Ruedii

This is listed as only compatable for 1.3.x has anyone tested it for 1.4 or is that just because it's untested?

Still not working on Linux. Different errors this time. I posted a new bug report because of drastically different errors. Feel free to link the two errors, of course. Specifically, It's trying to call libcurl despite the curl code being removed.

This is more a me-too than a fix. Yes, I experienced this bug under Linux as well. It's also appeared in several other releases. I suspect it's a common layer occlusion bug in either the menu renderer or the display. I don't know if there is a work around but it usually is fixed almost immediately. If you want, you can go search for the bug on the bug tracker and mark it as having a regression in 1.4.

Those ones are rebalanced for SETI though, I'm talking about the originals, only with meshswitch added.

The partoverhaul ones released slightly before Porkjet left Squad. I happened to find a modified version that adds meshswitch: https://github.com/Angel-125/KSPRocketPartOverhaul/releases

I'll put it on my todo list. However, my todo list is long. I will write up a blueprint so whomever gets the contributions in first can put the individual pull requests for each engine immediately. The other feature request is adding the upgrade function. (Across the board.) Also, if license permits, we need to see about getting the porkjet parts onto this pack set.

Considering the availability of coreheat function and radiators now, should many of the atomic engines be updated to utilize coreheat and have more overheat tendencies? Of course the RTG-Based atomic engines should have built in radiators sufficient for their purpose unless various conditions prevent proper usage of said radiators.

If you really want to keep players happy when making expansions, release free promo goodies whenever you release an expansions, and make sure expansions are big complete deals with new types of gameplay. (This one clearly qualifies with an nice easy mission builder.) If you could, make a mission exchange, preferably allow 4 methods of sharing missions: Steam Workshop, CurseForge, Dropbox and "export to file." Also allow mods to add additional sharing methods. Any method that can easily be done in one simple click once the login information is loaded, should be done as such, sharing to all methods configured in one simple step. Make sure exported missions are human readable and human editable with some basic programming/configuration experience. (Either use Javascript Block Format or an XML format, and make sure it's easy to understand and fault tolerant. This means making a lint script and quirks script for the loader to attempt to correct common editing errors.)

No, I'm saying the engineering IS capable of it according to my research. The SoC board in the switch (Tegra X1) is engineered to handle all 8 cores at 2.1Ghz However, the Switch has an insufficient power supply or cooling to run them at over 1Ghz. This is different than hyperthreading. It is basically closer to processor scaling. Simply put, it's like if you stuck a 150W TDP 8-core CPU when you have a power supply and heat sink that can only handle a 65W TDP CPU. You would simply need to reclock the CPU for it to produce less heat and draw less power. Basically, the CPU can handle 2.1Ghz clock speed, the motherboard just can't provide enough power. There are two ways to handle this. The first is to underclock the clock speed, the second is to turn off cores. Nintendo chose to underclock the clock speed. I'm saying they should give game developers the option of which to do. Many cell phones have the same issue and use the same workaround I said they should use. The other technique that can be used is to increase the speed for very short durations. Some chips have mechanism to do this internally either manually triggered or automatically triggered by detecting a full command queue with no power-save idle commands, a sufficiently low heat value, and sufficiently high voltage from the power supply. It is commonly called "Boost" or "Turbo" function and often runs even higher than the default maximum processor speed. In recent years, Intel has made their chips handle power saving governors entirely using this mechanism, but most SoCs still use manual firmware-based manipulation. As a note, the switch only downclocks it's GPU when off base, not the CPU. Considering the base output is 1080p and the device screen is 720p30, this is understandable.

The problem isn't so much that the Switch is weak, so much as that it's power is spread over 8 cores. The upstream platform for Kerbal Space Program, on the other hand, is severely limited in multi-threading capability (as are many other game dev kits.) This has been the limiting factor in any third party devs getting decent performance out of the Switch. The projects necessary to fix all the world thread locks on Unity3D have been stalled on their timeline since before Kerbal Space Program started, with only minimal progress. Unfortunately, the KSP Devs can't switch horses mid-race hear, and are thus pretty much stuck with Unity3D. Personally, I wish that Nintendo had allowed the switch's CPU to operate in either 4 core 2Ghz or 8 core 1Ghz mode. This would have retained reasonable power requirements while providing more processing power to low-thread-count games.

