem (which crashes the game as soon as you load in).

So my best guess would be that the system is in the search index but it returns an id that's not actually in the game.

I managed to find one exact mention of Sabiyhan as a proper name on the internet using duckduckgo and it appeared to be a turkish/kazakh name. Stretching it and looking for similar names you can get to Sabiya which apparently means brilliant/splendid. Would need someone who is more familiar with that language group to tell us if it means anything or makes any sense at all.

It's probably still tinfoil, but some other named systems I've seen in-game seem to draw from pretty obscure languages and I'm really curious how FDev got them.
 

VR Jay Le Chardon

VR Jay Le Chardon
How is it this game manages to whip up such a frenzy for something that is probably as innocuous as a little typo? But seeing as we are tinfoiling the heck outta this:

Crimson Echo said:
Systems like some of the PSR systems way below the galaxy do not crash the game when searched.(It is even possible to bookmark them if you are fast enough) Some procedural systems way below the galaxy are only searchable with a full system sting search. (partial searches do not work, but typing out the whole system names does work.)
Click to expand...
Could you tell me more about these PSR systems - I've never heard of them?

Turd Ferguson said:
this is BY FAR the most interesting development in Elite Dangerous I've seen in the last year and a half!
Click to expand...
It's kinda tragic, but also true, and this could be the frist Raxxla development in my 5 years in Elite...

Leeloo Catfish said:
It methods might be forbidden but could still have been done in 3307 if indeed the pilots federation and UC have an influence on what galmap presents its user. I'm not saying I endorse sneaky conduct , but that kind of programming judo gets me 'excited'. Also I'm jealous that I couldn't do it myself :whistle:
Click to expand...
IRL there will always be a balance to the ongoing battle between cyber-law-makers and cyber-law-breakers, essentially as new technology becomes available, through capitalism and market forces, both sides get access to the new toys, and often times through academic and patent filing/publication channels the inner workings are laid bare on public record. It's kind of silly that a child can google how a nuclear bomb works, and could do "a David Hahn" (google "nuclear boy scout"). Short of something game-changing being stomped out by one state, as in developed in a secret military/government lab and not patented, blocked from academic papers publication, and not announced, the tech will trickle down, thats how Jihadists are now able to launch milatary style drones or "UCAVs". So in that vein, the same tech that is available in 3307 to the likes of Pilots Federation, will be available to whoever has the spending power and inclination to buy it. So if the PF quantum encrypted the database, Sirius Corp could buy/build a quantum computer to decrypt it, meaning that in 3307, "file dipping" will probably be a thing just as it is now...

LeKeno said:
The crash appears to be happening in the audio engine part of the code. It could just be an oversight, an odd way of intentionally crashing the game to get a dump and debug something, or a broken scenario which was supposed to play a "congratulatory music / effect" for discovering something special.

Not sure if it's consistently crashing in the same way.
Click to expand...
Correct me if I am wrong, but doesn't every celestial body in galaxy map have the same sound? If so “Sabiyhan” must be super special if it has it's own sound(s).
 

LeKeno

LeKeno
Christopher Roberto said:
I don't think so. AudioKinetic gets a lot of blame for crashes in many games because of people knowing just enough to run tools they don't understand the output of.
Click to expand...

If you can do better, be my guest.
The audio engine is supposed to be shutdown in the stack trace but it's still pumping things and results in reading from a null pointer / access violation error.
As to why this is happening, it could be anything, including being a side-effect of something else going wrong upstream (which is quite likely).
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom