Jump to content

Steam Engine is Super Efficent? Bug or


Recommended Posts

My steam engine is very SLOWLY losing heat with my loop, which makes keeping the loop at 100% insanely easy. Has to be a bug right? I stared at this heat for a few mins and it barely changes.

 

"edit" it looks like my radiator doesn't absorb heat either, but still keeps everything cool, so it's a bug, it transfers the coldness but doesn't absorb heat ^^

 

Spacejunk Cycle 324.sav

20190716201441_1.jpg

2 minutes ago, paranoid945 said:

because radiant pipe totally not work,  since latest update. I will wait until next update.

I guest the game developer doesn't know the meaning of unit testing right?

Yeah everything worked fine till this new update haha I didn't even realize till my steam engine was running at 100% all the time

3 minutes ago, Oozinator said:

You know what "preview" and "beta" means?
You can do the testing, or play stable branch.
 

As a 15 years experienced software engineer/ system architect, I can say the core system should be always automate tested/ unit tested, because these functions will not change easily in the long run, as game function continues to grow, you just can't hire more dev to get rid of the code, and the game code will finally explode and totally out of control, I don't want that to happen.

I love looking at people's save games - thank you!  It's great how you use fertilizer makers to power your NG generator full time.  You definitely got my attention with the title and I was looking forward to @mathmanican min/maxxing this setup.  Unfortunately I wasn't able to see it running, maybe because I spent too much time gawking at your base?

EDIT:  I see - I wasn't running the experimental branch.  I can confirm that the radiant pipes don't work in this version.  Now neither do any of my old save games, even when uninstalling and reinstalling the game without beta branch enabled!  Bizarre.

1 hour ago, paranoid945 said:

As a 15 years experienced software engineer/ system architect, I can say the core system should be always automate tested/ unit tested, because these functions will not change easily in the long run, as game function continues to grow, you just can't hire more dev to get rid of the code, and the game code will finally explode and totally out of control, I don't want that to happen.

Save them! https://www.klei.com/careers

11 hours ago, paranoid945 said:

because radiant pipe totally not work,  since latest update. I will wait until next update.

I guest the game developer doesn't know the meaning of unit testing right?

This is funny.       Early Access = Testing.    Preview Branch = Testing.  This is all before posting it to "live" Early Access  and   before launching.       So we have  2 public testing.   Not to mention their internal testing.      3 Levels of Testing before going to launch.  

What exactly is the meaning of unit testing  again?      WE are the testers.

2 minutes ago, RonEmpire said:

What exactly is the meaning of unit testing  again?

The meaning of "unit testing" is that you have the code (unit tests) which tests your actual code. The purpose of unit tests in particular is to have logic and calculations tested, so that your code doesn't add instead of multiply because of the typo. Now I am not sure to which extent this can be applied to a game code on the Unity engine, but this is a pretty common standard in the software development. In general, automated tests can be added to pretty much anything with enough stubbornness applied, but whenever the time spent to write and update them will be less than the time needed to test and fix without them is still a debatable topic. Although I know that a lot of colleagues will say that there's no doubt and having no unit tests is insanity.

Being said that, even with an extensive suite of unit tests implemented, bugs like this can still go through. So we don't know if Klei uses unit tests or not. I would actually assume they do, I can't imagine running a codebase like this with a quality we have (the game is playable for most of the time) without automated tests.

21 minutes ago, miauly said:

Now I am not sure to which extent this can be applied to a game code on the Unity engine, but this is a pretty common standard in the software development.

Based on personal experience, this is likely the problem. When you look at how much everything hooks into everything, it becomes hard to isolate a single piece of code and write a unit test for it. Sure you can likely find a bunch of examples where it's somewhat doable, like liquid bridge transporting liquids in a certain way.

However when it comes to more complex systems, like temperature and gasses moving around, writing a unit test becomes tricky because suddenly everything seems to depend on so many other parts of the code.

