|
Ciaphas posted:i know literally absolutely nothing about javascript and js related technologies like react javascript is okay for a certain range of problems and applications, but a lot of people don't have the sense to determine whether js is right for the thing they want to use it for react is good, most other js related technologies are bad it's probably good to have at least a basic understanding of it, since the web is everywhere and it's probably never going away. it's a huge pain to decipher which tools actually solve good problems and which ones are stupid pointless poo poo, though
|
#
?
Nov 1, 2017 22:08
|
|
|
|
| # ? Jun 8, 2024 01:39 |
|
|
cinci zoo sniper posted:
lol its all the worst stuff together in one stack!
|
|
#
?
Nov 1, 2017 22:09
|
|
please dont
|
Main Paineframe posted:javascript is okay for a certain range of problems and applications, but a lot of people don't have the sense to determine whether js is right for the thing they want to use it for javascript is suitable for: opening a popup window closing the popup window that's about it.
|
|
#
?
Nov 1, 2017 22:09
|
|
|
Shaggar posted:javascript is suitable for: what about a button on my homepage that changes the background color?
|
|
#
?
Nov 1, 2017 22:29
|
|
|
unacceptable.
|
|
#
?
Nov 1, 2017 22:31
|
|
|
i liked to add snow flakes in winter
|
|
#
?
Nov 1, 2017 22:32
|
|
|
a very small horse posted:so a question i guess: every computer programmer person who's not an academic I know makes some sort of web service thing. so there are other things out there. working for a games company sounds hard and not that great by some accounts. so what else? embedded is a thing that sounds cool but what are my options if I don't want to be a web applications programmer for the next 10 years and how do I do that thing? there's a lot of ETL work out there, but it's trendy to try and microservice it up too which means all the same garbage as webservices
|
|
#
?
Nov 1, 2017 22:52
|
|
|
GenJoe posted:weird esoteric blow ups across different configurations do happen but when they do it's nice to know that you have a large enough community using your poo poo that there's a significant chance someone else has run into said blow-up and has found a solution before you waste time on it, there are a dozen different issues for that problem on google's first page for "docker npm update" i mean it's literally the long term support release. and it doesn't work without a workaround that amounts to `don't use this ever` how to reproduce: install the node 6.x lts release, type `npm install npm@latest` the talent deficit fucked around with this message at 23:05 on Nov 1, 2017 |
|
#
?
Nov 1, 2017 23:00
|
|
|
the talent deficit posted:there's a lot of ETL work out there, but it's trendy to try and microservice it up too which means all the same garbage as webservices ETL is a pretty good starting point for a dev career cause its important but conceptually easy.
|
|
#
?
Nov 1, 2017 23:05
|
|
|
ETL can be a hugbox where the walls are made of comfortable db schema, or a hell pit of application based dynamic torment
|
|
#
?
Nov 1, 2017 23:06
|
|
|
Shaggar posted:ETL can be a hugbox where the walls are made of comfortable db schema, or a hell pit of application based dynamic torment What do you mean by OR. It can be both at the same time depending on the job.
|
|
#
?
Nov 1, 2017 23:37
|
|
|
the talent deficit posted:i mean it's literally the long term support release. and it doesn't work without a workaround that amounts to `don't use this ever` npm is a separate thing from Node and the 6.x lts release, and while yes it would be really cool if we could live in a world where no software was broken ever, this is an issue that's limited to an (albeit popular) configuration, and it's extremely easy to find and implement the workaround for it -- if npm the org has any kind of restraint on developer resources then there are probably other things out there that are more attractive candidates for patching. like there are a lot of things to gripe on re: npm but this particular one is an issue that just kind of happens in software sometimes. it's not a good look for npm but it's not something you can point to and go SEE JAVASCRIPT BAD
|
|
#
?
Nov 1, 2017 23:56
|
|
|
Shaggar posted:javascript is suitable for: also posting stuff and injecting the rendered result into the page using partial views in .Net because that poo poo ruuuulleees edit:and manipulating validation rules when you change poo poo on page to do dynamic validation
|
|
#
?
