Dave Copeland :ruby:

Author of BrutRB, our last, best hope for Rubykind. Former CTO @ Mood Health. Former Stitch Fix, LivingSocial. Posts about Ruby, Tech, Maybe Cats

Public Key

npub1p2v0e36yqzyk6697anan9k3rp3jkmdhu026rcs3gka64xa7jjneq5cjyrl

NIP-05 Address

davetron5000@ruby-social.mostr.pub

Profile Code

nprofile1qqsq4x8ucazqpztddzlwe7ejmg3scetdkm784dpugg5twa2nwlfffuspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds0c2ds5

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1p2v0e36yqzyk6697anan9k3rp3jkmdhu026rcs3gka64xa7jjneq5cjyrl

Show more details

Last Notes

2026-08-07 13:40:02 UTC

Ep 2 is my least favorite of any sci-fi, which is everyone is hallucinating and nothing you are watching is actually real. I fucking HATE dream sequence shit and the like.
Ep 3 is a type that I like on occasion - main cast take on different personas - but they must be used sparingly. SNW leans too heavily on this style - S01E05, S01E08, S02E05, S02E09, S03E02, S03E04, S03E08 - those are all generally bad episodes (though I liked the musical :)

2026-08-03 17:09:04 UTC

- reply

@nprofile…zs9u Yeah, I have worked in healthcare and navigated it deeply with parents. Incentives are fundamentally broken but quality can be variable (often high!). I would not consider any health plan I've ever heard of being both "inexpensive" and "high quality", esp including meds, copays, and co-insurance. Anything from an employer is usually masking the true price (I think when you leave and pay COBRA, that is roughly the true price of the plan you had with the company?)

2026-08-03 13:56:33 UTC

I'd love some of this inexpensive health insurance / health care. I have good insurance - rarely denied and rarely any trouble in the last 25 years. It's my second largest expense after my mortgage and creeping up. I’m healthy and rarely need specialists or uncommon meds.
The only truly good insurance is for those lucky to be on Tricare4life which covers everything medicare doesn't. If my dad didn't have that, he'd have been broke or dead long ago. Only because he was military does he have it.

2026-07-26 18:08:07 UTC

Grades from left to right:
A-, B-, F-, F, C, C
If Apple hired people with hands, imagine how awesome their remotes could be.

2026-07-13 10:42:53 UTC

Looks like Firefox on iOS now obscures the app while app switching, which means that it feels incredible slow. I can't figure out how to turn this off (or why it's even on)

2026-07-01 19:37:03 UTC

- reply

@nprofile…nsyu @nprofile…pf23 Yeah, the "owners" seem to be nonexistent (Leylan Labs is parked on namecheap) and the repo doesn't seem to have had substantial updates in years. But it's as helluva domain name :)

2026-07-01 19:35:31 UTC

- reply

* useful formatter - I have no idea what this is saying. I guess use Fuubar?
And that's it! Almost entirely bad advice, and I guarantee you that a codebase that doesn't follow these guidelines will garner less complaints from the RSpec Hesitant on your team.
Bonus rspec thing to avoid: predicate matchers! Your code doesn't have the method be_well_formed in it, so why is your test calling it? .well_informed?).to eq(true) is fine.

2026-07-01 19:32:47 UTC

- reply

* faster tests - Do zeus/spin/spork even work? My experience with stuff like this is you waste all your saved time because something sat in memory when it shouldn't and you are testing old code.
* Stubbing HTTP requests - i.e. use Webmock. I'm not a fan of webmock/VCR - I use service wrappers https://naildrivin5.com/blog/2022/10/31/wrap-third-party-apis-in-service-wrappers-to-simplify-your-code.html I know some rails devs are allergic to putting code outside active record, but read the post, it's way easier to test code that hits APIs if you design it this way.
6/

2026-07-01 19:30:45 UTC

- reply

* shared examples (to DRY your test suite up) - HARD PASS, do not do this ever. Repetition in tests is fine, especially given how convoluted shared examples are. never use them.
* test what you see - unclear what is meant here. Don't test controllers? Agreed, sorta? no points.
* don't use should - we covered that already, no point for this either
* automatic tests with guard - again woefully naive, what apps test suite runs in seconds that you would want this?
5/

