/tg/Station 13

Bottom post of the previous page:

Jaredfogle wrote:

Ayy Lemoh wrote:how the hell do real devs prevent this from happening (in a way that makes it feel like it isn't a common occurrence)? sure, you can playtest it however if there was a game like this then how the fuck would someone think to playtest "oh yeah chems do IN FACT still work in the stomach lol" or some shit? you could say that it would be tested after a similar change however it's clearly not something obvious if we're having this issue right now.

You've been told a couple times now already, but yes--unit + integration tests. Of which we are slowly getting more and more of.

https://github.com/tgstation/tgstation/ ... unit_tests

unit tests and daily build tests by QA

Unit tests are not equivalent to live playtesting, but they can help stop some of the more egregious issues.

The real issue is that the bugs were really basic and gameplay impeding.

I'm available pretty often to testmerge your big prs and make sure to always heavily test locally!

its really funny that c*ders are all WE DO THIS FOR FREE SO YOU SHOULD TEST IT FOR FREE and then when people provide bug reports and feedback for free they shit themselves and screech that they're underappreciated and can't be expected to fix everything

Super Aggro Crag wrote:its really funny that c*ders are all WE DO THIS FOR FREE SO YOU SHOULD TEST IT FOR FREE and then when people provide bug reports and feedback for free they shit themselves and screech that they're underappreciated and can't be expected to fix everything

coders do not have enough commonality for you to refer to them as a group

woah did someone say
OH GOD NO SOMEONE SAID IT

SOMEONE SAID THE CURSED PHRASE
OH GOD THERE IT IS AGAIN

Fuck Discord, fuck Github. I offered what can I do with ease, commiting three reverts via direct repo access. Did not look like my pubkey was given access and some people actually picked up fixing those bugs instead of shitposting here and defending merges of untested code - as seen with other FOSS projects, most basic levels of "QA" is not too much to ask for.

The real issue is that the bugs were really basic and gameplay impeding.

This.

Ivuchnu wrote:Fuck Discord, fuck Github. I offered what can I do with ease, commiting three reverts via direct repo access. Did not look like my pubkey was given access and some people actually picked up fixing those bugs instead of shitposting here and defending merges of untested code - as seen with other FOSS projects, most basic levels of "QA" is not too much to ask for.

The real issue is that the bugs were really basic and gameplay impeding.

This.

Just open github issues documenting what is broke and people will see them and pr fixes.

We don't give out direct repo access to random shitposters with zero history of contributions to said repo and it's hilarious you think that would ever be entertained.

For a volunteer project bugs are really easy to overlook or miss even when testing, I've lost track of the number of times I've overlooked something really simple on my own code, in the scheme of things these were really minor issues.

Game you play for free, developed and maintained by other people for free, occasionally has buggy code introduced for short periods of time (for free).
Oh the humanity, what an outrage...

I also get the impression that anyone complaining about this sort of thing here, has not played much on other codebases, because I'm pretty sure TG is light-years away from being the worst offender.

Armhulen wrote:

Super Aggro Crag wrote:its really funny that c*ders are all WE DO THIS FOR FREE SO YOU SHOULD TEST IT FOR FREE and then when people provide bug reports and feedback for free they shit themselves and screech that they're underappreciated and can't be expected to fix everything

coders do not have enough commonality for you to refer to them as a group

when I say c*der I mean c*ders, not "people who happen to code." there's a difference, it's like the difference between someone who poops, and a shitter

I don’t get it.

It's not fun to introduce bugs, especially serious ones. It's extremely stressful and feels incredibly shitty. I'm not even talking about the community response here, I mean just the literal knowledge that someone's game is being ruined and it's all your fault is one of the worst feelings.

Yes, it's bad, but it's also inevitable. Nobody's perfect and shit happens. But please, just consider how often this happens. It's not that often in the scheme of things.

MisterPerson wrote:It's not fun to introduce bugs, especially serious ones. It's extremely stressful and feels incredibly shitty. I'm not even talking about the community response here, I mean just the literal knowledge that someone's game is being ruined and it's all your fault is one of the worst feelings.

Yes, it's bad, but it's also inevitable. Nobody's perfect and shit happens. But please, just consider how often this happens. It's not that often in the scheme of things.

i'm going to hug every coder who makes a bug

it's hilarious you think that would ever be entertained

It's not. Why bad man asks me to fix those bugs as I reported them in initial post? I've been given unreasonable request, I returned the favor.

hey man seems like you're really struggling to get into DM script, I can offer you a personal lesson? just add me on discord.

