magico13

April development live stream is going to start in about 10 minutes (at 12pm EDT) and will run for up to 3 hours depending on how I'm feeling (got a bit of a headache starting) and if anybody shows up As I've mentioned in the past, the development itself probably won't be super interesting (unless you like staring at shoddy code, in which case it will be VERY interesting ) but you can chat with me live about stuff (KSP related or not). This month I'm going to be focusing on ScrapYard and KCT. https://www.twitch.tv/themagico13 Edit: Stream finished! Thanks for stopping by!

Yeah, it's been an intermittent problem. With the latest build it's doing it on my end (and was also reported by someone else), but on previous builds it was fine. But that code hasn't been touched in those builds. That particular code has always been troublesome unfortunately. E: I just pushed an update that might help. When I was working on SimuLite I saw a similar issue with scene changes, where requesting a scene change would cause a crash to desktop. This code is a bit different, but still ultimately results in a scene change (I never actually request one, KSP does it itself). I applied the same fix and it seems to be working better for me. Try out build #16 or later if you get a chance.

It was previously reported and I saw it myself. Not sure when I'm going to get a chance to work on that, but for now that functionality is just broken, sorry

I just realized I said Saturday the 7th. That's not correct this year. I meant the 8th! It's tomorrow!

Just a heads up. Tomorrow (Saturday April 8th) [edit: accidentally said 7th] at 12pm EDT (16:00 UTC) I'm going to be streaming some dev work. The dev work itself probably won't be that exciting, but it's a chance to come chat with me live about stuff. I'm planning on working on some of the core stuff for ScrapYard, aka all the backend that will make it work but not yet a UI or integration with KCT. I've got other plans that popped up at 4pm so it'll probably just go between 12 and 3 pm. http://twitch.tv/themagico13

I did something like this back when the physics range was 2.5km instead of 22.5km, except instead of probes for science I was dropping RemoteTech relay stations halfway across Kerbin. Drop it close to the ground (like less than 500 meters, the lower the better) at as low a speed as possible (max 300 m/s), deploy chutes right away, then do a big circle (circle can be much, much bigger now than when I did it). Drop all your probes, then land the plane and visit each probe from the tracking station. Or use FMRS, drop the probes, then manually watch each one descend. Either way you've got to take the amount of time it takes to fly out there and the time for each probe to touch the ground. Even circling is fast in comparison to flying a plane out there each time and landing.

In the flight scene, while landed on Kerbin, open the KCT window and choose either the VAB or SPH. A button will be in the middle of the window saying "Recover To Storage" or something like that. What it does is takes the current vessel and puts it right back into the "warehouse" (after a short recovery delay). You can then edit it to refill the tanks and replace any missing/broken parts. It's mostly useful for planes since you avoid having to rebuild it with new parts. It's an error prone process since it takes a Vessel that's had all sorts of things happen to it, and tries to turn it back into a craft file that you can load in the editor. Sometimes it works. Sometimes it doesn't. Since 1.2 it's decided it wants to completely crash the game with no errors, but only sometimes and without warning. It's a feature I added after a lot of requests for it, but it's always been a giant pain. If/when KCT gets rewritten I might try to avoid it entirely by putting more value into the part inventory (such as making vessels that are 100% from the inventory take really small amounts of time instead of "mostly small" amounts).

This is an interesting mod. I haven't tried it out but I wish KCT could work with it (KCT's not written in a way that would allow it right now). StageRecovery could work with it though since StageRecovery's API announces when a vessel is recovered, which could allow this mod to consider those costs. As for your issue of loading vessels in orbit, have you looked at Contract Configurator's SpawnVessel code? It can take a ShipConstruct and put it basically anywhere. I'm using a slightly simplified version of it in a simulation mod called SimuLite and it's working really well, except for landed simulations where the orientation is messed up.

I've seen that too but couldn't tell if it was a result of changes I made or just KSP weirdness. I am not sure if I'll be able to fix it because it because it happens without any errors being logged and for no discernible reason... I'll see what I can come up with.

Looks like they're where the probe core is placed in the actual launches (so presumably on top of the probe core). Could it be because the stage is still firing? The one that didn't save is still shooting off somewhere and the stages that did save have the engine(s) deactivated. Just a guess based on the images.

