[1.8.x] Oh Scrap!- A ScrapYard based Part Failure and Reliability Mod 2.0.1 (07/12/2019)

[WuphonsReach]

Was the code to automatically trim out low-reliability parts from Scrapyard ever implemented?  (First post is a bit out of date. The posts that I've +1'd this week had information not in the first post.)  I don't see anything in DefaultSettings.cfg which would let me say "auto-recycle parts with a reliability under 2/10".  I'd still want the funds from StageRecovery, but would want them to be auto-trimmed instead of being put into Scrapyard's inventory.

I think SRBs are the lightning rod because of how they get used (radial, multiple) and how fatal to the launch that loss-of-thrust failure at lift-off is (due to imbalanced thrust combined with low altitude giving little to no time to recover).

I've had a bit of luck adding more verns/RCS ports to the tail of the rocket.  In situations where I have 6 radial SRBs, it can be enough to overcome the loss of one off the pad.  But staging early can still be a better choice (and praying that the SRBs don't impact the stack as they fly away).

[DasValdez]

@severedsolo... So, we actually started UHC Career today, and seem to be having some weird behavior with the mod I didn't observe before. Using your version 1.4 based on the changelog file.

SRBs don't seem to be properly incrementing their generation. I fly a mission, decouple the SRB, recover the pod, go back to build a new rocket with a new SRB... and the SRB is still gen 1. No matter what, it always stuck at gen 1. Previously, we got our SRB gens up to 30+ by flying tons of suborbital tourism missions... decoupled and destroyed the SRB every time but it would still up the gen. Have not actually confirm if this is SRBs only, it was just the glaring issue so far.

The only way I've gotten an SRB to increase it's gen is by literally setting up a rig that rolled it off the launchpad, using the stock recovery button to recover the actual SRB itself, and then going back to build a new SRB. Do you actually have to recover the part itself to increase the gen now? Unfortunately, that's not really workable in a real career. Am I missing something?

----------

We're also using OhScrap with KRASH, which is maddening with low-reliability SRBs. I wasted stacks of kerbucks setting up a sim to test the flight profile of a rocket, only to have the SRB fail at launch and get no actual data from the sim. That, coupled with the fact that I couldn't actually upgrade the generation on the SRBs was driving me insane.

In a sim, there should be no random failures. There should be intentional failures, ie, right click on a part and force it to fail... but no random failures. It seems like there may be a line of code to prevent sim failures, but maybe it doesn't apply to SRBs?

 

If I'm completely missing something, let me know, but whew! Was a rough first day of Ultra Hardcore Career due to mod weirdness. 

[severedsolo]

Oh Scrap 1.4.1 Released

• Fixed SRBs being allowed to fail during KRASH simulations

@DasValdez - I wasn't able to reproduce the first issue you had with Generation not incrementing on SRBs, it's working on a near stock (Just Oh Scrap and dependencies) install. I've added some extra debug stuff in the 1.4.1 release though for the code that refreshes it, so maybe you can ping me a log if it happens with 1.4.1 and I'll take a look.

[DasValdez]

7 hours ago, severedsolo said:

Oh Scrap 1.4.1 Released

• Fixed SRBs being allowed to fail during KRASH simulations

@DasValdez - I wasn't able to reproduce the first issue you had with Generation not incrementing on SRBs, it's working on a near stock (Just Oh Scrap and dependencies) install. I've added some extra debug stuff in the 1.4.1 release though for the code that refreshes it, so maybe you can ping me a log if it happens with 1.4.1 and I'll take a look.

Gotcha, will install 1.4.1 and see if behavior continues. 

To confirm, no part failures should occur in simulation mode. I understand that "right click, force failure" is out of scope currently. 

 

[severedsolo]

1 hour ago, DasValdez said:

To confirm, no part failures should occur in simulation mode.

