Looks like the Utah snapshots no longer include a “current” 64bit macOS BC version:
https://github.com/Bogdanp/setup-racket/runs/3649309725?check_suite_focus=true
Is this expected? I see one on the Northwestern server.
Is this still correct?
> Sometimes you need a virtual display when running racket headless: xvfb-run racket -l handin-server
https://github.com/racket/racket/wiki/Headless-with-virtual-display\|https://github.com/racket/racket/wiki/Headless-with-virtual-display
It looks like the build slowed down enough to cross a threshold, and there have been timeouts. (The Racket build doesn’t appear to be getting slower in general, based on http://build-plot.racket-lang.org\|build-plot.racket-lang.org); something about the old machine or its configuration may be making it slower. I’ve upped the timeout, but that won’t fix the problem for another day or so.
yeah, thanks guys! its working now. how do i make it public via domain name/url since students will need to connect to it… its running on localhost on the vps right now, so i suppose its just https://domainname.com:7979 (which is the port it is running on)
did it work?
@jestarray Is the dns of your domain name set to point to the server?
@soegaard2 yep, already hosting a few websites on there right now
nope, that is not working
the server is up though, unsure whats causing it… do i have to edit the config file of the handin server with my domain name or something ? reading the docs and its kind of unclear how to set this up beyond localhost
What did you do to get the other websites running?
i serve some static files using nginx
I don’t remember anything specific to the handin server. Is the server inside a NAT or requires local/university IP?
I am using nginx in front of the Racket web-server.
Can the handin client connect to the server? I think the web needs to be enabled explicitly
The nginx instance simply forwards the traffic to the racket web-server.
handin client cannot connect to the server no… i edited the server:port correctly but it says it cant connect
https://docs.racket-lang.org/handin-server/server-setup.html A student can download his/her own submissions through the handin dialog. This can also be done through a web server that runs concurrently with the handin server (on the same port) if you set the use-https option in the configuration file to #t (the default). The starting URL is
<https://SERVER>:PORT/
to obtain a list of all assignments, or
<https://SERVER>:PORT/?handin=ASSIGNMENT
You are right. The web interface should’ve been on by default. Is the firewall blocking incoming connection to that port?
ohhh right, think i have ufw filrewall enabled
lemmie see if i can let it through
sweet! it seems like i can connect! love you guys :heart:
was gonna make my students learn git and upload their solutions there lol
@shu—hung curious, are you a TA for a racket course?
Yes. But I moved away from the handin server a long time ago because I want to just use Canvas like other courses.
Well, not that I’m TAing this quarter.
canvas does have a nicer ui lol… but i dont think can get access to it since im teaching this as an independent instructor non formally…. maybe i should move away from handin server also so that i can integrate it with my main website
Yeah, for me it’s a trade off between asking the students to learn an entirely separate system (with its various behaviors) versus having instant syntax checks.
And yes it’s abou the UI too. I just want the submissions and the grading report to be in the same place as the course materials and the grades.
do you guys use autograding tools for the desgin reecipe stuff? e.g checking how many tests they should have and if any steps are missing ?
Nope. The course I was TAing for is not based on HtDP. I don’t know if there is any auto-grading tool for HtDP.
I keep seeing answers to questions I don’t have a the moment, but I’m likely to have in the future, and the Slack message limit is bumming me out. It seems something like https://www.discourse.org/\|Discourse would provide a better knowledge base. It would be nice to combine the relative permanence & searchability of the mailing list with the interaction of Slack, and I think https://www.discourse.org/\|Discourse might be able to do that. It’s free to run the <https://github.com/discourse/discourse|open source version> on your own server.
I’ve been thinking about using discourse for Racket, but I would say that it’s not really going to change the dynamic you describe. It’s a lot more like a mailing list than like Slack, and using it to replace Racket’s mailing lists would not make Slack (or Discord) go away.
Good point - the social/cultural aspect would be as important to change as the technical aspect.
With that said, I do think mailing lists are no longer a cutting edge discussion platform :slightly_smiling_face: and thus we may move away from them towards something like Discourse.
@samth https://blog.discourse.org/2016/03/free-discourse-forum-hosting-for-community-friendly-github-projects/ I mentioned this before in the past but discourse provides free hosting for open source projects! maybe its time we retire google groups
Yes, I know about that.
And, importantly, “Absolutely zero lock-in! You are free to download a complete Discourse export and migrate away from our free hosting at any time.”
someone from the racket team should try n claim it so we can test if migrating from google groups will go smooth
I’d be more excited about the free hosting if it allowed a custom domain for more permanence.
Sounds like to import from Google Groups, we’d need to setup a local server, run their import scripts, export the site, and then import to the free hosted site. So maybe the first step is to setup a local server.
This is from reading the comments https://blog.discourse.org/2018/11/free-hosting-for-open-source-v2/\|here
A more cost effective* approach might be a racket-slack-bot that detects a specific reaction and archives it to the #slack-archive created by @ben.knoble E.g. maybe the :file_cabinet:
Isn’t everything archived there?
Why would we need a special reaction?
I thought that stopped?
I’m glad it hasn’t
It is ongoing! I dont remember the last time we did an export (within the last month or two). I’m looking to improve the system still, but lacking time (new job). The actual archiver is public oss (mostly racket): https://github.com/benknoble/slack-archive-viewer\|https://github.com/benknoble/slack-archive-viewer (for example i want to make it faster for our large archives, and add a search option, probably through google’s site:
queries). The actual archive itself is private, but served to a public website (https://benknoble.github.io/racket-slack-archive\|https://benknoble.github.io/racket-slack-archive).