Just modify it using the standard KSP modifier on Windows: Alt. Perfect solution, definitely won't cause problems

This is for an in-progress vessel? Ok, I didn't test that case. Thought I got everything but apparently something still slipped through. Now that I (presumably) know what's wrong I can probably knock this out really quick when I get home. Thanks for all your help, you've done a great job at testing this. Edit: Alright, it should force a refind of the KSC all the time now, which should fix this issue (in #15). The KSC still gets cached for calculating rates and such, so that'll help performance still, especially when KSC Switcher is installed. Edit 2: I've added ScrapYard to the OP. It is by no means ready for testing but is my primary focus for April. This thread should be used for discussion until a release thread is made (when it's been through a few betas). If you want to track progress you can do so by reading the github commits, or more easily by visiting the ScrapYard Trello board. This is the first time I've used Trello for a project. If it goes well I'll probably start using it more often. Also, based on what's currently on the board, assuming no delays and not a whole lot new added, I'm targeting a May 1st release of a "Minimum Viable Product" that'll be fully functional but missing a few features I consider important but not required for functionality (MVP features are red, "Strongly Desired" are orange)

Shouldn't be... Hmm. I'll have to experiment. My Linux Mint install is totally goofed up right now because Cinammon is crashing for no reason (on both of my computers no less) so I can't test Linux as easily right now. But I don't see why that would be the cause of this. I definitely saw it not removing the vessel on launch and then definitely did in the code that #13 is built on. Though I didn't try the actual build from the build server. I'll let you know what I figure out. Edit: what... Yeah, even the rollout issue is happening for me now. This is gonna be fun... Edit2: I bet I know why. I started caching the KSC so it doesn't have to be searched for every time (likely a good performance hit). Ah, the 2 main problems in programming: cache invalidation (this one), naming things, and off by one errors. Edit3: @Pand5461 try out build 14 and let me know if it's still happening. Tested both cases on my end and it looks cleared up. Based on the logs you provided that definitely was the problem (ironically because it didn't fail to remove the vessel from a build list like it normally does, which normally triggers a forced search in the current lists). Basically it was still holding a reference to the old KSC which no longer existed because of the scene change, so when KCT held onto a reference to the "old" ship it removed it from the list in the "old" KSC instead of removing the corresponding "new" ship from the "new" KSC. In other news, I really need to just rewrite everything because maintaining this mess is way harder than it needs to be...

I'll take a look tonight. Is the bug with the vessel still being in the list after launch still there? That one stopped for me in #13 but I forgot to test editing.

Not yet. It'll take a little while to finish partly because I'm not 100% sure how I want to handle some things. Once it's ready for testing then I'll make a post in the KCT/SR dev thread and if you want to help with testing then that would be appreciated. I started setting up a Trello last night for it that I'll make public when I've got it ready so that people can see what I plan on doing/how progress is going and comment on things.

I once made a mod that let you do this. You could bring the active vessel into the editor, edit it, and then it would be back wherever it was when you exit with all the changes made. It was for when you forget an antenna/solar panel or want to reduce part count and stability on a station after docking. It had tons of issues, especially with docked vessels. KSP doesn't really like to take a vessel and turn it back into a craft file, but it is possible to do. There's just a lot of holes for bugs to get into. I don't have any intention of trying to bring that back, it was going to be a lot of work to get it to something remotely stable. With that said, there's always hyperedit (or save file editing, or the debug menu) for those times when you really need to fix a stupid mistake. Or there's KIS and/or KAS if you want kerbals to have to take parts up to fix something (I've replaced antenna that way on more than one occasion...)

I'm taking a look at this now, thank you for the report (with video and log no less!). Small request, in the future can you upload the log to a site like dropbox or google drive? It makes it easier to download and run through my log managing tool. Also, the output_log.txt file from the KSP_x64_Data folder (you're on Linux so it might be somewhere else) is usually preferred, but these days either log is ultimately fine. KCT is reporting that it removed the ship, so now I just need to figure out why it's not actually doing that. Edit: Build #13 should fix that issue. I moved DelayedStart to just be called at the end of Start since Start was happening before DelayedStart since DelayedStart was being called by OnLoad, which used to happen after Start but now happens earlier. Woo.

