-
Posts
2,953 -
Joined
-
Last visited
Content Type
Profiles
Forums
Developer Articles
KSP2 Release Notes
Everything posted by magico13
-
The official update for 0.90 is version 1.5.2. KerbalStuff and Curse don't realize 0.90 has released yet, so the KSP versions are wrong on there. Here's a quick rundown of the changes (sorry, not a whole lot. I haven't gotten a chance to work on SR in a while): 1.5.2 - (12/15/2014) - Compatibility update for KSP 0.90 - Automatic recovery of launch clamps when they are unloaded. - Right clicking on a stage in the flight GUI will now delete it. - Added indicator to flight GUI showing which stage is selected. - Several bug fixes. - Contains a bug where kerbals will lose experience if they are in the craft when it's "recovered". Will be fixed soon. I also just realized that while I added an option to use a custom TWR requirement for powered recovery, I didn't actually make that do anything. Oops. I'll fix that at the same time I fix the above mentioned bug. This wasn't thoroughly tested, so there may be some bugs. Let me know how it goes! @MartGonzo That's not something I've seen happen before. Perhaps it was just a fluke error. Should it happen again, grab me the log file as well and maybe I can find something in there that might offer some explanation.
-
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Gonna release SR as official then, KCT will need some work that I might just save for after I get home tomorrow. 0.90 changed a non-trivial amount of things internally that has gummed up things. I'll keep you posted, but I might just play 0.90 tonight and have some fun (shock and awe!) -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Not seeing any issue here. In the editor, the SR icon runs the check on the editor vessel so you can see if it will be recovered. Worked fine for me in the SPH just now. -
I'm gonna try to get a basic "compatibility + things I've been working on" update out within the next few days, then I'm gonna start tackling the recent KSC changes of 0.25 and 0.90 (upgrades & repair). I definitely like the idea of tying into the upgradeable buildings, but I haven't decided to what extent. I could make it so each level of VAB has set rates (both speed and quantity, so a lvl 1 VAB has 1 rate at 1.0, a lvl 2 has 2 rates at 1.5 and 0.5, etc) or I could just set the maximums for rates based on upgrade lvl (lvl 1 = 1 rate w/ 1.0 BP/s max, lvl 2 = 2 rates w/ 2.0 BP/s max, lvl 3 = unlimited). I do like the idea of limiting the number of techs you can research at once, but I might use the Stock limits of what you can spend at once as the total amount of tech you can have researching at once. I'll have better ideas once I play around with 0.90 more, but am open to hearing what others think. I'll open a github issue for it: https://github.com/magico13/KCT/issues/25
-
KCT is currently INCOMPATIBLE with 0.90. I'm working on StageRecovery now, and will get to KCT soon. KCT looks like it'll need a lot of fixes unfortunately, so it may be a while, and the dev version has some things it needs sorted out before I feel comfortable with it anyway. It may take a few days for a full release! The dev version will be updated first, and when that's been tested I'll make it the full release. I'll try to get a dev update out tonight.
-
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Working on an SR update anyway, I'll get a KCT recompile later tonight. Definitely gonna have issues in KCT based on some of the changes I've seen. Will post link to updated SR for testing as well in a few minutes (up to half hour). Edit: Wait. Running into issues with kerbals losing experience and need to transfer some fixed from KCT. It might be a while. Development version of StageRecovery: https://dl.dropboxusercontent.com/u/25340912/StageRecovery.dll Changelog: 1.5.2 - (12/15/2014) - Compatibility update for KSP 0.90 - Automatic recovery of launch clamps when they are unloaded. - Right clicking on a stage in the flight GUI will now delete it. - Several bug fixes. - Contains a bug where kerbals will lose experience if they are in the craft when it's "recovered". Will be fixed soon. Edit: BTW, there's also a 0.90 version of KCT in existence now, ready for testing. -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Could someone please check StageRecovery and KCT with 0.90? I wasn't expecting the release today and won't be home to work on updating for up to several hours. -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
I haven't considered the uses of xp with StageRecovery. I'm curious as to what you were considering. Off the top of my head I could imagine greater returns for better pilots (kerbal or probe). -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Welp, looks like 0.90 will be out on Tuesday since they've released the NDA and are letting youtubers release videos, that means you're all gonna get roped into testing the 0.90 version of KCT after 0.90 releases. KCT 1.1 will only be fully released for 0.90 I guess Other than simple compatibility, I'm going to try to put time limits on upgrading and repairing buildings. If that's too difficult then I'll release 1.1 without it and add it into either 1.2 or 1.1.1. Squad, if you're listening I'd love to get some early access to 0.90 tomorrow (Monday) so I can make KCT available ASAP for the full release Otherwise, looks like we'll be playing catch up. -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
JefferyCor, if you could take a look at that for me just to confirm, but it wouldn't surprise me at all if that happened. I should be able to get back to dev work after Monday, but until then don't expect any further progress. Thanks for helping with the testing Juggernaut! -
Well, I'm going to be doing a revamp of the upgrade system somewhat soon (within the next couple months) so things will get shifted around a bit then, but for the most part I haven't felt the current settings to be that unbalanced, especially if you unlock the tech tree one "layer" at a time and purchase one tech upgrade each time you move up a level. As for the times, the very first build tends to be a bit high, but it very rapidly drops down to only take a couple days for new ships as you use and reuse parts. I even turn up the overall multiplier a bit because it's too low A suborbital flight should only be taking like 10-ish kerbin days to build. Even my Mun base only takes like 30. Also, being only suborbital with third level tech seems weird to me.