2026-07-01 19:27:58 UTC

- reply

* mock or not to mock - this isn't given good advice, but yeah, don't mock always, but also, don't never mock! mocks are great!
* create only the data you need - this is woefully naive advice. Yeah, don't create data you don't need, but if you are doing rails, it's REALLY hard to make realistic data for a test that is valid and be saved to the DB.
* use factories not fixtures - yes. that's 2!
* easy to read matchers - I don't get this one - do people really use lambda instead of expect?
4/

2026-07-01 19:25:49 UTC

- reply

* expect vs should - yes, use expect. OK, that's 1 point for.
* use subject - NO. RSpec tests are _examples_ and you'd never name a variable "subject" in your code, so you shouldn't in your test. See also: "don't use it { is_expected.to }"
* Use let and let! - @nprofile…65fy told me once at his job these methods were forbidden. That sounds amazing to me - let and let! are another chapter in the rubyist obsession with "spread logic over a zillion places…a lot"
3/

2026-07-01 19:23:43 UTC

- reply

* keep your description short - why? make it as long as it needs to be. That's why it's a string and not a method name
* single expectation test -NO NO NO NO NO - do not run a test 5 times to test 5 things. Run it once and test 5 things. aggregate_failures if you want to see all expectations evaluated per test
* test all possible cases - reasonable advice, but the example shows my favorite rails anti-pattern: spread controller logic across hooks
2/

2026-07-01 19:21:54 UTC

I think if you did the opposite of every single thing in https://www.betterspecs.org, your RSpec tests would be better. Maybe not all of these are questionable but on the whole blindly doing the opposite would be far better:
* describe methods - actually is fine if you are testing a method, but you can and should test more complex things
* use contexts (which is really showing "use it { is_expected_to }" which is the worst. the WORST. DO not use this ever
1/

2026-07-01 19:18:15 UTC

- reply

@nprofile…pf23 Oof that site - I come across every few years and there's just a lot of "How to make RSpec Your Team Will Hate" advice in it.

2026-06-11 21:15:55 UTC

I have been getting so much SendGrid spam/phishing. Basically emails from other companies telling me my Sendgrid account is either cancelled, in arrears, or upgraded.
The thing is, I very recently closed my SendGrid account (which is a whole other dumb story that can be summarized as "Fuck Twilio"), so my monkey brain is trying to connect those two events.

2026-06-05 19:30:31 UTC

- reply

@nprofile…7fqd is it some sort of handoff thing? When I had that set up such that sms and voice came through to my Mac, so much stuff acted wierd. Phone calls unanswerable or not ringing anywhere, sms lost, etc. I just disbaled all handoff features and life has been more predictable.
Other option is delete the app? Reinstall or use web?

2026-05-13 16:21:43 UTC

This is going to be a problem with LLM-generated code going forward. And this problem isn't the one I identified in Death of a Software Craftsman[1]
We really should not be making websites by sending gobs of React to the browser and having it generate HTML. We should not be creating convoluted systems of server-generation to pretend React works on the server. We should be embracing the web platform as it evolves. LLMs will not be able to keep up.
[1]: https://naildrivin5.com/blog/2026/02/23/the-death-of-the-software-craftsman.html

2026-05-13 16:14:07 UTC

LLM-generated non-framework front-end code is such a pile of hacks, divs, and spans. Not surprising since most front-end code is either MBs of React or a pile of hacks/divs/spans. Most newer code I've seen using WebComponents tries to recreate React by having HTML generated inside JS, e.g. el.innerHTML = `a big ole string`.
If you actually USE CSS, custom element attributes, semantic markup and aria roles, you can achieve a lot with a LOT less code and very little client side HTML generation.

2026-05-07 17:47:01 UTC

- reply

@nprofile…4stt @nprofile…6mvt Then my snarky name for all such browsers is "Waiting on Google" :)

2026-05-07 17:38:11 UTC

- reply

@nprofile…6mvt Is Vivaldi a chromium-based browser?

2026-05-07 14:54:41 UTC

Browser choice still sucks and you must choose one of:
- spies on you
- ugly
- doesn't work
I am sticking with Ugly for now, but keep Doesn't Work around for Apple Pay and Spies On You for websites built by shitty engineers.