Unity already exists for the Switch. As of the part cap, it's more likely to be around 256, 512, or 1024, not 100. If you are having trouble building things under 100 parts you really need to rethink your designs. I can get a basic satellite into orbit with 10 parts, including the satellite itself and the fairing. (Satalite is 6 parts, decoupler is one, fairing is one, fuel tank is one and lifter engine is one.) As of the myth that the X-Box and PS4 contain "PC Hardware." The X-Box One and PS4 are not what could be described as standard PC hardware. They are X86-64 embedded hardware. Their hardware does not have a PC Standard OS or BIOS, and they have limited subset rendering libraries. Porting a game to them is not simply copying it over to the disk and reducing the rendering quality. They also do not have many of the standard hardware and firmware functionality elements that a standard PC has. (X-Box One being a variant of Windows Mobile with a limited subset of Direct3D 12, removing most legacy extensions and the PS4 being a variant of FreeBSD with a subset of OpenGL, with many extensions removed, and others mandating to be thread-safe when the GL standard says otherwise. Unity does not support Windows Mobile or FreeBSD.) Neither of them have the standard APIC subsystem or ACPI firmware setup required for any modern PC OS. They also lack any form of BIOS, instead relying on a thin boot-loader closer to what is found in your smart phone. Additionally, it is worth noting that Kerbal Space Program is not written in system specific code. That is the whole point of using a platform such as Unity.

Yes, they would have to cap at 30fps, and likely be a part cap per vessel. Texture quality might not be too much of an issue considering the fact that there are 4GB of RAM, and it should be able to run with GL-ES 3.x and ASTC texture compression. This means far lower texture overhead than D3D. Direct CPU access to the texture memory also reduces texture memory usage further. The use of flash memory for storage should reduce problems with processor time in ground quad loading ground height-map through the use of precomputed cache. This comes into a bigger advantage if sparse texture support is added. It is notable that the processor on the switch gets more instructions per cycle than your average PC processor. I am not sure if this is usable, though, considering the limitations of the Unity engine.

Waypoint manager is causing substantial framerate loss when it's window is open. Is it still using the old Legacy UI or something? If not is it set to update every frame, as it should be set to update on a timer instead (5 times a second should be sufficient). There are also a handful of other ways UI can be adjusted to reduce load. You can look into those.

Yeah, I've found normal difficulty is too hard and easy difficulty too easy. I might want to add that feedback to the bug tracker.

No, I mean I want objects to be able to be mirrored BOTH on the vertical axis (ventral to dorsal) AND the horizontal axis (starboard to port), in a vertical, horizontal or BOTH.

Could we get vertical symmetry and horizontal+vertical symmetry as well as the horizontal symmetry?

Thanks anyway. It's good to get a reply either way.

Using 1.22.6 on Linux it seems to work now. It may have been an issue with my previous setup. I've replaced my entire install and am using a clean user directory as well. I'll get back and report any errors if DNS problems star occurring again.

Technically you just use FireSpitter fuel switch and texture switch on one, and hide the other (not disable it entirely, so pre-mod crafts work.). Intersteller partswitch would allow you to actually combine a reasonably large number of parts into one if you want to switch to using intersteller partswitch instead of firespitter. (For instance you could combine all the 1.25m to Mk2 adapters into a single object and even add more utilizing single dimension scaling for the lengthwise dimension.)

Any chance you could integrate the fuel tank parts on the Mk2 and Mk3 to be a single part with fuelswitch?

On my DNS issues with CKAN, it might be something in my system settings. My system is all full of cruft and lint from 4 years of upgrades through Ubuntu revisions, so I'm going to be doing a clean reinstall soon. I finally have real motivation to do it, since the latest Mesa install fixes the problems I have in ARK: Survival Evolved (should for other Linux Radeon users out there.)

Could we get biome specific bases on celestial bodies (one for each biome) as well as having one station in high orbit another in low orbit?

There should be a function built in to CKAN for recovery, that basically deletes and reinstalls everything but your installed mod list. As a note on my own DNS issues. It seems to work fine with the KSP_CKAN_USE_CURL="1" variable so my issue is clearly an issue with either the new network code.

I was wondering if this could be modified to work for making non-ablative heat shield. Idealy make 3 varieties. High-temp conductive (High heat capacity coefficient, High heat tolerance, high conductivity, high durability) High-Temp insulator (Normal heat capacity coefficient, High heat tolerance, low conductivity, lightweight, fragile), and normal (Normal heat capacity coefficient, normal heat tolerance, normal conductivity, high durability, lighter and cheaper.) lightweight (Low heat capacity coefficient, low heat tolerance, low durability, ultralight, cheap.)

You could simply retry by parsing the error and running it again after a delay if it is on a list of errors that might be resolved with a retry. I like the idea of running the DNS resolver first. The DNS resolver routine should have a longer timeout.