-
There's plenty of solutions, so whichever one I'm feeling when I go to fix it is likely the one I'll use Not simulating an uncontrolled craft is probably the easiest, but there might be that random time where you might want to. Or, since you can change craft during simulations, you might "simulate" a structural pylon so you can run tests on your lander that's orbiting Duna without using quicksaves. I might make that a viable option in a later update, where you don't have to simulate a particular craft and can just pay a base cost for the simulation time. Switching craft will probably also be an upgrade you have to purchase now that I think of it. I really want to get started on the new upgrades, but I need to sort out the last few remaining bugs with 1.1 and be done with finals (I have 2 on Monday, 3 hours each, and I'm pretty close to the border between passing and failing them. Grad school is hard compared to undergrad...)
-
I can see why this would occur. I'll have to add in some checks to avoid that. From my side, what's happening is that you are given the cost of the vessel when you simulate a craft (so that you can do it even with 0 funds) but KSP prevents the simulation from starting. Normally those funds are then taken back after you revert (I have to give you the funds before the persistence is saved, so KCT keeps track of the cost of the vessel and takes it back after revert), but since the simulation doesn't actually initiate it just keeps giving you funds. I can't think of an appropriate fix off the top of my head just yet, so I'll have to mull over some ideas. My first thought is to keep track of the total funds given and take all of that back on scene change, but then you could still build a ship you can't normally afford. I can add some more checks when you go to build a ship, but it feels like there should be a better way. I'll make a github issue to remind myself to fix this.
-
^This. Just go out, grab the data, and enter the command pod. Like magic all your data is in one easy to return place! Been that way since 0.23 I think. Works the same with crew reports, btw, which let's you grab more than one at a time (take crew report, eva, grab all data from capsule, reenter capsule, take another crew report somewhere else). Only condition is that you can't have two of the same exact data in the same pod (so you can't have two surface samples from the same biome, for instance).
-
Simulator facility
magico13 replied to GoSlash27's topic in KSP1 Suggestions & Development Discussion
Kerbal Construction Time has similar functionality (if you only want to use the simulation feature, just turn build times off in the settings). It's going to go through a bit of a revamp soon, but currently you can do simulations for a large range of time limits (with increasing costs), can start them at any time, and can start them either landed on Kerbin or in orbit around any planet/moon you've visited. When I rework them in the next update you will also be able to start simulations landed on any planet. Simulations cost funds based on the time, celestial body chosen, and cost of the vessel (smaller ones are cheaper), and anything that happens during a simulation doesn't get kept. The only limitations are the time limits and that you can't quicksave or quickload. -
Well it was a little bit of both. The inventory system was one of the very first things I created (it's present in Pre-Release 1) and was designed such that you could reuse (space)planes or other reusable craft without having to take the full time to build a new ship. The original version of KCT (when EkkuZakku was still working on it) didn't have any systems to reduce build times, so you'd have to pay the full time every single launch (and you could only have one ship building at a time and couldn't switch to a different craft, it was a good time ). It wasn't until recently when I was trying to add a way to recover craft directly into storage, thus bypassing the whole refurbishing and refueling stage and its associated time, that I discovered that there is virtually no simple way of turning an existing craft into a craft file. The conversion is definitely not as straightforward as it should be, but changing one or the other on Squad's side would likely break save games. I did find one way, and can get valid craft files, they're just pretty messed up. Which I might actually keep as-is (once I figure out the rotation issues) since it's a side effect of not going through the refurb phases that your wheels or solar panels are still broken and your tanks are empty. You can just use the edit function to fix those yourself, many of which won't require any extra game time other than what you take to manually fix things. I personally don't care so much for the "recover to storage" thing since the part inventory was designed to satisfy that requirement, but it was requested enough that I've spent quite a bit of time on it despite the many issues. The particular github issue for this is number 10 and you can view some of the progress I had on this by reading the posts on the dev forum (linked in the bottom of the first post). As I've mentioned before, it won't be a feature present in the "normal" 1.1 release since it's got a lot of issues (primarily with resetting things), but will be available if you have the "Debug" setting enabled in the KCT settings. If you (you being anyone reading this) wants to test it out now, the dev version is actually pretty stable otherwise (but we've only got one other person actually testing it out, so there's likely many bugs him and I can't anticipate that only more eyes will find). All the warnings on the dev thread about it breaking your save are there to ensure you know the risks before trying, especially because the next version uses a radically different save mechanism, but we're pretty sure old saves convert properly now. 1.1 required much, much save breaking to get to the current state (even between individual builds). For comparison, the update branch is 31 commits ahead of the current release (plus or minus one or two since the master branch has some very minor stuff not actually in the release) while the master branch only has 116 commits TOTAL since PR4 which was 8 months ago. That's over 25% of the total current commits all in one update, mostly because we had to keep retrying things. I don't know the overall additions/deletions until I merge it into the master branch. TL;DR: Rambling! I don't want to work on this paper about CuAAC click chemistry!
-
Nope, not a config setting. The EXTREMELY reduced build times are representative of refueling/repairing (default settings result in a build time of about 10% compared to a brand new ship). You can always increase the InventoryEffect (5% build times would be 400) if you feel it isn't reducing your times to something you would consider satisfactory. The development version (aka, the pretty much 100% stable except for some weirdness with Procedural Parts) has the ability to recover straight to the inventory without going through any build times first, but since KSP provides ZERO way to easily turn an existing vessel into a valid craft file (and straight copying it from the save doesn't work because suddenly everything is named differently or has a totally different meaning) it is currently pretty buggy (since I have to manually parse the ship and fix its rotation, remove science from containers, repack parachutes, turn off engines, etc). It works ok for rockets, but planes tend to get put tail down on the runway due to lost rotation information. And then you get to play with vessel recovery time instead, with a really wacked out frankenplane that may or may not be missing modules Squad doesn't make anything easy for us it seems. You also can do it manually by keeping the ship next to the runway and refuelling it with a tanker. I know of people who actually do that.
-
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
I'm wondering if the costs are perhaps being set to zero for the parts themselves. What's baffling me is that the debug info I put in wasn't being printed, despite the fact that you were running the latest version. I want to do some testing myself, which is why I haven't pushed this as the full release, but this is the last week of classes and I have finals next week, so I am currently very preoccupied -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
Kerbal Konstructs (and thus Kerbinside) are not currently explicitly supported, but they are also not incompatible (unless there was recently a change in how they worked so that they also use the launch button). To pick launch sites you use the button on the Stock toolbar. Whichever launch site is the active one when you press KCT's launch button is the one that will be used. Just go into the VAB/SPH, change the launch site to the one you want, then launch your craft. I plan on adding explicit support for KK later, but they made it so you can't open the launch selector from any view other than the editor, which will make it much harder. Also, disable reconditioning since it doesn't make sense with multiple launch sites. -
I actually was planning on moving away from the tech tree entirely. The original plan (many many months ago) was to tie upgrades to specific nodes of the tree, but with modded trees and other recent changes to the community that is becoming less feasible. It also would have made it so KCT didn't work in Sandbox mode at all. The new plan is my own tech tree for upgrades that cost funds/science/reputation to unlock, or upgrade points (or something similar) for sandbox mode. I haven't started work on that yet because I want to see what 0.90's upgradeable buildings will cover (which may render some upgrades that I had planned useless). The times don't quite sync up, EPL typically has smaller build times despite the fact that the KSC has actual facilities for constructing rockets. Also, since EPL times are based on mass and not cost (like KCT is) you'll never quite get them to line up. Otherwise I don't know of any incompatibilities. I obviously snipped that quote down a bit, but don't worry, I have a habit of rambling too Rollout times are implemented already in the dev build (which is actually quite stable now and only really needs some tweaks for procedural parts). As for customizability, I tend to prefer to give people the tools to play however they want. With the rollout times (which are treated the same way as reconditioning in that it's mass