2026-04-28 18:30:32 UTC

As a paying subscriber to YouTube, I would fucking LOVE if: a) they do not randomly turn on closed captions, b) did not show me shorts if I have indicated as such, which I have about 1,000,000 times, and c) ibid for games. I'm paying for this shit.

2026-04-17 16:04:44 UTC

- reply

@nprofile…vlm2 And all respect to the work you have done - I don't mean to imply it's all horrible and I'm not as mad as some about the bundler stuff - I really do appreciate the difficulty the org has faced over the years.

2026-04-17 16:03:46 UTC

- reply

@nprofile…vlm2 Based on Richard's post mortem, it seems that Andre and Mike made decisions based on DHH being invited to RailsConf and that precipitated the recent unpleasantness. Maybe I'm mis-understanding (and I understand there is more to everything).
But "volunteer or be silent" is a false dichotomy - I'm having a hard time getting excited about a nebulously defined AI thing. And I have offered to help the org over the years. Was not taken up on it. I'm not sure what else I'm supposed to do.

2026-04-17 15:42:42 UTC

What a wild butterfly effect: Ruby Central invites DHH to RailsConf and now:
* Almost out of money
* Bad feelings/animosity with at least some of the major Ruby open source contributors
* ED fired
* board almost totally turned over
* primary function (rubygems.org maintenance) now being hopefully outsourced to volunteers
* some AI bullshit
All RC had to do was simply NOT platform him. No big statement, no virtue signaling, nothing. All they had to do was have RailsConf.

2026-04-07 16:01:43 UTC

Writing a post on a Discourse-based forum for the first time in years and I just cannot understand how bad the UX is after all these years. TONS of popups nagging me to do stuff, and my post is given one of the tiniest windows I've ever seen to write in. 0% of the buttons have tooltips/titles to help understand what they do.
As I hover over "Create Topic", I have zero idea what it's going to actually do. PPHBB was awesome and I weep for its decline :'(

2026-02-23 20:22:41 UTC

- reply

@nprofile…s50v I have been a depressive haze for the last few months because of this. I'm not sure if writing this post helps, but I think it does. Though still not sure what this means for me personally.
Craft{er,sman,sperson} - I don't love the term, but "maker" doesn't seem right to me, nor does "artisan", but maybe it's just "coder"?

2026-02-23 17:53:50 UTC

The Death of a Software Craftsman (it happens a lot 'round here)
https://naildrivin5.com/blog/2026/02/23/the-death-of-the-software-craftsman.html
AI Abstinence? All in on Agents? Or resign yourself to becoming a niche craftsperson?

2026-02-21 10:31:40 UTC

The killer feature of Kagi that made me sign up and pay for it was that you could tell it not to show Pinterst results.

2026-02-20 00:07:30 UTC

In other news, I think I'm entering my "apps I use need to work" era. This means abandoning a lot of Apple sodtware.
Example: GBoard as my keyboard. I hate google but Apple's iOS keyboard simply doesn't work. Ibid Apple Maps, Notes, Mail.app

2026-02-09 20:39:22 UTC

- reply

@nprofile…7fqd @nprofile…p334 I would love if Ruby had a protocol for invoking a method on a constant like that. It would break a million things probably, but "turn a thing into me or at least try" is a really handy thing to have on all objects.

2026-02-09 19:20:24 UTC

Holy shit Obsidian has vi mode.

2026-02-08 18:39:57 UTC

- reply

@nprofile…qad4 You are luckier than me - I have a huge archive (back to 2000's, I never delete things) and mail search often never returns a result, but gmail and fastmail instantly return pretty much every time. (this is on iOS mail - haven't used macOS mail in a while so if it got better I may not have noticed)

2026-02-08 17:51:49 UTC

I'm switching from Apple Notes to Obsidian. I just can't take Apple notes incessant thirstiness for engagement† any longer. I'll still use it for the shared lists with my wife.
That said, having set up sync with Obsidian…no regular person would ever be able to use this in a million years. It's so convoluted and confusing, I barely got it set up and working. No normal person will ever persist through it.
†constant opening of activity pane, constant linking of names to contacts, etc.

