Welcome to MUGS ⚄♠♞🏹 (Multi-User Gaming Services)! | github.com/Raku-MUGS | v0.1.1 has been released! (github.com/Raku-MUGS/MUGS/blob/mai...v0.1.1.md)
Set by japhb on 16 April 2021.
raydiak and yeah I noticed development on it lost some steam, kinda figured that's where you were at. I know how it goes, that giant burst of energy in the beginning and the later deflated feeling when you put in a bunch of work and nobody's using it. that's also what I'm hoping to help with as I implement this web game: something really in-your-face shiny and exciting and compelling that attracts attention 00:09
it's even got music and sounds and stuff, will have animated effects and/or a particle system, that kind of stuff. even if a lot of the client-side glitz isn't from MUGS itself just a part of that game's UI, it says "hey, look! you can make amazing stuff on top of MUGS!" 00:10
and hopefully also be fun enough to attract people to installing MUGS just to play the game. which is sometimes also a good way to get some of them proving they can do it better than you did, leading to more developers :) 00:13
japhb Oooh, sneaky 00:15
Actually, the biggest delay has been that I couldn't release any of the new MUGS work because it depends on CBOR serialization, which depends on Cro::CBOR, which depends on the just-released Cro (I'd been running a patched version for a month) 00:16
Looks like the build-from-scratch worked, now to start bumping and pushing. 00:17
raydiak cool deal, I'll try installing it later. afk for now, nice chatting! 00:47
japhb You too! :-) 00:48
04:15 japhb left 04:17 japhb joined
japhb OK, today I managed to get CBOR::Simple 0.0.8 and Cro::CBOR 0.0.2 released, both depending on the recently-released Cro 0.8.5. This unblocks lots of different performance improvements, including CBOR auto-serialization for Cro. 08:33
I also got a lot of release prep for MUGS 0.1.2 done, but still a bit left to do before 0.1.2 release.
I added a bunch of docs to github.com/Raku-MUGS/MUGS/tree/main/docs including several in the design/ subtree: 08:35
* Add two "Life of a ..." docs from local tree (both need updates)
* Add Game Classes orientation doc
* Add Data Validation design doc
* Add Supporting "Bullet Hell" Games rumination doc
* Add Logo Ideas historical doc
raydiak japhb: the first time I tried mugs-cli, at the "Select a game type" prompt, it tooks a very long time before it accepted input. Eventually it did respond to keystrokes, and this is what happened: gist.github.com/raydiak/6b6d67a98a...d4336de81b 10:59
then I tried a second time, figured I'd try a different game, and it's been sitting at that prompt for several minutes now, still not responding to keypresses (I even tried just typing something and hitting enter in case it was just a terminal echo issue, but no response). ctrl+c doesn't interrupt it and ctrl+z doesn't suspend it. no cpu usage or disk activity, just totally unresponsive 11:04
this one isn't your issue, but see github.com/croservices/cro-tls/pull/7 . if anyone else asks you about a similar error while installing, they need to manually upgrade Cro::Core before trying to install MUGS 11:06
mugs-web-simple starts up fine, but trying to access it gives gist.github.com/raydiak/9d5fbcc568...a635e333ef 11:10
also, firefox dev tools say your cookies soon won't work: "Cookie “k6sgTIZhrjsfQa586Z4PUO375vF7gvFNFchoSvZgjno8ZcEn4NnryaieteBI6UHf” will be soon rejected because it has the “SameSite” attribute set to “None” or an invalid value, without the “secure” attribute. To know more about the “SameSite“ attribute, read developer.mozilla.org/docs/Web/HTT...e/SameSite 11:11
localhost:20000 11:12
oops, didn't mean the multiline paste. you get the idea anyway
also warns about the lack of a character encoding in the response headers, but idk if that's just in the 500 response which is all I get, or all responses 11:15
I understand you're already in bed, just assuming you'll see these messages tomorrow 11:17
japhb Wow, OK, lots of great feedback. Give me a bit to chew through it all. :-) 17:52
raydiak, what MUGS version are you using? Are you installing from fez (and thus getting 0.1.1) or are you installing from GitHub URLs or checkouts (in which case you should be getting HEAD)? 17:55
That first error looks like one I'd already fixed ... 17:56
Hmmm, OK, from that cro-tls problem, looks like MUGS-Core needs to explicitly depend on all the Cro versions for now. 17:58
OK, depends fix pushed 18:01
lizmat japhb : well, that's a good policy anyway 18:02
even inside the code, if I can help it, not just in the META
because if you only put it in the META, and something else installs a higher version, your program will pick that one, instead of the one you're sure of will work 18:03
japhb lizmat: I was already depending on the leaf versions, but github.com/croservices/release/issues/2 indicates that (for now) I need to depend on all the versions in the transitive dependency chain