Correct I'd actually already written "Krash Simulations never fail" into the "am I allowed to fail" method, I just forgot to actually call that method anywhere XD - luckily it was a very easy fix, literally one word.

1 hour ago, DasValdez said:

I understand that "right click, force failure" is out of scope currently. 

Well, actually it's in the code, it just only works when I compile in Debug Mode - and you don't want to see that, it's information overload.

Edited December 19, 2018 by severedsolo

[DasValdez]

SRBs

https://clips.twitch.tv/QuaintJazzyClipzBrainSlug

Still

https://clips.twitch.tv/YummyMagnificentScorpionCharlietheUnicorn\

Failing

https://clips.twitch.tv/ImportantExpensiveSwallowKappa

...in KRASH.

1.4.1 Fresh off the download. On the plus side, the release/reinstall did seem to resolve my issue with parts not gen upping. 

[severedsolo]

1 hour ago, DasValdez said:

SRBs

https://clips.twitch.tv/QuaintJazzyClipzBrainSlug

Still

https://clips.twitch.tv/YummyMagnificentScorpionCharlietheUnicorn\

Failing

https://clips.twitch.tv/ImportantExpensiveSwallowKappa

...in KRASH.

1.4.1 Fresh off the download. On the plus side, the release/reinstall did seem to resolve my issue with parts not gen upping. 

Ugh, thats what I get for deploying a fix without testing it, that should have done it though.

Will take a look tonight if I get time, I have alot on today - can I assume that you are still on 1.5.1? - I have no intention of updating anything to 1.6 right now.

[severedsolo]

Oh Scrap 1.4.2 Released

• Fixed SRBs failing during KRASH sims (for real this time)

[DasValdez]

12 hours ago, severedsolo said:

Ugh, thats what I get for deploying a fix without testing it, that should have done it though.

Will take a look tonight if I get time, I have alot on today - can I assume that you are still on 1.5.1? - I have no intention of updating anything to 1.6 right now.

Yes, 1.5.1, no intention of moving to 1.6 until 1.6.+, whatever that turns out to be. 

Also, thanks. I'll be your test team anytime. Will try the new version, and please excuse the repeated look of abject defeat. Just keep swimming... we worked around it. 

[DasValdez]

@severedsolo Heads up, Github file is missing the .zip extension. Renaming resolved on my end. 

[severedsolo]

4 minutes ago, DasValdez said:

@severedsolo Heads up, Github file is missing the .zip extension. Renaming resolved on my end. 

Thanks - fixed now.

[gamerscircle]

Would sepratron fall under this?  I am getting a lot of failures on Generation 51' with a safety rating of 10, they are highlighted in red, but no message.

[severedsolo]

59 minutes ago, gamerscircle said:

Would sepratron fall under this?  I am getting a lot of failures on Generation 51' with a safety rating of 10, they are highlighted in red, but no message.

Yes sepratrons would probably count - there is a bigger issue with SRBs - I just can't find it right now, the code to fail them is being called from somewhere unexpected (this probably explains the too high SRB failure rates people are reporting)

[DasValdez]

Update, on 1.4.2 we didn't see any failures during KRASH sims, and the people rejoiced (yay). Thanks again! 

We did see an issue where the failure roll was being called multiple times per second before launch, but it may be a conflict with World Stabilizer... something is causing our rockets to roll for fail maybe twice per second while they're sitting on the pad. Not a showstopper, and potentially would be resolved once we get to launch clamps.

One growing frustration is that we can't see part generation in Scrapyard (as discussed before)... it's getting difficult to find the higher gen reliable parts, or at least clean up the lower gen stuff.

---------

Feedback: Doing the "preflight test" for parts begins to get a little tedious when you're combining reusable parts with new parts. Scenario: a reusable crew module with a disposable SRB. If you integrate the whole rocket and roll it out to the pad, your booster will be "untested" and thus lower the reliability. If you recover the entire ship, now the booster becomes tested (increasing reliability) but the crew module gets one more previous use (decreasing reliability). 