VexingRaven wrote: "basic shit like mech icons not changing" in what world is this "basic"? I think I've gotten in a mech like twice in my life, you're acting like this is some core function that the game is unplayable without.

in the context of a mech rework that is pretty basic
i dont even care about the thread i just want to point out your dumb

Basic on the user end of things but not on the functionality. The mechs still worked, they just looked jank as fuck. And given just how big an undertaking it is to take one of the shittiest pieces of code we have and make it not shitty is, I think it turned out great. Besides the...few minor issues initially. (Exploding mechs).

But that's all fixed now. Mechs work great.

(Except they just explode on death but that's not a mech problem that's a vehicle problem and maybe an unintentional revert of a balance pr.)

Sometimes shit is testmerged but the catatrophic bugs never see any actual issue reporting at all. I am pretty sure when material datum stuff being added to all items was initially tested on Bagil, it never saw any reporting despite some of the absolutely horrific bugs attached to that. Like plasma objects being bombs and uranium objects being radioactive. I think a lot of focus was being directed towards uranium ore itself being radioactive that nobody seemed to be aware that suddenly all their plasma gear was explosive (I think miners even misattributed their radiation poisoning to the ore and not what was really killing them. Their gamer gear).

And this was because of how the testing from Floyds end was likely being done that he missed it. He was I believe spawning stuff via admin tools and not constructing via the lathe.

Typical players are not going to spot it because they don't have the same perspective or eye for dysfunctionality if it's something subtle and a change they aren't fully aware about. The only reason I discovered most of the bugs the day it was because I was an admin at the time and I could see what was happening via logging, and put two and two together. I was amazed this stuff got missed as well but in hindsight it's pretty obvious why.

It's basically a little lacking in perspective to claim every coder should be able to perfectly test every aspect of their changes when most probably don't know how to conduct a proper and thorough test, or their test methodology is ineffective and has gaps, even if it might catch most of the functionality bugs. In a professional scene, this kind of quality testing is done by a team of professionals going over a game with a fine tooth comb. Our contributors are a few people or just one person doing this as a hobby. And even a professional team can miss bugs. Because all of it is a series of human errors, even the code.

A thought I had just after seeing this entire thread is, why not use testmerging more often? Testmerge a PR for about a day, and then ask players to play with whatever mechanic is being changed and report any unintended concequences. If there are not true gamebreaking/gameplay impeding bugs. Finalize the merge and have a cleanup PR grab the bugs found in the testmerge.

We do test merge the big changes and big additions, we ask players for feedbacks.. Most of the times we don't get an answer. Just see my healium gas, when I was adding the new gases we did a week of TM, I opened a thread on the forum, asked on discord, asked in game often, even a post on digg.. Nobody reported shit, we now have a PR to fix healium knocking you down without warnings or counterplay. Later I discovered that some players knew this but didn't bother to report it.
We do many tests, we ask for help from other coders, we do test merges, but players should understand that we can't try every single interaction, that's why we rely on them for the massive testing for bug and balance issues, but if they don't bother to report problems, they should just be quiet and stop complaining.

atmos stuff is rarely ever done and usually by the same people, they probably tought it was a feature for wasting 20 min lol

see TGMC for why test merging tons of stuff at once is hell and not an actual solution.

Ghilker wrote:Later I discovered that some players knew this but didn't bother to report it.

There are three reasons why, and they are equally valid.

1) They didn't think it was broken, it was working as intended, but it was obviously quite powerful. Basically, literally ignorant of this being a problem but accepting of it as the new state of things. Complacency with the state of things is easily a massive detriment that can really do more damage than active reporting. I've seen some bugs elsewhere exist for literally months that ruin the game but completely go unfixed because people just accept it as 'the way of things'. We don't seem to have that too much unless it's a particularly benign thing that provides more QoL and game-flow than detracts. Like uh the bug ATHATH made into a feature with the shuttle curse. Even I was surprised to learn it was a bug.

2) They did know it was broken but wanted to keep it quiet to abuse it. Rarer but people like this do exist, people will flat out lie about what they're doing even if you tell them you know exactly what they're doing, because having a closely guarded secret is a tool to which you can exploit in-game, even if it is because of bugs, and reasonable deniability is going to prolong getting away with it before anyone realizes what they've been doing all this time. Some stuff goes entirely unnoticed for an extremely long time due to these kinds of people, who are likely making note of problems in the code but not moving to fix them for their own benefit. It's the unfortunate consequence of the game being treated with such a competitive fervor.

3) They did know but they just didn't want to report for some misc. reason, most likely because of either laziness or because they can't for some reason. They probably will admit it if questioned about it. They likely just aren't the proactive type about dealing with bug reports.