egg changed the topic of #principia to: Logs: https://esper.irclog.whitequark.org/principia | <scott_manley> anyone that doubts the wisdom of retrograde bop needs to get the hell out | https://xkcd.com/323/ | <egg> calculating the influence of lamont on Pluto is a bit silly…
<queqiao-> ⟨sichelgaita⟩ A lot of our crashes are actually suicide, so in that case we would know.
<queqiao-> ⟨egg⟩ Though they are very deep. Maybe we can do something in ProcessExit?
Jesin has quit [Ping timeout: 189 seconds]
<queqiao-> ⟨sichelgaita⟩ The above suggestions are valid, and I am sure that technical solutions could be found, but as I see it it's an opportunity cost: should we perfect crash reporting or implement feature X that could benefit users? I must confess that I am not particularly excited by the perspective of spending a month or two doing boring, complicated work on this topic.
<queqiao-> ⟨Al2Me6⟩ I’d argue that a basic crash popup is desirable, simply so that users actually know what caused a mysterious crash.
<queqiao-> I’ve played enough to recognize what’s a Principia crash, new users can’t.
<queqiao-> ⟨siimav⟩ > should we perfect crash reportingCertainly not but something as easy as opening the browser with some Principia wiki page or showing a simple dialog would be a substantial improvement.
JanskyRaptor is now known as UmbralRaptor
<queqiao-> ⟨sichelgaita⟩ This is where the First Axiom of Software Engineering ("Nothing is simple") kicks in. The crashes come from deep within C++ libraries that we don't control, and which may call "_exit" or "abort" depending on the phase of the Moon. Catching these is not easy (signal handling, anyone?) and opening the browser from within C++ (not to mention a signal handler) doesn't look trivial to me.
<queqiao-> ⟨lpg⟩ I thought we were (siimav was) talking about the classes of crashes that are triggered by a CHECK() or similar, I would that thought that was completely under your control
<queqiao-> ⟨lpg⟩ * have
<queqiao-> ⟨lpg⟩ * class
<queqiao-> ⟨lpg⟩ (I was exclusively talking about the save-is-corrupted case, where a checksum check would exist in a very different place than the current location corruption is detected)
<queqiao-> ⟨egg⟩ ⟪lpg⟫ I thought we were (siimav was) talking […] ⮪ CHECK is from glog.
<queqiao-> ⟨Transparent_Elemental⟩ thank god I don't have to think too much about the low level stuff like signal handling anymore, that was probably the hardest part about my programmer life and I ain't coming back to it, it's almost as managing time zones
UmbralRaptor has quit [Quit: leaving]
<queqiao-> ⟨siimav⟩ Pff, time zones are piece of cake compared to calendars
<queqiao-> ⟨egg⟩ _grins and waves the French Republican calendar around_
<queqiao-> ⟨lpg⟩ you say that like calendars and timezones don't overlap
<queqiao-> ⟨egg⟩ _grins and waves the French Republican […] ⮪ (which has leap days based on astronomical observations)
raptop has joined #principia
<queqiao-> ⟨rnlahaye⟩ ⟪egg⟫ CHECK is from glog. ⮪ GLOG has "InstallFailureFunction" which might be useful here
<queqiao-> ⟨egg⟩ ⟪rnlahaye⟫ GLOG has `InstallFailureFunction` which […] ⮪ Yeah, we spotted that; but from C++ doing anything nice like spawning a browser is a mess, so you would have to leave a turd under a rock in that failure function, and then figure out a way to do signal handling from C# to pick it back up in a way that is compatible with either _exit or abort, cross-platform.
<queqiao-> ⟨Alex6511⟩ Opening a browser from C# is non-trivial if you're not working within some sort of cross-platform framework that doesn't have such a feature already implemented.
<queqiao-> ⟨Alex6511⟩ I'm not sure what KSP's dependency stack looks like, you might be able to leverage something KSP is already using to do it (given KSP itself can open browser windows, I assume it should be possible)
<queqiao-> ⟨Alex6511⟩ If you're looking for a centralized way to collect information on crashes, that is not too difficult in dotnet world though. Like if your C++ code is returning an exception and you want to capture that, there's a library made by Microsoft that does that.
<queqiao-> ⟨rnlahaye⟩ "std::system" is able to open a browser on macOS using the "open" command. Windows seems to have an equivalent command ("start"), but Linux does not.
<queqiao-> ⟨Alex6511⟩ Yep, pair for the course haha
<queqiao-> ⟨Alex6511⟩ This is why I stay in C# land.
<queqiao-> ⟨Alex6511⟩ * ~C#~ dotnet
Jesin has joined #principia
<queqiao-> ⟨Paculino (ŝi/ri/she/they)⟩ ⟪egg⟫ _grins and waves the French Republican […] ⮪ That is a really cool calendar though. Have you ever done anything with it? You have mentioned it before.
<queqiao-> It is currently 29 Pluviôse by all methods except the 4-128 rule (in certain time zones)
<queqiao-> ⟨egg⟩ yeah, for year CCXXX, the Romme system happens to agree with the actual equinox; but that is not the case for year CCXXXI.
<queqiao-> ⟨egg⟩ The IMCCE mentions when the year starts in its calendar: https://www.imcce.fr/newsletter/medias/2022/01/docs/Calendrier_Paris_2022.pdf, bottom of last page.