2026-02-08 16:09:34 UTC

- reply

* Calendar sync hasn't worked reliably in years (I've long since abandoned it)
* Mail continues to add my email to reply-alls even though it knows what my email is. Have to manually remove it every time
* Mail search doesn't work now, and never has ever
* Creating a new Numbers sheet has a hard limit to a tiny number of rows/columns requiring fussy dragging in order to add rows/columns. WHYYYYY
* Opening non-native docs in Number shows an undismissable dialog about fonts. EVERY. FUCKING. TIME.

2026-02-08 16:07:19 UTC

The most recent @nprofile…hq9d has caused me to critically question why I'm using any given Apple solution. Apple's software has just become so full of papercuts:
* Maps search is SUPER SLOW
* Maps "pinned items" only shows the first 3 as pinned, the rest behind a menu
* Opening a shared note always opens up the activity tab which I give a fuck about
* Safari on iOS is a nightmare of tedious animations to get to anythning
* Safari on macOS just doesn't work with Kagi or 1Password and hasn't in years…

2026-01-27 19:07:47 UTC

I want some of what they are smoking over there at Mozilla HQ
https://mas.to/@zzt/115968498211919805

2026-01-07 01:07:30 UTC

Both chili and cornbread waffles freeze great! Made weeks ago, perfect for a cold evening. /cc @nprofile…s50v
https://cdn.masto.host/rubysocial/media_attachments/files/115/851/135/976/814/590/original/e8a6dde5be021dd1.jpeg

2025-12-17 21:03:43 UTC

- reply

@nprofile…pzz0 whew! let me know how it goes!

2025-12-17 19:48:46 UTC

- reply

@nprofile…pzz0 if that is happening even with nonstick spray, you might either be cooking too long or need it to be more liquidy (add milk or water to the mix). i swear i thought this was an easy win.

2025-12-17 19:47:10 UTC

- reply

@nprofile…pzz0 oh shit I’m so sorry :( I spray mine with nonstick spray (which creates another mess but easier to clean). i hope i didn’t ruin your evening :(

2025-12-13 22:10:44 UTC

- reply

@nprofile…pzz0 @nprofile…2ft4 with the chili instead of notmal cornbread. i serve the chili kind on top or as a dipper but use the same recipe, just waffle it.

2025-12-13 20:11:32 UTC

- reply

@nprofile…pzz0 a solid meal for a group since it’s hard to mess up, still requires chopping and whatnot, plus cornbread is GOOD.
if you want to up the challenge and you have a waffle iron, you can do cornbread waffles.

2025-12-12 19:40:03 UTC

this sounds absolutley dystopian. maybe it makes the job so awful that salaries fo up? or creates a new programmer underclass that just reviews AI-generated PRs. https://thoughtbot.social/@thoughtbot/115707504356691614

2025-12-05 17:04:59 UTC

As a…owner of a grocery store in the USA
When…anyone tries to buy something
Then…assume they are shoplifting literally everything.

2025-12-01 18:59:58 UTC

- reply

@nprofile…wtxa sounds like you went straight to 26.1? On 26.0 it required force quitting the phone app to even roll the dice on accessing it.

2025-11-30 21:35:12 UTC

This pathological payment pattern I can't wrap my head around why it exists, because it just happened with STRIPE, who should know better: 1) default method is expired and has failed, 2) add new method, charge goes through, 3) expired payment method is still default, fails next time.

2025-11-30 21:35:12 UTC

- reply

I just can't think of why on earth you would ever want to maintain the expired payment method as default. At least ask if I want the new one to be default? But every single recurring payment system I've ever used does this. WHYYYYYYY

2025-11-29 01:37:14 UTC

More and more websites don’t work on iOS without disabling content blockers. PayPal is the latest and most surprising.

2025-11-26 15:29:36 UTC

1 hour into getting 1password + firefox to work. Main sticking point so far is getting it to start when I log in. It just always produces a crash dialog, but then seems to actually start. (Please don't tell me about other password managers, I'm not switching)

2025-11-23 19:42:41 UTC

- reply

Some people really want a higher quality product (and can often tell the difference). Those products can't be produced by industrialized methods. They require a skilled artisan/craftsperson. Such artisans are highly skilled, meticulous in their work, and often sought-after.

