|
Oh my loving god it's just doing it function by function I cannot even step through.
|
#
?
Nov 6, 2013 02:18
|
|
|
|
| # ? Jun 11, 2024 04:58 |
|
|
PleasingFungus posted:I'm the operator that only exists to trip up language newbies & people who miss a character when typing yeah backwards compatibility is a bitch, same w/ the thing under my av it's unpleasant and should be better, but isn't really a big deal in practice. hell jshint even yells at you by default if you use == instead of === any time it could matter
|
|
#
?
Nov 6, 2013 02:19
|
|
|
2banks1swap.avi posted:Oh my loving god it's just doing it function by function I cannot even step through. hell is other people('s code)
|
|
#
?
Nov 6, 2013 02:27
|
|
|
I understand now 
|
|
#
?
Nov 6, 2013 02:33
|
|
|
this is some frontend meltdown over here
|
|
#
?
Nov 6, 2013 02:39
|
|
|
I really really wanna be a back end programmer again.
|
|
#
?
Nov 6, 2013 02:40
|
|
|
hepatizon posted:dependency resolution is always deterministic. dependency configuration -- which Gemfile.lock handles -- is not deterministic. that's why you check it in, because it's a config file. when people talk about "generating" Gemfile.lock, they mean auto-configuring a new application i think we're not speaking quite the same language. when i make a pom, i put my dependencies and their versions in there. when maven reads the pom it resolves my transitive dependencies including (possibly) version conflicts. the result is a dependency graph of artifacts and their resolved versions. this is a deterministic process. bundler must have some equivalent mechanism, and I ask: how can this reasonably be non-deterministic
|
|
#
?
Nov 6, 2013 03:15
|
|
|
2banks1swap.avi posted:I really really wanna be a back end programmer again. do you know java and want to live in nova maybe u can get me a referral bonus
|
|
#
?
Nov 6, 2013 03:15
|
|
|
Nomnom Cookie posted:do you know java and want to live in nova maybe u can get me a referral bonus I'm doing .NET right now and got into it because I did Java so yes I do  Can you wait a few days? I hired that Resume writer duder but didn't pay to have it done extra fast so it'll be done like friday/next monday. Also how close to DC are we talking.
|
|
#
?
Nov 6, 2013 03:47
|
|
|
PleasingFungus posted:I'm the operator that only exists to trip up language newbies & people who miss a character when typing no you are actually the operator that exists so people don't have to write a bunch of dumb boilerplate code. you save everyone tons of time and make code prettier and neurotypicals love you.
|
|
#
?
Nov 6, 2013 03:47
|
|
|
tef posted:i'm sorry shaggar thanks. it will be rough but I think i'll be able to make it.
|
|
#
?
Nov 6, 2013 04:04
|
|
|
2banks1swap.avi posted:why the do people think making the part of a code that a person clicks and looks at is the worst thing ever for my part it's 2 things: i have zero eye for design so i feel like one of those painting elephants anytime i have to make ~*pixels*~ dance. incidentally the only devs i've known that like ui also have a strong design aesthetic. in my experience ui dev more than any other domain seems to have more than its fair share of hacky (or at least unusual and specific) idioms. i think there's something deep about graphics-based user interaction that maps badly to our mental models of computation resulting in every ui framework being some weird sideways thing that doesn't really behave like anything else in programming. maybe it's just me but it always feels like i'm fighting between how to make the ui do a thing vs how i think the ui should do a thing.
|
|
#
?
Nov 6, 2013 04:31
|
|
|
I just don't like how every browser does CSS different, callbacks in general (why use these???) and such. Throw in debugging JS being a pain in the rear end and you have a nice little party of poo poo and failure. Events make enough sense. It's just a loop abstracted away and "if this is true execute more poo poo lol."
|
|
#
?
Nov 6, 2013 05:12
|
|
|
Why are callbacks so loving popular anyway. What's wrong with regular events.
|
|
#
?
Nov 6, 2013 05:22
|
|
|
callbacks go great with exceptions! boy that stack trace sure is meaningful
|
|
#
?