Considering the importance of temperature handling, I would likely make a mod with sensor like buildings, which once a second report to a log file. I would then make a savegame and spam those sensors around. Once in a while I would start the savegame and run it for a cycle or two. If the log entries are formatted correctly for scripted reading, some script or program could then analyze if temperatures are more or less the same as they were prior to whatever change they made. Not really unit testing, but at least it's somewhat an automated test. Other sensors could log pipe flow and stuff at the same time, possibly to a different log file. It would get around the issue of isolating code despite all code hooking into each other.

We don't know if Klei does this or something similar. All we know that whatever test they do, they miss some bugs once in a while. That's unavoidable and we have to admit even if they catch 99% of the bugs, we would still be talking about the remaining 1% with people asking if Klei never test their software. Bugs occur, particularly in the test branch, which acts mostly like a nightly test build. Yes it can be annoying when the game suddenly breaks. However the better they test what they release, the longer it takes. The longer they have to spend on releasing frequently, the less bugs they have time to fix on a daily basis.

It's also worth adding that whenever people declare the test branch as unplayable, people stop playing and suddenly have way more time for the forum. I wouldn't be surprised if you can somewhat accurately detect game breaking bugs by reading a graph of the forum activity.

3 hours ago, Nightinggale said:

Considering the importance of temperature handling, I would likely make a mod with sensor like buildings, which once a second report to a log file. I would then make a savegame and spam those sensors around. Once in a while I would start the savegame and run it for a cycle or two. If the log entries are formatted correctly for scripted reading, some script or program could then analyze if temperatures are more or less the same as they were prior to whatever change they made.

This is integration testing and indeed it's not possible to test correct integration with unit tests. And using Unity is in fact integration. Not surprisingly, when I googled both unit and integration tests for Unity, it turned out they exist and look pretty similarly to what we have in completely different and totally boring conventional software development. This is completely off-topic of course, but I thought you might be interested to take a look as you implement mods, and these tools might be applicable for your code as well. (this is where you get on forums when pipes are broken, indeed)

14 hours ago, __IvoCZE__ said:

You are literally on the frikking beta branch where you know... things are tested?

Sometimes it is very curios! They add something and if they would test their own creations, they would notice, it doesn't work. No tester needed, just test what you have written (code). Other EA titles have also beta and preview branches like Factorio, but there, you must hardly search the errors! If you find one, the developer had just overlooked a possibility. Here, at ONI you test buildings they never worked at the beginning. What is the apologize? Hadn't time to test it? I thought they have a big savegame with all buildings and test by themselves, if something is broken or not.

Your point of view is hardly odd. We all know that errors are probably included. But the most errors are errors they could find by themselves it they would test their own code/buildings! Not so difficult. No excuses for lazy coders!

Factorio has basically no physics engine.  It's not really that comparable imo.  Source software dev with a physics degree.  Based on the frequency of updates the devs are definitely very hard working.  Thanks devs :).

Edit: I do share your pain in not being able to play today :(

4 hours ago, Nightinggale said:

It's also worth adding that whenever people declare the test branch as unplayable, people stop playing and suddenly have way more time for the forum. I wouldn't be surprised if you can somewhat accurately detect game breaking bugs by reading a graph of the forum activity.

LOL thats why im cruising the forums. I started an early base on Rime just to mess with.

1 hour ago, DustFireSky said:

Your point of view is hardly odd.

Bruh you talking to a guy that has a realistic maxwell pfp.

1 hour ago, DustFireSky said:

We all know that errors are probably included. But the most errors are errors they could find by themselves it they would test their own code/buildings! Not so difficult. No excuses for lazy coders!

It's on a selectable branch so that way it gets tested, bugs get removed and it gets shipped to the public branch, also klei isn't a big studio to have a crap ton of people working on just one game.

12 hours ago, paranoid945 said:

As a 15 years experienced software engineer/ system architect, I can say the core system should be always automate tested/ unit tested, because these functions will not change easily in the long run, as game function continues to grow, you just can't hire more dev to get rid of the code, and the game code will finally explode and totally out of control, I don't want that to happen.

You write unit tests every time you write temporary in progress code? I'm surprised you get anything done at all.

Archived

This topic is now archived and is closed to further replies.

Please be aware that the content of this thread may be outdated and no longer applicable.

×
  • Create New...