2025-11-23 19:42:41 UTC

- reply

The new software crafter isn't something each professional programmer aspires to, but instead a description of someone that hand-creates their software, taking more care than one using LLMs. The goal is to produce a higher-quality product, when that is requested.

2025-11-23 19:42:41 UTC

- reply

But there will always be a small group of highly skilled programmers who achieve far better outcomes—fault tolerant, secure, high-performing, etc—by careful application of a skill long-honed by experience. There just won't be nearly as many as their are developers now.

2025-11-23 19:42:41 UTC

- reply

I'm now wondering if this is the future of software? Industrialized software created by moderately-skilled operators (prompting and managing LLMs) make most software, because the quality—bugs, security, performance—just don't matter that much.

2025-11-23 19:42:41 UTC

- reply

What LLMs have demonstrated—at least thusfar—is what I think many experienced devs have suspected for a while: in most situations, quality does not matter. Most people don't care about their furniture, wine, food, music, etc. They just need something that works. But SOME people want more

2025-11-23 19:42:41 UTC

I initially bought into "software craftsmanship", but soon soured on it, as it seemed to be couched as moral imperative, not a practical/outcome-based one. But now that LLMs allow a lot of rank&file devs to get work done in a much more industrialized way, I'm wondering if "craft" has new meaning?

2025-11-23 19:42:41 UTC

- reply

"Software quality" is somewhat nebulous, but then so is "quality" in regard to any field that has both craftspeople and industrialized production, e.g. furniture, art, music, decorating, food, etc.

2025-11-20 15:55:40 UTC

I got 5, but not a bingo https://mastodon.social/@BasicAppleGuy/115582711438367134

2025-11-19 15:17:18 UTC

If I'm understanding this Cloudflare thing right, it looks like the smug use† of monads took down the entire internet?
†is there any other kind?

2025-11-17 19:14:37 UTC

I can't help thinking when I read posts like the one linked below that it would just be easier to learn how to program than to do all this incessant meta-work and quality control over something that not only cannot program well but will never learn how to, either.
https://cosocial.ca/@timbray/115566455140045758

2025-11-05 18:03:25 UTC

Reason # 467 to not rely on software produced by 37 Signals:
https://github.com/hotwired/turbo/issues/1456
My only surprise is that David Rails didn't just post a picture of the fuck you slide and actually wrote a response

2025-10-29 21:08:37 UTC

- reply

@nprofile…t0hf Do you think they are committed to maintaining it? Despite the fact that Apple seems to stop maintaining tons of their software, they do seem to keep Logic updated and have for a very long time

2025-10-29 02:39:44 UTC

OK, did today’s #looptober. 4 in one day, but I think this is the most interesting one.
NGL, using a single-cycle saw on the Digitakt sounds better to me than the actual analog East Beast.
https://bandwagon.fm/69017ddc5b9e33de2569c27e/audio?format=mp3

2025-10-29 01:41:49 UTC

- reply

End of day, Apple's leadership is falling down on the job. Hardware seems to be the only thing that isn't terrible, but that means that Tim Cook, C-Fed, and just about everyone else is checked out and not doing their job very well. Alan Dye, as previously discussed, sucks.

2025-10-29 01:40:28 UTC

- reply

And you can no longer book a genius bar appointment through the apple website - so I'm on a chatbot to try to do it. What a mess.

2025-10-29 01:30:27 UTC

New iPhone Air USB port became unusable 11 months ahead of schedule. Air, needle, etc. not doing it. FML I will probably have to send it in for replacement.

2025-10-14 23:47:08 UTC

- reply

@nprofile…wtxa Do you think this is viable in Rack apps as well? It seems way cleaner and easier than the whole token/session thing that is in use in Rack/Rails.

2025-10-11 03:02:25 UTC

- reply

@nprofile…83zn @nprofile…p3s2 @nprofile…66kr you can do automations on texts?! Or is it more like you run it on a text yourself?