<queqiao-> ⟨Paculino (ŝi/ri/she/they)⟩ The initial legislation saying both that the leap years would be every fourth year and that they would be scheduled according to the autumnal equinox as observed in Paris is bothersome. I wish it was in continued use past year 20 so we could see what method they would have switched to (or not switched to).
<raptop> 💭April 1st release that lets you set dates in the Republican Calendar
Jesin has quit [Ping timeout: 192 seconds]
Jesin has joined #principia
_whitelogger has joined #principia
_whitelogger has joined #principia
<queqiao-> ⟨Prophet1313⟩ Hey, I was redirected here from #ro-support about an issue that appears to be principia related. After updating to the new version the game started up and ran normally once. After about an hour it crashed and since then it hasn't been able to load the save. The logs from /glog/Principia are attached bellow. Thanks in advance for any help
<queqiao-> ⟨von Kerman⟩ Looks familiar 😦
<queqiao-> ⟨Al2Me6⟩ (Your save is likely corrupt.)
<queqiao-> ⟨von Kerman⟩ The question is, when, where, how and why the corruption happens
<queqiao-> ⟨von Kerman⟩ Because in my case, it actually happened more than once before I posted the log
<queqiao-> ⟨Prophet1313⟩ Im afraid I cant offer any info there. Needed to hop off for 20 minutes so I just left the game running and unpaused
<queqiao-> ⟨siimav⟩ Is only the Principia PluginAdapter section corrupt or test of it as well?
<queqiao-> ⟨siimav⟩ * rest
<queqiao-> ⟨von Kerman⟩ The rest doesn't look corrupt, but no way to tell for sure
<queqiao-> ⟨von Kerman⟩ I wonder though - is it possible to load a principia save in an install without principia 🤔
<queqiao-> ⟨siimav⟩ Yes
<queqiao-> ⟨von Kerman⟩ I might try that
<queqiao-> ⟨Prophet1313⟩ I also tried removing the craft that was loaded before it crashed from the persistent.sfs but that didnt help
<queqiao-> ⟨siimav⟩ Alternatively delete the entire pluginadapter section
<queqiao-> ⟨Prophet1313⟩ ⟪von Kerman⟫ I wonder though - is it possible to […] ⮪ One sec, ill try that
<queqiao-> ⟨Prophet1313⟩ Also unrelated since I read something in the messages above, my save is 1.2 gb, is that not supposed to happen?
<queqiao-> ⟨von Kerman⟩ 1.2 GB?
<queqiao-> ⟨von Kerman⟩ How long does a scene change take???
<queqiao-> ⟨Prophet1313⟩ 30 sec maybe?
<queqiao-> ⟨von Kerman⟩ Wait, do you mean the entire save folder or just persistent.sfs?
<queqiao-> ⟨Prophet1313⟩ Save folder
<queqiao-> ⟨von Kerman⟩ ok, this doesn't matter, it contains a lot of backups. Only persistent.sfs matters
<queqiao-> ⟨Prophet1313⟩ Aha, right. Its 120mb so I'm assuming its fine 😆
<queqiao-> ⟨von Kerman⟩ I ... don't think so
<queqiao-> ⟨von Kerman⟩ But there doesn't seem to be a direct connection between savefile size and crash. Mine was 67Mb when the crash happened.
<queqiao-> ⟨sichelgaita⟩ I must confess that I find it mildly concerning that we have two save corruptions in 24 hours. I am mildly concerned.
<queqiao-> ⟨von Kerman⟩ My save corruption actually happened more than once. But I'm unable to reproduce it at will. Just tried with the previous backup before the crash, with no success.
<queqiao-> ⟨Prophet1313⟩ Update, game loaded normally with principia removed and other than once solar probe everything is more or less where I left it
<queqiao-> ⟨Prophet1313⟩ Also unrelated I can't remember the last time the game ran at above 30 fps xd
<queqiao-> ⟨von Kerman⟩ Same here (the crashing save loads fine without principia in the install)
<queqiao-> ⟨Prophet1313⟩ Just for the sake of completeness, Im now restarting the game with principia back on and I'll try loading the save
raptop has quit [*.net *.split]
Jesin has quit [*.net *.split]
bees has quit [*.net *.split]
bees has joined #principia
raptop has joined #principia
Jesin has joined #principia
Jesin has quit [*.net *.split]
raptop has quit [*.net *.split]
Jesin has joined #principia
raptop has joined #principia
<queqiao-> ⟨Prophet1313⟩ And nope, still crashing with both the new and old versions
<queqiao-> ⟨Prophet1313⟩ What I found interesting is that one probe in heliocentric orbit was knocked WAY out of plane when I tried loading with principia off. Everything else was where I left it except that probe that got kicked up to almost 40 deg relative above the plane of the Solar system. When I last checked on it it was on the way to Mars and very much in plane with the rest. Hopefully that helps, I'll keep the install if the logs can be of any...
<queqiao-> ... help
<queqiao-> ⟨von Kerman⟩ Relocated probe in heliocentric orbit is normal and expected when you add or remove principia mid-game
<_whitenotifier-1f03> [Principia] columbus15 opened issue #3332: Game crashes when opening map view from vessels orbiting Minmus, or clicking on them in the tracking station - https://github.com/mockingbirdnest/Principia/issues/3332
raptop has quit [Ping timeout: 189 seconds]
raptop has joined #principia
<_whitenotifier-1f03> [Principia] pleroy commented on issue #3332: Game crashes when opening map view from vessels orbiting Minmus, or clicking on them in the tracking station - https://github.com/mockingbirdnest/Principia/issues/3332#issuecomment-1043435719