Nov 6, 2013 05:32
|
|
|
oddly enough web ui has been both the most and least nicest ui tech ive used in that html and js are both fine and work well together and css is literally the sloppiest thing ever done with computers "hmm yes keep style and structure separate" *sacrifices 6 divs just to make the box model happy* "why won't this column just loving float???" css is just *just* reaching the point that js reached 5 years ago where frameworks are abstracting away all the poo poo work but you still have to go poking your dick in the markup too often just to make the style engine do its goddamn job Dr Monkeysee fucked around with this message at 05:40 on Nov 6, 2013 |
|
#
?
Nov 6, 2013 05:36
|
|
|
js is not fine and a new js framework comes out every month and they are all poo poo
|
|
#
?
Nov 6, 2013 05:37
|
|
|
jquery just works don't download fad frameworks every week i guess (also avoid most jquery plugins)
|
|
#
?
Nov 6, 2013 05:40
|
|
|
css is permanently hosed cause print design nerds made it
|
|
#
?
Nov 6, 2013 05:45
|
|
|
JewKiller 3000 posted:callbacks go great with exceptions! boy that stack trace sure is meaningful fun fact: promise error handling is a loving nightmare if you go in blind .then seems like an intuitive api but oh man they are fraught with terror (i know everyone immediately tl;drs at a blog post link in yospos but if you do read that make sure to read the first comment below too) spent a ton of time this morning debugging a horrifying sinon+backbone+q.js issue that kept silently failing because i put in a success handler on .done() but not an error handler 
|
|
#
?
Nov 6, 2013 05:46
|
|
|
Nomnom Cookie posted:i think we're not speaking quite the same language. when i make a pom, i put my dependencies and their versions in there. when maven reads the pom it resolves my transitive dependencies including (possibly) version conflicts. the result is a dependency graph of artifacts and their resolved versions. this is a deterministic process. pom = Gemfile.lock. the non-deterministic part is creating the file since you can pick arbitrary version numbers. that's why you have to check them both in hepatizon posted:when i think of a build artifact, i think of something that is generated determinstically. Gemfile.lock is not generated deterministically hepatizon fucked around with this message at 06:00 on Nov 6, 2013 |
|
#
?
Nov 6, 2013 05:58
|
|
|
2banks1swap.avi posted:why the do people think making the part of a code that a person clicks and looks at is the worst thing ever because the tooling and frameworks are terrible? because, unlike say other programming, it isn't influenced by algorithm choice, data structures, but almost entirely dominated by workarounds, hacks, polyfills, and boilerplate. because the asset pipeline, because css frameworks, css preprocessors, because cross browser behaviour, because every junior programmer who has come along has written a series of regular expressions to avoid writing a few extra lines of code, and because of the constant facile separation of code which is mvc. because good ux is hard work, and most of the tools are riddled with assumptions you have to circumvent
|
|
#
?
Nov 6, 2013 10:25
|
|
|
Tiny Bug Child posted:css is permanently hosed cause print design nerds made it nah, css wasn't made by print nerds, what happened is that browser vendors didn't want to implement it. it took many many years for CSS2 to take off
|
|
#
?
Nov 6, 2013 10:29
|
|
|
2banks1swap.avi posted:why the do people think making the part of a code that a person clicks and looks at is the worst thing ever bc ui are a mix of code and artwork and there's no good solution around to properly separate the two especially if there's a big emphasis on aesthetics in vidya games the industry standard is to use flash for the ui just imagine having to interface with action script code made by artists its an absolute mess bc they have zero notion of software engineering as it is outside of their qualifications and yet you do want them to have full control of the action script side of things bc they need to have detailed control on the behavior and animations of everything and it would be hell if they needed to go through programmers to tweak anything
|
|
#
?
Nov 6, 2013 10:35
|
|
|
tef posted:because the tooling and frameworks are terrible? because, unlike say other programming, it isn't influenced by algorithm choice, data structures, but almost entirely dominated by workarounds, hacks, polyfills, and boilerplate. ok, but other than that
|
|
#
?
Nov 6, 2013 11:13
|
|
|
minidracula posted:I'm actually using only a little vendor packaging, but outside of that virtualenv + pip has more or less worked for me so far. This is on both Windows and Linux. i've generally used OS vendor packages, and if i have to use a library, i will either fold it in to the repo, or ensure there are pre-built packages available i don't have to maintain
|
|
#
?