The solution is to create the booster separately, roll it out to the pad alone, recover it, scrap it, then create a new ship using the reused crew module and the "tested" booster parts... doable, but tedious. We were actually building booster 5 packs, but at the end of the day it's still just rolling out a partial ship and recovering it immediately. On top of the tedium, you're also only really getting 4 uses out of a part, since you have to burn the first use on testing. 

An elegant solution would be to allow "Preflight Inspection" to the flight scene PAW. Right click the part, and click a button to do a preflight inspection, which flips it to tested, gives it the reliability boost of a tested part, and incentivizes you to have an actual pre-launch checklist you need to run, ESPECIALLY focusing on newly built parts. Forget your preflight, or miss a part, and you run with the lower reliability. Also gives you a good reason to do a final check on part tweakables, since there's the weird issue where a part will lose it's settings sometimes if you replace it from Scrapyard inventory, or click "use new part". Chutes lose altitude config, engines lose throttle limit, etc.

[severedsolo]

1 minute ago, DasValdez said:

We did see an issue where the failure roll was being called multiple times per second before launch, but it may be a conflict with World Stabilizer... something is causing our rockets to roll for fail maybe twice per second while they're sitting on the pad. Not a showstopper, and potentially would be resolved once we get to launch clamps.

This is probably the same issue that I'm trying to track down with the SRBs - but that does help, because it tells me where to start looking.

2 minutes ago, DasValdez said:

Feedback: Doing the "preflight test" for parts begins to get a little tedious when you're combining reusable parts with new parts.

I'll take that one under advisement - Oh Scrap is kind of on the back burner at the moment, other mods need some serious attention., but I do like the idea

3 minutes ago, DasValdez said:

One growing frustration is that we can't see part generation in Scrapyard (as discussed before)... it's getting difficult to find the higher gen reliable parts, or at least clean up the lower gen stuff. 

ScrapYard is 2nd on the list, after Monthly Budgets - don't worry I'll get to it

[Beetlecat]

On 12/22/2018 at 11:19 AM, DasValdez said:

Yes, 1.5.1, no intention of moving to 1.6 until 1.6.+, whatever that turns out to be. 

KSP has moved into the realm of how I update my computer OS, or get a new phone anymore -- *no way* do I run out and get the latest; I wait for everyone else to suffer the initial bugs and woes.

---This is largely due to the fact that Kopernicus rarely releases right when a new KSP version does, and I have a hard time *not* playing with OPM installed.

[strudo76]

11 hours ago, DasValdez said:

An elegant solution would be to allow "Preflight Inspection" to the flight scene PAW. Right click the part, and click a button to do a preflight inspection, which flips it to tested, gives it the reliability boost of a tested part, and incentivizes you to have an actual pre-launch checklist you need to run, ESPECIALLY focusing on newly built parts. Forget your preflight, or miss a part, and you run with the lower reliability. Also gives you a good reason to do a final check on part tweakables, since there's the weird issue where a part will lose it's settings sometimes if you replace it from Scrapyard inventory, or click "use new part". Chutes lose altitude config, engines lose throttle limit, etc.

 

11 hours ago, severedsolo said:

I'll take that one under advisement - Oh Scrap is kind of on the back burner at the moment, other mods need some serious attention., but I do like the idea

This could also be integrated with KCT, where pre-flight check would appear as an individual item what takes time to complete. That way it might lead to some interesting situations where you want to rush a launch by cancelling the pre-flight check item, and run the risk of a higher chance of certain parts failing.

[Beetlecat]

13 hours ago, strudo76 said:

This could also be integrated with KCT, where pre-flight check would appear as an individual item what takes time to complete. That way it might lead to some interesting situations where you want to rush a launch by cancelling the pre-flight check item, and run the risk of a higher chance of certain parts failing.

