Legacy Of Kain Soul Reaver 1 & 2 Remastered has been released.

It contains a lot of bonus material, including a few areas that have been scrapped before release.

Exploration of one of those areas may end up _crashing_ the game...

...Imagine that...

In case anyone is wondering:

Bonus Content -> Lost Levels -> Smokestack -> Turel's Audience Chamber.

0. Approach the doors at the top to trigger the cutscene && wait till it's finished.

1. Shoot a projectile at the doors at the top to open them.

2. Travel the corridor up as far as you can, until you've reached the dead end.

3. Try going back to the area w/ Turel.

...Boom.

...at least on Nintendo Switch.

This one has made my day.

Proof:

https://devrant.molodetz.nl/LOK_SR1...

@retoor: Thanks for providing the storage space - finally got to use it.

Feel free to delete the video if you ever need the space or something.

Side note: two previous uploads ended w/ 500 error for some reason.

Files in question:

LOK_SR1_2_Remaster_Turel_Crash_Nintendo_Switch_v1.0.1.mp4

LOK_SR1_2_Remaster_Turel_Crash_Nintendo_Switch_v1.0.1_reupload.mp4

Just in case you'd like to know.

You can delete those as they are corrupted, lacking the ending.

...actually all it takes to crash the game is to make the boss follow you outside the initial area.

Sounds like something Bethesda would do. Modders have spent a decades fixing their bugs. A good modpack for Skyrim is actually a better experience than the original game. Am playing one now that tries to capture the original game feel, but fixes everything and adds a bit more content. It has crashed once in about 100 hours of play.

@D-4got10-01 finally one uses my great storage service! What a nice high quality video on such a fast server! :) I will raise upload limit if it's used for video to 300Mb.

@retoor Edit: Upgraded to 300Mb now. I have to configure some limits especially with the 'hacker wars' going on. It's probably not smart to have a no-signup upload service now :P

@retoor Much appreciated.

The thought of hitting an upload limit has crossed my mind.

If I find any 'gems' in the future, I'll be sure to share them as well.

@retoor Yup.

Wise decision.

@Demolishun Yeah.

The 'modders fixing buggy games' also reminds me of the Vampire The Masquerade Bloodlines.

The game is great, but pretty much unplayable w/o the unofficial patch.

Too bad the current trend in video games is

'Let's release the buggy game. We can patch it later.'

@D-4got10-01 that's all software these days. Half assed because it can be patched later. It's especially the mentallity of web developers (at least where I worked) -> everything is OK and fixable -> continues deployment whopaaa. I've worked with sentry servers reporting over a million issues (cloud shit). Mass bug fixing became task for me because I cared, later other developers willingly joined and we got a team squashing bugs. We time boxed finding root causes to a hour, if we couldn't find it, we just fixed the data corruption it caused. Many bug fixes resolved in other issues getting resolved as well. The fun of working at my team was that you made a LOT of progress and made stake holders happy. But we also prioritized the 'new' bugs because older bugs - those customers were already pissed off and gone of course.

@retoor Yeah. The prioritization of new vs old bugs sounds about right.

Same here.

The constant patching of things is a sad side effect of the ubiquitous Internet nowadays.

...or at least the assumption of its ubiquity.

@D-4got10-01 not sure about other games, but designing a game to be moddable helps a lot with the bugs. But holy crap. The modding community around TES games is writing c++ code that injects itself into the engine to add more scripting options. Way beyond the original modding design of the games. There are also a bunch of dll mods that fix memory issues and stuff. Just wild IMO.

@Demolishun yeah, but I think making much things moddable is a lot of work. All those interfaces, probably a mapping between C++ or C# to Python or LUA or some other scripting language. DLL's are great for mods. I also made a dynamic DLL loader for one of my scripting languages with a certain interface. Put .so file in directory and whoops, part of the programming language. Maybe I'll implement that in my current language again, but I'm still at parser stage.

@D-4got10-01 also, old bugs were many times not reproducible due not related software changes. So the old bugs took a lot of investigation time too. Another reason why the old bugs were less popular. For most we just did data corrections or whatever needed since no customer was waiting on it anymore.

@Demolishun Damn. People modding via code injection?

That's just awesome.

@retoor yeah, I play around with SKSE plugins which end up being DLLs. There is even a completely game version agnostic middleware for writing these plugins for SKSE.

@retoor This one is interesting.

Bugs disappearing w/ unrelated changes.

Though I've seen many memory-related issues causing all sorts of weird behavior.

Fixing the memory issues would resolve those.

Also, sometimes there are other factors at play.

Quite recently we closed a report not because we had fixed it, but because it was an OS-level issue.

