[01:52:54] * bitmapper [bitmapper!~bitmapper@142.162.167.202] has quit (Ping timeout: 240 seconds). [08:43:23] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has quit (Read error: Connection reset by peer). [11:55:11] * ColonelJ [ColonelJ!~ColonelJ@unaffiliated/colonelj] has joined the channel. [15:23:41] * bitmapper [bitmapper!~bitmapper@142.162.167.202] has joined the channel. [17:03:10] so I found a way to freeze factor from git [17:04:11] I'm playing around with version management-y things in factor because I didn't see one and because it sounds like a fun project [17:04:28] and I tried to run `"https://downloads.factorcode.org" http-get` [17:04:37] and it locks up [17:04:54] any ideas for debugging? maybe I can fix it with some direcdtion [17:16:51] * itmuckel [itmuckel!~itmuckel@mue-88-130-56-099.dsl.tropolys.de] has joined the channel. [18:05:17] Lemme try. [18:05:53] Works for me jfhbrook . What exact version are you on? [18:05:57] And OS for that matter. [18:06:18] Hmm, I only see tow new patches, neither of which should be relevant here. [18:18:46] uh how can I check? I'm on arch linux and using the factor-git aur package [18:19:55] oh got it [18:20:03] Factor 0.99 x86.64 (1982, heads/master-6c02569916, Jun 17 2020 04:42:11) [18:20:03] [GCC 10.1.0] on linux [18:26:34] Okay, so the only two differences are I built with clang and I'm on a Mac. [18:26:42] Biggest suspect given the URL is gonna be Libre/OpenSSL. [18:26:49] I can grab my Linux box, but it'll be a few. [18:26:58] Like an hour. [18:27:00] yeah I was thinking https too, I tried https://google.com and that was fine [18:27:03] yeah sounds good, take your time [18:27:09] appreciate you looking [18:27:13] oh [18:27:17] Are you on the CLI or in the GUI? [18:27:17] one thing to note [18:27:19] gui [18:27:23] Okay, so not that either. [18:27:28] but one thing to note - on my machine ipv6 lookups are a problem [18:27:36] and I don't know why [18:27:43] but maybe it's an ipv6 issue [18:27:49] like dns resolution [18:27:52] I'm on IPv6 and factorcode.org is (was?) on IPv4, so I wouldn't assume anything. [18:28:15] yeah no assumptions - but yeah, it's when the AAAA record doesn't exist [18:28:27] Gooot it, and there is indeed no AAAA record. [18:29:28] I'm gonna try on a machine I have that I know has no AAAA record and can reproduce that problem [18:29:44] yup that's it [18:29:51] frick [18:30:04] Okay. So is your current gut that it's your box, not Factor? Because I don't mind trying to debug this if it's a Factor-on-Linux thing. [18:30:42] my gut is that it's both [18:30:47] that factor should time it out [18:30:52] but that the underlying issue is my network [18:31:05] Roger. Mind filing a bug, and I'll see if I can't repro it later? [18:31:19] yeah sure - tracker is gh issues? [18:31:20] I'm on the clock at the moment, so I need to turn back to bringing home the kosher bacon, but I'll try to poke this evening. [18:31:24] yeah gotcha [18:31:27] Yes, GitHub issues please. [18:31:36] I'm gonna see if I can reproduce the ipv6 issue on my windows machine [18:31:46] and shake down some arch linux people if it's linux only [18:31:49] I find it unlikely, since the networking stack there is completely different, but mebbe. [18:32:32] yeah I'm guessing that a dns lookup w/ the windows tooling in powershell reproduces but that factor doesn't - but just a guess [18:32:51] * Sgeo [Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net] has joined the channel. [18:33:37] That'd be unlikely, but I'll be curious about that, too. [18:35:16] yeah I'm really curious to see what happens [18:39:40] ok, Resolve-DnsName is able to look up AAAA records without barfing, though I don't understand its output [18:41:18] getting a little lazy and trying 0.98 from the downloads page instead of the git build but that should be good enough for me [18:44:35] That'll be easier due to bootstrapping issues anyway, but it "should" have the same behavior, doing a quick glance. [18:44:51] Factor actually has a DNS vocab BTW that isn't used for `http-get`, but I'm curious if it has the same issue or not. [18:55:57] I was able to sort out the deal with my linux stuff, networkmanager was ignoring my global dns settings completely and insisting on using the completely broken dns inside my router [18:56:07] and now factor works [18:56:48] so I'll file an issue w/ factor and leave the arch people alone, since all those people are basically hostile to anyone dealing with bad networking to begin with [18:56:52] "sounds like a personal problem" [19:04:04] https://github.com/factor/factor/issues/2312 I filed an issue - a pretty solid nice-to-have I figure [19:07:36] I mean...have you considered not using Arch?.... [19:07:45] Or filling a ticket with the GNOME