Nov 6, 2013 11:24
|
|
|
tef posted:because the asset pipeline, because css frameworks, css preprocessors, because cross browser behaviour, because every junior programmer who has come along has written a series of regular expressions to avoid writing a few extra lines of code, and because of the constant facile separation of code which is mvc. you should have kept going here; i started to hear "lust for life" playing in the background as you ran through the streets of edinburgh
|
|
#
?
Nov 6, 2013 12:20
|
|
|
Rust developers express their 64-bit privilege the only way they know how: https://mail.mozilla.org/pipermail/rust-dev/2013-November/006314.html
|
|
#
?
Nov 6, 2013 12:30
|
|
|
tef posted:i've generally used OS vendor packages, and if i have to use a library, i will either fold it in to the repo, or ensure there are pre-built packages available i don't have to maintain this is the only sane method of python deployment
|
|
#
?
Nov 6, 2013 14:26
|
|
|
tef posted:because the tooling and frameworks are terrible? because, unlike say other programming, it isn't influenced by algorithm choice, data structures, but almost entirely dominated by workarounds, hacks, polyfills, and boilerplate. and its doubly annoying because most of the development problems were solved by Microsoft in wpf but people are idiots and want to try to make ui work on the browser instead. eventually html will be replaced w/ a xaml-like and javascript will be replaced w/ language agnostic bytecode and everything will be good.
|
|
#
?
Nov 6, 2013 15:22
|
|
|
Shaggar posted:and its doubly annoying because most of the development problems were solved by Microsoft in wpf but people are idiots and want to try to make ui work on the browser instead. eventually html will be replaced w/ a xaml-like and javascript will be replaced w/ language agnostic bytecode and everything will be good. keep going, almost there
|
|
#
?
Nov 6, 2013 15:23
|
|
|
web "languages" are all terrible and as bad as p-langs because they are used and promoted by the same community: hobbyists who want to re-solve the problem every time.
|
|
#
?
Nov 6, 2013 15:25
|
|
|
img-shaggar-was-right.gif
|
|
#
?
Nov 6, 2013 15:28
|
|
|
I fully remember in the mid-2000s when I was doing actual web dev work, the word was 'package all that poo poo server-side' because the bottlenecks were always a) the DB and b) the network. The idea was to make pages load as fast as possible to get better results because any other optimization was worthless. Somehow JS devs made it go the entire opposite way and now it's "make as many loving calls as possible", "render poo poo on the front-end" and probably flies in the way of doing it properly because Ruby-on-Rails masters found out that their framework was a huge god drat behemoth that managed to be the bottleneck of the entire stack when PHP, Perl, or Python and their collective frameworks were never the problem on their own end. Somehow, Ruby on Rails and node.js guys (well, RoR guys now doing JS) decided to turn this around and saw huge benefits, which was probably related to removing the workload from a dog-slow framework with a bad server in front of it, I guess?
|
|
#
?
Nov 6, 2013 15:52
|
|
|
no. if i'm amazon, either i render an entire page on the server for every link with all of the header and footer and navigation junk, or i send you over the raw data for the part of the page that changed, and you apply the view logic yourself. there's less data in the latter case, and it's faster to the user since the navigation won't blank out while their new page loads.
|
|
#
?
Nov 6, 2013 16:00
|
|
|
Shaggar posted:eventually html will be replaced w/ a xaml-like and javascript will be replaced w/ language agnostic bytecode and everything will be good. and they'll be implemented on top of html5 canvases and asm.js
|
|
#
?
Nov 6, 2013 16:01
|
|
|
yeah of course they will. what was I thinking. it cant actually work if its on the web and universal bytecode makes way too much sense.
|
|
#
?
Nov 6, 2013 16:02
|
|
|
but what universal bytecode will you pick
|
|
#
?
Nov 6, 2013 16:04
|
|
|
there is literally no such thing as a "universal bytecode". every bytecode platform makes tradeoffs related to its language design
|
|
#
?
Nov 6, 2013 16:05
|
|
|
|
| # ? Jun 11, 2024 04:58 |
|
|
java or cli bytecode.
|
|
#
?
Nov 6, 2013 16:07
|
|