Oh, nice -- I rather like this idea. A bit of a way of "automating" the expenditure of time a little bit. I do enjoy setting up test flights and coordinating that in the longer-term goals, but it'd be cool to have a team of kerbals take over that for me in the background after a while.

[overkill13]

I'm looking to replace BARIS as my part failure mod.  I've had a lot of crashes, pun intended, of KSP while using it, but I have a few questions.

Does Oh Scrap! model failures during launch and while in space?  Does it perform background calculations for unloaded vessels?

[severedsolo]

1 minute ago, overkill13 said:

Does Oh Scrap! model failures during launch and while in space?

Yes - launch failures are weighted a little heavier than space failures because you spend less time doing them.

3 minutes ago, overkill13 said:

Does it perform background calculations for unloaded vessels? 

No, but its been requested, and I'm considering it.

[TGApples]

Am I correct in saying that damaged communication parts still can act as relays on inactive (ie. out of physics range) craft? It seems that way. I guess that's a bug?

[severedsolo]

Uh... yeah that sounds like a bug, thanks for the report - will look into it.

6 minutes ago, TGApples said:

Am I correct in saying that damaged communication parts still can act as relays on inactive (ie. out of physics range) craft? It seems that way. I guess that's a bug?

 

[DasValdez]

Ok, Status Update! 

1. It looks like craft recovery via KCT might not be incrementing part generations at all. Fly and recover a plane 5 times using the KCT Recover to SPH, go to retrofit it with new parts, and they come back as the same generation. It seems to only increment if you recover the craft via the stock "recover" function, then build it again using the same parts? @severedsolo can you give us any details on the logic by which generations are increased? Also, I'm curious as to whether or not parts which are launched but not recovered (e.g. expendable stages) should be incrementing. 
    
2. Still hanging in there with having to "test" newly built parts by rolling them to pad and then recovering them to scrapyard. Getting to be more and more annoying... the best fix we've been able to come up with is being able to right click parts on the flight scene and "run a preflight check" to get them past their "untested" state. Scenario we've been using a lot: Have a plane that's recovered and reused via KCT. Pre-build new "payload, which is "completed" but "not tested", and integrate with plane. IF you roll and recover to "test" payload, the plane gets penalized a use for all it's parts, which is especially impactful since plane parts only last 5 uses before they start to fail... only two payload flights per rebuild that way. We can deploy the payload to get the test done separately, but that it tedious and time consuming... not fun to do or watch.
    
3. Spamming on part failure rolls before launch. We're still encountering the part failure attempt spam when we deploy a plane to the runway. Scenario is we roll out a plane, Alt-F12 for the console, and see Oh Scrap spamming failure rolls twice per second. We can stop it by firing up the engines and getting the plane rolling. Can you help us understand why it would EVER/in what craft situation want to spam multiple rolls per second? Again, not sure if this is related to the "World Stabilizer" mod confusing the craft situation, but it isn't clear to me why Oh Scrap would ever roll twice per second.
    
4. MINOR: Still struggling with part inventory/Scrapyard not displaying or grouping by the generations. Manageable but the larger our inventory gets, the more of a pain it is to find usable parts. Would be great it the UI (and this might be scrapyard feedback) allowed us to group by gen AND reliability rating... sometimes I don't care the gen as long as it's high reliability

[scerion]

17 minutes ago, DasValdez said:

     3.  Spamming on part failure rolls before launch. We're still encountering the part failure attempt spam when we deploy a plane to the runway. Scenario is we roll out a plane, Alt-F12 for the console, and see Oh Scrap spamming failure rolls twice per second. We can stop it by firing up the engines and getting the plane rolling. Can you help us understand why it would EVER/in what craft situation want to spam multiple rolls per second? Again, not sure if this is related to the "World Stabilizer" mod confusing the craft situation, but it isn't clear to me why Oh Scrap would ever roll twice per second.

Clip with example of what the log looks like:

 

[jospanner]

Just a quick question, does this enable catastrophic failures of engines? *boom*