|
zuh steed
|
# ¿ Mar 18, 2023 22:28 |
|
|
# ¿ Apr 26, 2024 01:08 |
|
well that one time there was a bug in binary search so you see
|
# ¿ Mar 21, 2023 23:39 |
|
Powerful Two-Hander posted:p.s. the auditors are still confused about how code deployments work and why a screenshot is not evidence that the same code is in all environments send them a video of you web inspectoring a tweet to say something else and then taking a screenshot
|
# ¿ Mar 23, 2023 01:02 |
|
could be one of the "gl hf" parts of the json spec, like duplicate keys in an object
|
# ¿ Mar 23, 2023 04:01 |
|
cinci zoo sniper posted:
six months later "did someone change the foo comment field? downstream was scraping it for type information. we need to revert"
|
# ¿ Mar 24, 2023 18:50 |
|
real pros use win1252 so every byte is mapped (ish)
|
# ¿ Mar 27, 2023 21:51 |
|
there's a good ios image fetch transform cache display library called nuke
|
# ¿ Apr 5, 2023 00:23 |
|
efficient persistent async loading processor pipeline
|
# ¿ Apr 5, 2023 04:45 |
|
sounds like it's xlst time!
|
# ¿ Apr 5, 2023 18:58 |
|
reminds me of apple's c stdlib shelling out to perl for something
|
# ¿ Apr 6, 2023 18:11 |
|
idk if that's a great example as the comparison to nullptr is redundant though I learned the hard way that if you do provide a condition, you have to check for falsey yourself. which I guess makes sense when you see how it's specified but also gently caress c++
|
# ¿ Apr 8, 2023 19:38 |
|
I was surprised when apple added a swiftui chart library, but this discussion plus some unearthed memories is making it seem exceptionally reasonable
|
# ¿ Apr 13, 2023 19:08 |
|
Soricidus posted:I’ve been writing php again and holy poo poo how does this language manage to be so bad, it’s actually impressive something like this is on like every third manual page in php, assuming you need to look up every single function (like I do) (because there's zero consistency so intuition is impossible) sometimes it's a pleasant surprise. last time I touched php I wanted to iterate through a couple arrays at the same time. the massive hierarchy of "iterator" classes were a dead end, but turns out every "array" has its own "current index", which was very convenient! Share Bear posted:maybe more of a confessions thing but i honestly dont know why oop took over as the dominant programming paradigm, I think the plan was for artisanal worksmiths to handcraft libraries of objects (not "code", objects) for you to build upon. idgi but it sounds nice
|
# ¿ Apr 14, 2023 16:16 |
|
cinci zoo sniper posted:auto reject on formatting errors, but reject on linting errors is a topic for 2024 what's the difference? or is this a language "without a compiler" so you're using a linter
|
# ¿ Apr 15, 2023 05:20 |
|
susan b buffering posted:the op is using python op oops I missed that
|
# ¿ Apr 15, 2023 06:57 |
|
Its a Rolex posted:i also set up some CI that managed dependency versioning, and nobody wants to use it because they don't like that PRs are opened by a bot which tell them to set compatibility bounds. every other week i am putting out a fire where a public package releases a breaking change, and the 0 compatibility bounds in place mean that breaking release gets included automatically in builds and breaks an environment (most recently this made its way to prod) I hate this poo poo too. your "compatibility bounds" don't mean anything (unless you're writing elm) pin exact or nothing
|
# ¿ Apr 16, 2023 18:48 |
|
"semantic" "versioning" is for suckers
|
# ¿ Apr 16, 2023 19:16 |
|
cinci zoo sniper posted:serious question: what's the case against semantic versioning? any big brain blog posts (medium brain is fine too) i'll also appreciate, if there are such unless it's enforced by tooling (e.g. elm) it's entirely up to the whims of the maintainer. "oops should've bumped major, not minor, sorry" and even if you do it right for public api, there's all kinds of implementation details you can change without changing the "public api" that nonetheless fucks with library users. there's the usual dumb poo poo like scraping error messages that you could blame on the caller holding it wrong. but esoteric poo poo like "now we round down instead of up" is probably not gonna be thought of as public api it's probably good for maintainers to think about the changes they're making and how to describe them to library users. but unless you trust all 1-50000 (depending on ecosystem) maintainers you welcome into your project, and they never make a mistake, you're better off pinning exact versions
|
# ¿ Apr 16, 2023 19:28 |
|
I mean, I tend to specify a version range of "up to next major" and use a lockfile to avoid surprise bumps, which sounds like what you're doing? and I don't have a ton of transitive dependencies to worry about, so there's a small enough set of libraries involved that when I wanna bump something that I can glance over their changelogs first. it's not perfect, and would be much more difficult in node or wherever to keep up with all that. and I still miss things that weren't in the changelog but I draw the line at having a bot automatically bumping things
|
# ¿ Apr 16, 2023 20:35 |
|
Armitag3 posted:Thats throwing the baby out with the bathwater imo, I want security fixes and minor performance improvements to my deps without having to janitor for it. If someone fucks up versioning it’s infrequent enough of a headache that I can deal with it and roll back then use elm
|
# ¿ Apr 16, 2023 21:17 |
|
I am 80% joking about elm. I do like how its package manager rejects library updates that break/add public api without a major/minor version bump. only solves the "broke the build" problem though, not the rest of it and yeah I consider it a feature that giving myself more work to janitor dependencies means I am less inclined to depend on things
|
# ¿ Apr 16, 2023 21:36 |
|
for awhile I had a checked-in custom dictionary for android studio spellcheck but it seems to have stopped working at some point
|
# ¿ Apr 17, 2023 00:21 |
|
one of the nerds at language log reposts a webcomic like once a fortnight when it has something to do with linguistics. xkcd comes up like every tenth such post, and they include the alt text so I don't even have to remember to check that. it's about the right level of exposure for me still listen to weezer once in a while too!
