Peter Bhat Harkins @pushcx@octodon.social
Joined Apr 2018
On mumble I'm listening in as alynpost and @355E3B work on the ansible playbook for Lobsters. Working remotely, I forgot how nice it is to overheard fellow professionals puzzling through some work.
Tomorrow is #CodeAndCoffee Andersonville, 7 AM until we wander off (lunchish). Weekly time to work on your side project or studies with friendly chats. https://push.cx/code-and-coffee
Weird way to start the day. A guy I banned from Lobsters two years ago for calling for race war (against Kansas, it was not the most coherent racism) shot two cops: https://www.clickondetroit.com/news/2019/11/20/man-named-simba-lion-suspected-in-shootout-with-2-monroe-officers-police-say/
Because stubborn, I want to take another swing at packaging a Rails app in Nix again this weekend. Does anyone have advice or maybe even a worked example? (The ones I've found are incomplete/outdated.)
wtb github, but issues are in the repo for offline editing and I can directly edit PRs to my repo before merging
Really feeling the changing of the seasons. Every day my macos work laptop goes into dark mode a minute or two earlier.
Hi, Dallas - I'm around for a couple days, if anyone would like to grab a bite. Maybe #codeandcoffee tomorrow morning at Ascension Coffee in the Design District?
Next step is 'git blame' to see how we introduced this recently. Right?
Nope. This code is untouched since 2012-06-16. It really never worked.
It just took 7.5 *years* for a website to delete a cookie rather than set one. That's 60k submitted stories before 1 removed a cookie!
The exception came from deep in the common page-fetching code. In set_cookie:
if val.to_s == ""
@cookies[host][name] ? @cookies[host][name].delete : nil
else
@cookies[host][name] = val
end
Sites delete cookies by "setting" them to empty string, so the logic's right, but there's an obvious confusion in how to delete from a Ruby hash, should be .delete(:name). This indeed never worked.
Lobsters fetches web pages for a couple reasons:
* to prefill the title field on new links as a convenience
* to cache story text for the search engine
* to check for rel=canonical links
* to auth github/twitter/keybase accounts
* to send webmentions back to blogs
Interesting Lobsters exception in my inbox this morning: ArgumentError in `delete': wrong number of arguments.
What? We never get clever with *args, so how did this code ever work? Funny story about the rest of the web, actually...
Over on Barnacles, the site for bootstrappers: I closed signups, opened the invite queue, and caught up on 2.5y of code from Lobsters: https://barnacl.es/s/rvjtqz/signups_closed_invite_request_open
In Chrome, you now need to go to chrome://flags , search for "steady", and disable all three "experiments" to see the address of the page you're on in the address bar.
Nice analysis, to evaluate risk of dying during various sports/hobbies by active hours: https://chessintheair.com/the-risk-of-dying-doing-what-we-love/
just venting a bit - learned about the neat way it supports defining a controller in test (thanks emrox) and bumped into a limitation: https://github.com/lobsters/lobsters/commit/6c804eeea641d11fbbd9992435cbaee2197d9c32#diff-f803dcdbbef9948899992eb475f5af1eR17
more like arghspec amirite
In Illinois we they don't call it the DMV they call it CyberDriveIllinois.com and I think that's beautiful.
Doing some financial modeling with a spreadsheet and it has a bug *somewhere*. Trace Precedents/Dependents is a huge help, but I think I'm about to start writing property-based tests to enforce double-entry bookkeeping.
I just remembered that in 1998 I incorporated my first business and my lawyer refused to name it "International Cocaine Smugglers, Inc." and not because it was an LLC.
Joined Apr 2018