lizmat: Yeah, I've been thinking about that. For Cro I don't think the API is continuing to change, but it would be good if the community moved forward to depending on api rather than ver (except for "at least" versions) 18:04
I'm not sure mi6 is doing the right thing with api declarations yet, though. 18:05
lizmat well, even without api changes, things could work (probably unintended) but subtly different
japhb lizmat: True. It's a tradeoff of allowing people to pick up bug fixes quickly versus locking versions hard. Right now I'm moving to depending on :ver<major.minor.rev+>, which works for projects that are stable enough to do semantic versioning. Not so much for early-times projects. :-/ 18:07
raydiak: Were you using HTTPS or HTTP (TLS or plain TCP) to access the WebSimple UI? 18:11
raydiak japhb: ah I see. I just naively did a zef install MUGS, so I got 0.1.1. didn't realize that wouldn't give me the changes you made yesterday w/the new Cro version 18:19
and I was using regular http. explicitly using https gives me "PR_END_OF_FILE_ERROR"
that's the browser error. the server error is "Malformed request line", I assume you just don't support https 18:25
japhb raydiak: Yeah, I wasn't able to get all of the stuff ready for 0.1.2 last night, and I only upload full releases to fez/zef because I need to be able to specify strict version dependencies for now. I'll announce it here (and you'll see it in the topic! :-) ) when I've got everything released again. 19:39
I generally use TLS between client and mugs-ws-server, because I have sufficient debugging output to be able to see if something goes wrong there -- and I can simply have the client trust the server's certs, no biggie. 19:40
I support either HTTP or HTTPS from the browser to the mugs-web-simple server, though I have been in the habit of using HTTP because of constantly changing browser restrictions on cert trust. Still, that means I might have broken the HTTPS WebSimple support at some point, bah. 19:41
But ... the big thing here is I think session cookies should be SameSite=Lax explicitly, not SameSite=None. 19:44
That's on the Cro side, but I can look into a PR for that.
OH! I forgot to mention: mugs-web-simple only runs one protocol at a time right now. In other words, if you want to turn on HTTPS, you need to do `mugs-web-simple --secure` 19:46
(--secure is the default for mugs-ws-server already)
raydiak japhb: sorry I had to go for a while. I'm trying reinstalling everything from git clones, so I can at least give you more current feedback 23:36
japhb raydiak: Spelunking into the Cro::HTTP::Session::* code, it looks like session cookies are set without any SameSite at all. Could probably upgrade that, since the actual Cro::HTTP::Cookie code does understand SameSite. 23:37
raydiak: Ah, thank you!
raydiak I still get the same "Could not locate template 'logged-out-home.crotmp'" from mugs-web-simple 23:43
japhb raydiak: Figured that one out too. 23:44
Cro doesn't yet have templates-from-resources, though there's been discussion of it. 23:45
Right now it expects a directory to exist on disk ... which means running mugs-web-simple from the MUGS-UI-WebSimple checkout. :-/
raydiak ah, okay. I'll try that 23:46
japhb Until the ability to override the template loader is in place, I'll probably just have to doc that.
raydiak mugs-cli from HEAD still gives me the same unresponsiveness too
japhb WTH 23:47
That's just weird.
You're on Linux?
raydiak yeah, manjaro
japhb Hrmph.
So you actually get past "Loading MUGS." all the way to "Select a game type", and *then* it appears to hang? 23:48
raydiak yep
japhb suspects alien influences
raydiak known game types, no already active, all the way down to the prompt, but then nothing
japhb OH WAIT 23:49
What's your terminal?
(program and emulation)
raydiak bash in yakuake
japhb looks up yakuake ... 23:50
raydiak kde's konsole, but embedded in a guake/tilda-like dropdown 23:51
japhb Yeah, I just realized it may have weird behavior when I send it console size query strings 23:52
OK, do you have xterm or gnome-terminal on your system? I want to cut away some other possibilities. 23:53
Also it might be worth testing with plain old konsole as well, just in case all the problem is the roll-away behavior.
raydiak you're really grasping for straws, there. everything works exactly the same in it as konsole. I'll see about installing xterm, definitely not interested in pulling in the morass of gnome deps 23:55
doesn't work in konsole. does work in xterm. 23:57
japhb OK, that helps.
So konsole derivatives don't work, xterm does (and for me, gnome-terminal does) 23:58
Guess I get to learn more about terminal compatibility again. It must be a day ending in 'y' ....
raydiak it's not even a derivative, it's literally the konsole kpart just embedded in a different shaped window 23:59
are there some relevant env vars or something I can check for you?