2025-10-10 22:32:35 UTC

- reply

@nprofile…66kr STOP is handled at the carrier level. Yes, the dems can keep getting numbers because political spam is not subject to the same approvals but if you internalize doing it will slow down eventually. I always stop and block just to be safe

2025-10-07 22:37:04 UTC

- reply

@nprofile…a46r yeah, I think git used to take the command name and exec git-$commandName to allow extensibility.
I’m not sure how avoid clashing global vs command options but maybe avoiding that is a feature?

2025-10-07 19:37:09 UTC

Related to this (putting dev scripts in bin/), I realized that Brut's set up of making one-off scripts like bin/db, bin/scaffold, etc. is not viable long term. Ideally these would be packaged with the gem, but if a gem includes a script like "db" that would be available in the $PATH due to how gem install works. So, I plan to make it all `brut db`, `brut scaffold`, etc.
But I will still not use Rake for any of this.
https://ruby.social/@davetron5000/115334545191525677

2025-10-07 15:43:46 UTC

- reply

@nprofile…ktkr I call it a “Dev Enironment Protocol” in two parts: Foundation sets up DB servers and VMs, Workspace sets up the app itself.
https://devbox.computer/protocol/index.html

2025-10-07 15:41:42 UTC

This is so handy when you have it and maintain it. You need it on all levels. Brut has bin/setup fo this and it calls bin/db rebuild to blow away and recreate the database. https://ruby.social/@jardo/115333595497883106

2025-10-07 01:42:49 UTC

#looptober for Oct 6, this time with more techno (I think?) sounds. I just got a TR-8, and the combo of it + the Drumbrute Impact sounds great in my headphones.
Steampipe is packed up & awaiting my listing on Reverb. I was also going to sell my Microfreak, but I think I might try harder to love it.

2025-10-01 18:38:45 UTC

Y'all, I'm retiring from public life - ParkMobile had a massive data breach and a class action lawsuit was settled. THE SYSTEM WORKED! I'll be clearing a cool $1.00 (USD) that I can use only in 25 cent increments and only when I use ParkMobile. I've struck gold!!!!

2025-09-09 23:38:59 UTC

- reply

@nprofile…83zn I did have moderate success with a sewing needle - but I think I should’ve been more proactive. There’s something impacted in there

2025-09-09 18:15:34 UTC

Honestly, the main driver in getting a new iphone is that my current one—2 years old—has accumulated so much pocket lint that it rarely works with CarPlay or charges through the usb port. Taken it to the apple store where they got out dental tools to try to dig it out of there. It sorta worked? but not really.
Of course, Amazon has the solution:
https://www.amazon.com/dp/B0BXWQ5W7P
But I got it too late :(

2025-08-15 14:08:00 UTC

- reply

@nprofile…uzec isn’t a ranked choice algorithm basically an interview question?!?

2025-08-12 15:28:04 UTC

- reply

@nprofile…uzec @nprofile…cut6 I'd actually like it if the page refused to render when an id was used more than once 😂

2025-08-06 13:19:13 UTC

- reply

I assume it works better in VSCode? I know I'm using vim/containers, but it's a helluva SERVER that uses stdio and has no way to debug or even log anything.
This is a Ruby cultural norm pioneered by Rails: It Just Works Until It Doesn't And There's No Logging or Debugging Don't Worry Sweetheart

2025-08-06 13:19:12 UTC

Welp, ruby-lsp just stopped finding definitions. Nothing has changed - I have a completely isolated reproducible environment. ruby-lsp provides no debug abilities, does not log anything. --doctor just prints a list of files, --debug does nothing.

2025-07-29 15:29:59 UTC

Imagine cutting all the deals needed to make the Apple News+ app work and then letting it rot for years and allowing it to be a total garbage app.

2025-07-25 17:19:32 UTC

Linting causes more problems than it solves, Part 87
https://ruby.social/@soulcutter/114914897230922151

2025-07-24 17:50:00 UTC

- reply

It's especially important with "enterprise" vendors who are new to doing APIs and will be moving fast and breaking things. Having your code at the HTTP level, and the ability to easily debug literally anything that happens, is critical to resolving problems.