Console manufacturer fixed it on their side.

@D-4got10-01 for example: email sender crashes -> cause: invalid email address. Someone later added unrelated an validator on the email field (server side ofc) and bam - solution solved with no direct relation.

My nicest memory bug was smth like:

user_t * user = (user_t *)malloc(sizeof(user));

It caused bugs in other places in the application. I had SO much luck by just SEEING it very soon before I went hardcore debugging while it was even on a very unrelated place where the app crashed. Just dumb luck.

@retoor One of the more interesting issues that I remember hearing of

/* as it was in a project finished prior to me joining one of my past companies */

is an issue in a multiplayer game.

IIRC it was a sports game of sorts on one of the portable systems.

The game's designer had been reporting many issues in the MP mode to the lead programmer.

After roughly two weeks of debugging, the programmer _finally_ found && fixed the issue.

Turned out the root cause of _all_ those issues reported by the designer was the Random function called at the beginning of the match.

The thing was, the function was being called _on both machines_ - the Host's && the Client's, instead of just the Host's who would then send over the result to the Client so that the game would be in sync.

Effectively the Host && the Client had been playing two separate games.

Unfortunately since I only heard about this from the lead programmer I have no idea what introduced the issue in the first place.

...could've been because of the development hell.

I _think_ the programmer had been complaining about the designer constantly wanting design changes, but my memory might be misleading me.

@D-4got10-01 do you program games? For portable devices? I'm not into gaming, but creating one must be awesome and a real skill thing. You can make it as advanced as you wish. Making performance / code quality trade offs and stuff.

Haha, the random function. Causing also issues because:

1. Forgot to seed making it producing the same 'random' numbers over and over.

2. Not being so random as what was the case of the random function of MySQL for a while. It only had one job, hihi.

@retoor I've been dealing w/ portables some decade ago.

Nowadays I'm dealing w/ current gen consoles.

Game dev is interesting in that the whole thing is dynamic.

You can't really play two exact same games.

There's always going to be some difference between playthroughs.

My current project is a multiplayer game, as well.

A lot of issues due to networking.

Many caused by lack of sync due to pings.

It's wild.

...also, your point number 1 reminds me of:

https://xkcd.com/221/

I love that comic...

@D-4got10-01 haha, that comic is amazing. Will AI in the future test games? What languages are you using and is there much math involved?

Networking is a bitch. I had a socket fetish since I was young and still have. Since a few years ago I own the book "C network programming". Does also not contain everything to learn. Things like ignoring the signal PIPESIG, else your program crasshes when it happens. Also it doesn't explain you about draining (SHUTDOWN before seinding you last batch) so all data gets pushed before disconnecting. Good networking is appearantly not learnable by book or internet. It's a magic combination of both and fucking alot. You're probably using a low language. Higher languages (python e.g.) do have the issues like I described above for example fixed within already. They don't crash on pipe and drain well using sendall(). I just recently have learned that sockets have a buffer size you can set with setopt().

But yeah, interesting stuff.

@retoor Ah, apologies - must've gotten distracted, so I forgot to tackle that part.

I am sure that I am _bound_ to disappoint you && others here, but I spent most of my time in game dev as an internal tester for game dev studios in close cooperation w/ programmers, while programming remained my 'road !taken'.

I had a chance to become a programmer right after finishing 2y college, but I chickened out.

I do value honesty, as such I'm mentioning this.

I'm currently learning C++ thanks to 'Beginning C++17', as in game dev the C / C++ are dominant.

/* Yes - the newest one is 23 */

Java is also used, but it's more of a mobile games thing.

Python is being used mostly for scripting automated tests, IIRC.

Something I might also consider.

So programming skills-wise, I'm a beginner-in-the-making.

Anyway, over the years I've seen some funny / retarded things in game dev, which I'd like to share here over time hopefully to your amusement.

...

...Also, generally speaking most testers are retarded.

Very true of many external ones who are unwilling to learn, just 'play'.

I've always preferred hanging around programmers.

My best friends are programmers - the good ones.

Sure they do fuck up here && there, but we all do at some point.

I do too. I'm !perfect.

Programmers tend to be witty, have an interesting humor, create things.

I've always admired that about them.

Yes - there's quite a lot of math involved in game dev, although it also depends on the thing one programs.

Physics, graphics need a lot of it.

I'm also quite interested in networking - I have a nice setup for extensive network tests.

Another thing that interests me is hacking games, so learning Assembler could be highly beneficial.

Currently I know about 'nop', though.

I like hacking single player games for personal amusement, to learn how things work in the game's memory,

learn about protections && such.

...