|
# ¿ Apr 17, 2023 17:21 |
|
if you think about it, = is the ultimate "I really really mean it" equality operator
|
# ¿ Apr 18, 2023 19:41 |
|
cool av posted:changed a string constant. now to fix 8 tests that all check that that constant equals to a second copy of the constant, defined in the test. joke's on you, thanks to string interning it's not even a copy
|
# ¿ Apr 20, 2023 01:42 |
|
last time I touched mysql I couldn’t figure out how to atomically create things (tables, indexes) because create statements seemingly weren't rolled back on transaction failure/rollback. did I miss something or is that how it is? (the recent discussion reminded me of this, but I'm not trying to enter the discussion, just curious about mysql)
|
# ¿ Apr 22, 2023 17:22 |
|
it's also nice for grouping bytes in binary literals
|
# ¿ Apr 30, 2023 19:34 |
|
FlapYoJacks posted:Rust is the best-compiled language out there. is this like when the oscar for editing goes to the film not with "the best editing" but to "the most editing"
|
# ¿ May 2, 2023 04:11 |
|
oh great now there's a ninth python package manager I gotta learn
|
# ¿ May 3, 2023 17:17 |
|
if you think about it, 1. look up the one true officially blessed python package tool 2. see there still isn't one and pick whatever seems plausible for your use case is technically one, and only one, way to do python packaging
|
# ¿ May 3, 2023 19:05 |
|
oh sorry, by "officially blessed python package tool" I meant "tool that performs the expected functions of a package manager, such as 'resolve dependencies', and is labelled as the one true such tool on python.org". I see how that could be conflated with "tool labelled on python.org as the one true package manager but won't actually do expected functions of a package manager, such as 'resolve dependencies'"
|
# ¿ May 4, 2023 00:52 |
|
CPColin posted:Long ago at Experts Exchange, somebody realized that bulk inserts and updates were making some thing called the "materialized view log" get really big and solved the problem by truncating said log after every few thousand rows. This became standard operating procedure and continued well after that guy quit. Smash cut to our newly hired DBA years later looking at our materialized views and going "what in the absolute gently caress". lmao
|
# ¿ May 4, 2023 00:53 |
|
odd that the thread called "terrible programming" is sometimes down on the concept as a whole
|
# ¿ May 5, 2023 02:41 |
|
CPColin posted:The only python I like is the one in my pants there should be one, and preferably only one, way to do it
|
# ¿ May 5, 2023 17:01 |
|
gnatalie posted:thinking of converting a json rest api to grpc cuz it needs to send out a bunch of data to hundreds/thousands of devices with extremely spotty internet connections. what does grpc give you that you can't do with json over http?
|
# ¿ May 7, 2023 04:13 |
|
and if you still want it, you could put protobuf in your http response instead of rewriting everything
|
# ¿ May 8, 2023 18:45 |
|
love to use undocumented implementation details for my secure randomness
|
# ¿ May 16, 2023 23:24 |
|
if they don't want the docs, edit your issue op to contain the docs you wrote. then it'll show up in searches and share most of the repo url, so other people can still find it
|
# ¿ May 18, 2023 16:54 |
|
0 days since I last thought "figma balls"
|
# ¿ May 18, 2023 21:20 |
|
|
# ¿ Apr 26, 2024 01:08 |
|
that’s not a compiler warning? yeesh
|
# ¿ May 20, 2023 03:49 |