2025-07-24 17:50:00 UTC

Since Twilio rejected my business license and won't allow me to send reminder SMS, this means my SendGrid account is inaccessible, too. So I'm looking at other options, and MailGun is my current pick. It's "Ruby code examples" just use Net::HTTP and I think this is the way(ish)

2025-07-24 17:50:00 UTC

- reply

I'd use Faraday because it's a bit more ergonomic, but, after doing many many API integrations over the years, life is much better when you do it directly with an HTTP library instead of the vendor's library. Stripe is the only example I can think of where their library is better than just HTTP.

2025-07-24 17:50:00 UTC

- reply

The best way to integrate this sort of code into your app is with Service Wrappers, which create an API exactly and only for your use case. No third party dependencies, no weird abstractions…except the ones you take with you (🫱📦).
https://naildrivin5.com/blog/2022/10/31/wrap-third-party-apis-in-service-wrappers-to-simplify-your-code.html

2025-07-23 15:53:09 UTC

This is classic VC/Founder bullshit from Altman: “Our app/tech/company will be amazing for everyone and change the world! No, I won't be using it though, nor will my family”
https://mas.to/@carnage4life/114903338900466064

2025-07-23 15:04:01 UTC

Build a Blog in 15-ish minutes with Brut!
https://video.hardlimit.com/w/ae7EMhwjDq9kSH5dqQ9swV
Source code for those that don't want to watch a video: https://github.com/thirdtank/blog-demo

2025-07-23 13:44:37 UTC

Let's not forgot how just awful and user hostile the Apple Podcasts UI is - you can see exactly 1.75 podcasts you subscribe to, and then the rest of the screen is ads. Alan Dye does not care about your content. https://mastodon.macstories.net/@viticci/114898260376387566

2025-07-09 20:32:32 UTC

- reply

@nprofile…ecmu Yeah, I do allow(SomeClass).to receive(:new).and_return(some_object), and then allow(some_object).to receive just to make it explicit.
At any rate, mocks are a tool in the toolbox for me, not a way of life. I also don't think DI _just_ for testing is good. It’s fine if dependencies need to be configuration, but if they don't, I don't see value in making them configurable when, at least in Ruby, you can mock them by replacing .new.

2025-07-09 15:45:46 UTC

- reply

@nprofile…ecmu I should probably write this up more but the tldr is I like the mocking system better and I like custom matchers. There is a lot about RSpec I don’t like though so def a tradeoff.

2025-07-08 17:51:18 UTC

I'm not at #RailsConf, but I did just now release a new web framework I've been working on: BrutRB:
https://brutrb.com
Read about it on my blog, below. It's not like Rails at all. I had fun working on it, will keep working on it, and hope you like it!
https://naildrivin5.com/blog/2025/07/08/brut-a-new-web-framework-for-ruby.html

2025-07-03 17:35:17 UTC

- reply

@nprofile…383v @nprofile…9vce I understand, but it doesn't have to be that way. That is a choice made by the designers.
And I would nitpick that calling something a function where the ordering of the contents aren't relevant to its behavior IS new and weird in the context of most other popular programming languages, as your original post points out.

2025-07-03 15:49:37 UTC

- reply

@nprofile…9vce @nprofile…383v that was my assumption, too, but it doesn’t have to be this way. CSS could blow minds with awesomeness and not wierd semantics.

2025-07-03 15:21:52 UTC

I like that CSS is evolving. But, it would be nice if new features had less of this random flexibility. https://front-end.social/@css/114789914481764351

2025-06-27 00:56:59 UTC

- reply

@nprofile…v4c2 I was using cursor and had their rules file and literally pasted in the rules and mentioned the file in every prompt.

2025-06-27 00:56:12 UTC

- reply

@nprofile…am9h I’m about to make a minimal Rubocop set of rules to deal with this.

2025-06-26 22:47:51 UTC

- reply

I am particular - probably more than your average rank and file who don't care and probably shouldn't care and for whom caring will not get a promo and probably even keep them from getting promoted. But I want the code a certain way and I don't think it's that esoteric

2025-06-26 22:47:51 UTC

- reply

Also, I could NOT get it to avoid yoda expressions. The rules file had it in there written 40 different ways and my prompt included "no yoda expressions" and a ton of examples of what not to do. It still wanted `exit 1 if !call` ARRRRHGHGHGHG