...e.g. games made in older RPG Makers tend to have this protection where value displayed on the screen is stored in memory as displayed_value * 2 + 1;

Hacking leaderboards / MP games is fucked up && I'm highly against it, though.

Only losers who want to _feel_ superior do that, because they lack the skills to be the best on their own.

My recent interest in Linux came to be because of one of the issues that has been reported for a Steam Deck.

Testers around me said 'we don't have the HW, so sorry - can't verify'.

On the other hand, I though... 'Hey... Stem Deck runs on Linux... sure it's modified, but still.'

I have managed to create a setup proving them wrong, that we are actually capable of this w/o the real HW.

So this would be an example of uneducated testers !wanting to learn.

I want to know how things work, I do _a lot_ of research googling things && reading documentation about how the platforms work.

...

...There's also a lot of 'It's probably OK.' or 'I think' mentality among many testers who want to talk between themselves to _assure_ themselves that they're right.

If I just don't know something platform-specific that I should or would benefit from knowing, then I don't hesitate to ask on console forums.

I could continue, but wow... this has already become quite a long confession.

I do apologize for !mentioning this sooner && I hope that you don't mind me hanging around.

To answer your AI testing games question - I don't see it.

It could help w/ some automation e.g. for smoke tests, but there is _a lot_ of things that just need to be done manually.

...damn... what a huge wall of text this is...

@D-4got10-01 I missed all the text because I was mentioned by too many spam bots than the mention list could handle. Having a skilled tester is worth gold, hard to find I guess.

Hacking in general is interesting. I've read three books about it. This one is the favorite tho.

@retoor No worries.

You got around to them in the end.

Also - I know that book.

I've obtained a copy in *.pdf during one of the Humble Bundle promos.

@D-4got10-01 good purchase but I couldn't read such from a pdf. Tomorrow I'll have my new neural network book. I'm so excited. Had to wait two weeks. Happy hapoy happy.

@D-4got10-01 look what i'm working on: https://molodetz.nl/retoor/...

It's a project to replace content in http without making corruption by updating the content length. It's a proxy that is able to be implemented unnoticed by client or server. I use this for a good usecase tho. I have my site, I also want it with my real name. It will replace retoor with my real name life. So, somewhere else molodetz will run with my real name under different name.

@retoor The small pleasures in life.

I completely get your enthusiasm.

@D-4got10-01 What small pleasures? I'm only talking about the big ones currently :D

Yes, i'm enthousiast. But I've managed to make a memory leak in python somehow. I first have to fix that before I can start something else.

I'm scrumming myself. Last week I had 30 tasks. Not, I have have 30 tasks. DAMMIT.

But I do complete stuff good these days, it's only going forward most stuff. Not a lot of maintenance.

How're you doing?

@retoor Taking a look right now.

Should be helpful w/ that portfolio of yours.

I've seen you mentioning the desire to make a copy of the site w/ real name for those potential future employers in some other post, too.

@retoor Small pleasure as in new book -> new knowledge; new thing to play w/.

@retoor Yeah, that mem leak in python is weird.

Good hunting.

I'm doing fine.

Making slow, but steady progress each day.

@D-4got10-01 exactly. I call that a big pleasure :P I have the book here now. It is called "Neural Network Projects with Python". and I've skimmed it a bit and I think it's the correct level. I would be lost in the hardcore math stuff, this book as a bit easier, but still really training own models for recognition and stuff. Sadly, not a lot about text - what is kinda my goal. I will read this book in bed. I study better without a laptop in front of me. I spend much time on it but also have other things to do. I think it will take a month to study and do the examples.

@D-4got10-01 yes, I'm making an application that replaces content and updates the http content header length so there won't be any corruption. So, I did had to almost fully implement the basics of a http server with websocket support. I'm quite happy with the flow and quality of the application. Check the AsyncReader and AsyncWriter and Socket class. I'm happy with result. I test my application on any buffer size. It even works with a 1 byte buffer. I test everything with apache benchmark and it does concurrency with 500 connections easily. Every connection costs two threads tho. I configured 2500 threads, why not. I think they start lazy.

@retoor https://amazon.co.uk/Neural-Network...

?

£30 + shipping seems affordable.

@D-4got10-01 yes, exactly, many AI books are quite expensive and also, my interest in AI is quite limited. I just realize that a pretrained model is just very overkill for things I want so want to make my own models. I normally develop things in C because I want to learn everything under the hood. But for AI, that's not the case. I even doubt if i'm smart enough to understand what is really under the hood. The idea behind it. I think it's too complex for my level of interest. This book seems just about my level :) Lucky. I have a another one called "Machine Learning with Python". And i've never read it. It was only graphs and stuff, no practical examples like this one.