Nov 2, 2017 00:04
|
|
|
also I might be literally satan here but I detest popups and instead use modals if you open something you better do something with it, don't just ignore the popup and carry on doing whatever the gently caress.
|
|
#
?
Nov 2, 2017 00:07
|
|
|
GenJoe posted:npm is a separate thing from Node and the 6.x lts release, and while yes it would be really cool if we could live in a world where no software was broken ever, this is an issue that's limited to an (albeit popular) configuration, and it's extremely easy to find and implement the workaround for it -- if npm the org has any kind of restraint on developer resources then there are probably other things out there that are more attractive candidates for patching. you're not listening. it's the npm bundled with the node lts release. it's not a separate thing
|
|
#
?
Nov 2, 2017 00:09
|
|
|
"our package manager is unable to install this package" "eh, no big deal. people should just use the workaround of installing a totally different package manager and using that instead."
|
|
#
?
Nov 2, 2017 00:11
|
|
|
its a good solution to npm problems though 
|
|
#
?
Nov 2, 2017 00:16
|
|
|
is etl just "data comes from somewhere, gets transformed, then output somewhere else" with some bells and whistles?
|
|
#
?
Nov 2, 2017 00:29
|
|
|
it's a common kind of problem, but there's also absolutely no reason not to fix it
|
|
#
?
Nov 2, 2017 00:34
|
|
|
Powerful Two-Hander posted:also I might be literally satan here but I detest popups and instead use modals I don't like popups or modals but popups are all javascript was designed for.
|
|
#
?
Nov 2, 2017 00:36
|
|
|
redleader posted:is etl just "data comes from somewhere, gets transformed, then output somewhere else" with some bells and whistles? yeah. you extract data from somewhere, transform it in some way, and then load it into some other thing spark is a good starting point if you're interested in learning more
|
|
#
?
Nov 2, 2017 00:38
|
|
|
redleader posted:is etl just "data comes from somewhere, gets transformed, then output somewhere else" with some bells and whistles? yeah its extract, transform, load. irl its about taking the terrible garbage poo poo trash your client sends you and contorting it to suit your needs. its also about handling the many failures that will occur when the client sends the wrong data, sends invalid data, doesn't send data, or any of 100 other gently caress ups they'll explore. when the client wants to know why X didn't happen because it was definitely on the file load, you are the one that gets to explain to them how it wasn't actually on the file load and heres the client's file that they sent that clearly doesn't have X on it. but you get to use nice tools like SSIS which makes things easy
|
|
#
?
Nov 2, 2017 00:42
|
|
|
someone complained about "bad data" in a file breaking their loader the other day and it turned out their 'loader' was some sort of script that took a pipe delimited file and converted it to csv by removing all quoted thingies and the replacing all pipes with commas so an input of "foo" |"bar," | became foo, bar,, and it all imploded idk where the gently caress people get poo poo like this from edit: Dtserror:success
|
|
#
?
Nov 2, 2017 00:52
|
|
|
the talent deficit posted:you're not listening. it's the npm bundled with the node lts release. it's not a separate thing it's completely different software and it's versioned and patched separately from 6.x -- it's a separate thing that the core node team has almost nothing to do with. when you're triaging issues, it's not uncommon to have to sigh and say "okay well yeah we're broken for this use-case on docker, but on the other hand our reputation is getting massively hosed because people keep on maliciously taking over repositories and poo poo, so we're kind of going to have to let that one slide and build out 2fa etc to save our failing heap of an organization" if the issue were that installing /any/ package on docker failed then yeah it would be inexcusable if this wasn't patched like the next day
|
|
#
?
Nov 2, 2017 00:58
|
|
|
like clearly npm has a history of loving up in spectacular ways but let's at least point out those cases instead of hand wringing relatively minor bugs that are isolated to specific configurations
|
|
#
?