project, which is a lot kinder? [19:07:59] s/filling/filing [19:20:40] "sounds like a personal problem" lol [19:21:31] There are many projects where I have a technical agreement with what they're doing, but think their personal interaction is extremely caustic. Arch has pretty reliably been in that bucket. [19:22:24] Void and Alpine are similarly minded distros that seem to have more welcoming communities, but I know, from two and a half decades of this, that switching distros is really painful. [19:22:24] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has joined the channel. [19:23:27] :) [19:26:28] bmps: Regarding the language server. I've rebuilt it to use stdin/stdout and it's still not working. Neither with vscode nor with intellij. So I just wrote a small ConsoleApp in C# to rule out that factor on Windows is broken. Still not working. So now I created an issue at intellij-lsp asking how it actually communicates with a language server. I feel really stupid now. :( [19:27:25] It's okay. I really appreciate you spending all this time on figuring it out. I'll still try to help when I have some spare cycles. I'm going to *try* to take tomorrow off to split my time entirely between Factor and Mercurial, so...hopefully that'll be useful. [19:28:12] Is it helpful if I dig up some work I'd done from a year or two ago and see, at least at the integration level, whether my initial memory I'd used http was right, or stdio is what I did? [19:31:41] Depends: Did you get the initialization step / the communication working? Either way, I'd like to read some factor code doing what I'm trying to do for some inspiration. :D [19:37:34] I have no idea; I'll need to grab the laptop. [19:37:47] I think *all* I got working was convincing VS Code to load my server and then crash. [19:37:58] But that sounds like it's the reverse problem you're having, so I'll check in a few. [19:38:43] jfhbrook: I wanted to fully trace this before I made the following comment: Factor is just directly using `getaddrinfo`, so the issue is going to be entirely on the Arch side. [19:39:25] If you want Factor to time out, we'd need to yank the entire DNS querying in-house, which wouldn't actually really be a good thing (e.g., we wouldn't honor `/etc/hosts` on Linux systems, or AD settings on Windows) [19:40:01] The main word you want to look at is `resolve-host`. The implementation of that word is determined by platform, so it'll be different on Windows, but there you go. [19:40:51] The delay was mostly me making 100% sure that was the word being invoked via `http-get`, which involved me refamiliarizing myself with the sockets system. [19:50:29] bmps: yeah I agree that rewriting getaddrinfo is out of scope for factor [19:50:56] as far as arch, I kinda wanna switch to nix or guix or something like that [19:51:16] nix is cool and they're nicer than the arch people but things have their own idiosynracies there [19:54:23] * tgunr [tgunr!~davec@dsl-dhcp-conrtxxahre-72-26-18-39.consolidated.net] has quit (Quit: My MacBook has gone to sleep. ZZZzzz…). [19:55:15] I'm with you in principle, but I'm going to wait for them to iron out a lot of things first. [19:55:45] If they stick to, "don't use npm/go get/etc., just use Nix," they're not gonna get beyond niche. And I don't know what the Nix-friendly solution to those domains look like--and I don't think they do, either, which is, I think, why they're saying that. [19:55:46] yeahhhhh [19:55:50] So...I'm going to hold my breath. [19:56:05] a friend of mine really deep into it has forked it with a few dozen friends that are basically making an online commune [19:56:14] and it's like that's chill but yeah I gotta do stuff [19:56:40] Yerp. [19:57:39] I usually think in terms of what one boss I had called "innovation tokens," but which I've historically called "seat-of-pants points": any given project gets a certain fixed count of points to do something way off normal, but that's it. If you go over budget, you're going to fail, because you're going to be focused on solving problems you caused more than the problems you wanted to fix. [19:57:54] Factor is always where I spend my points when I get to pick. Factor + Guix/Nix is a lot of points. [19:58:11] I'm using it for a project and my experience was: when it works it slaps, it sorted out letsencrypt automatically after I set a config flag; but yeah eeeeverything is NIH and at one point I was like how do I deploy this in practice and some people were like, use nixops, others were like, install the nixos toolchain locally and use the --remote flag, and me? rsync with ansible haha [19:58:28] innovation tokens, that's good [19:58:33] and yeah agree [20:30:12] * itmuckel [itmuckel!~itmuckel@mue-88-130-56-099.dsl.tropolys.de] has quit.