based, they're actually two instances of the same internal object in the code) there's a slider to let you choose how much time for a rocket should be spent on rollout and how much should be spent on reconditioning. Set the slider to 0% rollout and it acts like it does now, 100% rollout and there's no reconditioning (you just pay that time prior to launch), the default is 25% if I recall correctly, meaning if the total "reconditioning" time is the sum of the rollout and post-launch reconditioning then 25% of that total is rollout and 75% is reconditioning. There's also a setting to set a maximum reconditioning value (which actually sets the "total reconditioning" and thus affects rollout and post-launch reconditioning). Reconditioning was originally going to be based on total active rocket thrust at launch, but that scales with mass so I decided that just dealing with mass is simpler. With the recent changes I've made it would be very very difficult to move away from a purely mass based system (since I now need to calculate reconditioning times before the rocket is put on the pad). Short answer is that I plan on doing this. I haven't spent a whole lot of time trying to figure it all out, but it is definitely my intention to have upgrades and repair take time. Additionally, due to Real Solar System discussions, individual part unlocks will take time. As always, most of these things will be optional toggles.
-
In the development version I have actually already added some basic time requirements for recovering vessels! It's mass and distance based, with the minimum time (so, if on the launchpad/runway) being the same as the vessel rollout time (for hopefully obvious reasons), up to twice that time if it's on the other side of the planet. I might change that to some fixed speed like 1 second per 150 meters to simulate an airplane flying there and back at 300 meters per second (plus the rollout time). Currently it's just for when you recover a ship directly to the inventory, but I might see about also having it for normal part recovery (so you don't get parts or funds until the time is up), but I'd have to try to do that without messing up other mods. Since recovering straight to inventory won't be included in the next version by default (but will be enabled if you've got debug mode active) there's a chance you won't see this.
-
Apparently the log only outputs the current locks when entering the editor, not when switching to the space center. The only out of the ordinary one I can find is conveniently named "202025866" which is absolutely useless (meaning it definitely could be Stock ) as opposed to KCT's nicely named "KCTKSCLock" and other "KCT${function}" locks. It only appeared active when editing a vessel the third time and had quite a few restrictions it appears based on the bitmask. bitmask with weird lock 101111101111111111100000000000000000010000000000 bitmask with just KCT 100000000001111000000000000000000000000000000000
-
I haven't had a chance yet to take a look at those logs, but will during work tomorrow. I suspect an InputLock is active that shouldn't be, likely in another mod and not KCT, but because KCT sometimes has abnormal scene changes some mods won't release locks when they should. Mechjeb used to do this. The log might tell me which locks are active during the scene change and I can hopefully pinpoint the mod that enables it. Next time it happens, open the debug menu (alt-f12) and press the "Show input locks" button. Take a screenshot of that if there's some listed and send that my way. Be very careful to make sure that when clicking a KCT window button (such as the save edits button) there are no other mod windows behind the button (even better, close all other mod windows). That was the issue with mechjeb in the past. If I find which mod has the issue I can send them a report and suggest that they clear locks when the scene changes away from the editor. It certainly is possible that KCT is at fault alone though. Such is the way of software development: crazy annoying bugs I swear half of the 30 or so commits made for 1.1 so far were just to fix bugs and then caused more unexpected bugs!
-
Pre-emptive support for Linux selection.
magico13 replied to Hannu's topic in KSP1 Technical Support (PC, modded installs)
Ubuntu is typically the friendliest to newcomers. It would be my first suggestion but I'm also partial to anything Debian based (which includes Ubuntu). My second suggestion would be Linux Mint. I've not had as much success running KSP on linux as I have with windows, but I also haven't tried all that hard. -
Kerbal Construction Time/StageRecovery Dev Thread
magico13 replied to magico13's topic in KSP1 Mod Development
I didn't try actually launching, just putting a pod with a RC parachute on it on the pad (and recovering through flight, the space center, and the tracking station) and was still getting parts correctly. Looking at the log it appeared to only be procedural parts for you for some reason, so I added a bit of debug info to see if the costs are being calculated correctly. I also probably made the editor display the amount of PP currently in use more properly. Otherwise not much else in build 25. It seems to be getting fairly stable, knock on wood...