Yes, StageRecovery v1.7.0 works with FMRS. If you don't have FMRS active then SR will attempt to recover it, and if you defer parachute-only stages to SR then it'll try to recover those. You can just increase the DR Max Velocity setting in the StageRecovery settings if you don't believe the default value of 2000 m/s is acceptable, or set it to 0 to disable that feature entirely. With the default value, stages going 2000 m/s at deletion start having a chance of burning up and at 3000 m/s have a 100% chance of not surviving. Adding a full heat shield shifts that to 3000 m/s to start having a chance and 4000 m/s to definitely burn up.

I mean, it's possible, but KCT also adds in a "Clear All" button so you can empty it with a single button press. If it's something you really want then I can see about adding it.

It's slow going and I haven't worked on it for a little while since I've been focused on other things, but it's going to be the main thing I focus on in April. I hope to have a beta with KCT support in the next few weeks but it depends on how many snags I run into. The KCT dev thread will likely talk about that more once I really get into it. It's currently at the "proof of concept complete" phase and I'm adding in things that KCT will need but the inventory itself doesn't (part tracking is one, letting mods trigger inventory processing rather than only at rollout), plus it still needs a UI, configuration, and lots of testing.

I can't fully respond right now, but the part inventory doesn't actually exist right now so procedural parts are ultimately treated the same as regular parts, they just go through different formulas. The part inventory is being rewritten from scratch and entire modules are stored with the parts so that one procedural part is completely different from another, though they can be considered as related for helping with the "build effect". Which modules are considered "different" is configurable by the user. Let me be clear that the variable [R] or whatever it ends up being won't be used by anything other than this extra integration mod, so there's no reason for it to exist within the core mod. Since the integrator will come with a required preset that includes the variable then if someone wants to use it they have to choose that preset. Otherwise the default preset just won't use it and nothing will change.

Well, be prepared to take it as a definite "yes" since I just got it working in the latest KCT dev build. Best part is that it isn't specific to BetterCrewAssignment, it works with manual assignments as well Also, semi-unrelated, but normally I have to fight with the constraints of Squad's code. In this case it was entirely fighting with my own disheveled mess of code that I wrote 3 years ago when I knew nothing about proper development. I need to just rewrite everything because it was/is a mess.

Speaking of crew assignment... The latest build (#11) now remembers your crew settings in the editor and uses those as the defaults when you assign crew through KCT. It isn't fully tested, so that's where I need your help! This means that if you put your scientists in your science labs and your pilots in your command modules, they should still be there when you go to launch. If they were sent off on a different ship between the build's start and its launch then that kerbal's seat will just be empty by default. If all crew are gone then it'll fall back to the regular method of throwing Jeb into the first seat he can find. If you want those assignments of scientists and such to be automatic, try out BetterCrewAssignment. This update was made with that mod specifically in mind, so I'd appreciate it if some people could test it. Thanks!

I haven't actually tested this mod out yet, but does it modify the CrewAssignmentDialog directly and before the launch button is pressed? If so then what I'm working on right now should work. If not then we can probably just talk through the craft file (is it just a module? That's easy, peasy, just go through the ConfigNode)

A combination of not a whole lot of time to work on it and a lack of desire to work on modding a game I haven't played for more than 15 minutes at a time in months/years (0.90 was the last time I really played), plus sizeable changes required due to changes added in 1.2 that knocked out some core functionality. The dev version hasn't gotten to a point where I am comfortable with releasing it as a proper release and I currently prefer the smaller support requirements of an open beta with fewer players. I've only very recently come back to modding and have been working on a handful of different things. KCT and the part inventory are going to be my focus this coming month. Grab the dev builds from the dev forum thread that @TheRagingIrishman linked (thanks! :)), from the build server. Those are made every time I commit to github, but note that a successful build doesn't mean it won't break your save. I've been tinkering around with the positioning of some code since the ordering of some functions (which KCT relied on happening in that order) got changed all of a sudden and broke things in 1.2 and I'm still figuring out what needs to happen where/when (edit: I mention this because that's one of the types of things that could unintentionally break things).