Nov 2, 2017 01:00
|
|
|
node.js is trash and npm is a dumpster firePowerful Two-Hander posted:someone complained about "bad data" in a file breaking their loader the other day and it turned out their 'loader' was some sort of script that took a pipe delimited file and converted it to csv by removing all quoted thingies and the replacing all pipes with commas so an input of "foo" |"bar," | became foo, bar,, and it all imploded from oracle dbas most likely
|
|
#
?
Nov 2, 2017 01:09
|
|
|
tab separated value is a decent way to store flat file data no escaping to worry about in non-degenerate cases so you can whip up a parser for it in like 10 loc, excel/libreoffice understand it, and it's also the native text i/o format for postgres in case you want to quickly copy out a dataset from the psql tool or you want to firehose data into the db faster than a prepared INSERT
|
|
#
?
Nov 2, 2017 01:14
|
|
|
escaping is explicitly for the degenerate cases, if you don't have to worry about escaping then sure go ahead and use whatever character you want as field/record separator. but in the general case, gently caress off with that poo poo and use rfc-style csv, postgres can read it natively too, and at least then the escaping rules are written down somewhere that both you and your client can access
|
|
#
?
Nov 2, 2017 01:17
|
|
|
there's an ascii code for record separator, everyone should use that
|
|
#
?
Nov 2, 2017 02:34
|
|
|
GenJoe posted:it's completely different software and it's versioned and patched separately from 6.x -- it's a separate thing that the core node team has almost nothing to do with.
|
|
#
?
Nov 2, 2017 02:34
|
|
|
necrotic posted:If we push them all to mongo/js id call it a win someone talk all the JavaScript people into replacing git on their projects with a bespoke system implemented in Node.js against MongoDB that way eventual consistency would mean eventually we wouldn’t have to care about them
|
|
#
?
Nov 2, 2017 02:58
|
|
|
CRIP EATIN BREAD posted:there's an ascii code for record separator, everyone should use that there are so many good ascii codes that nobody ever uses. sad!
|
|
#
?
Nov 2, 2017 03:01
|
|
|
I’m vertical tab
|
|
#
?
Nov 2, 2017 03:02
|
|
|
hive uses an ascii control character as a record separator but they used \001 ('start of heading') instead of the actual record separator char 
|
|
#
?
Nov 2, 2017 03:04
|
|
|
I only try to put aside my unreasonable stubbornness and write JS once every blue moon, so I've got to ask are you still on your own in JS land if you want to parse the query string from a URL? I assumed this was a standard old as time but when last I attempted to use a query string it turned out the answer was actually a wild west of 'go loving code it yourself' with fifteen answers on SO all arguing about how to do it best. Despite my project already having a bunch of libraries like jquery, underscore, and react all as hard requirements to even get a hello world page going, none of them supporting var value = window.url.query['key']; or something was frustrating enough for me to again give up and retreat down further into the stack. edit: admittedly though I am a terrible programmer.
|
|
#
?
Nov 2, 2017 04:53
|
|
|
You could use URLSearchParams, though you'll need a polyfill for IE.
|
|
#
?
Nov 2, 2017 05:12
|
|
|
it's listed on mdn as an experimental technology?? https://developer.mozilla.org/en-US/docs/Web/API/URLSearchParams/URLSearchParams
|
|
#
?
Nov 2, 2017 05:17
|
|
|
Whoops, missed that, I thought it was more widely supported. Though I suppose you could also use a polyfill with the assumption it'll eventually be supported, like with window.fetch.
|
|
#
?
Nov 2, 2017 05:22
|
|
|
redleader posted:is etl just "data comes from somewhere, gets transformed, then output somewhere else" with some bells and whistles? yes. basically you take something like pentaho that is mostly agnostic to input type and output type and fix your poo poo instead of having random scripts all over the loving place
|
|
#
?
Nov 2, 2017 07:33
|
|
|
|
| # ? Jun 8, 2024 01:39 |
|
|
if you want some incoming data horror stories, we have a provider that insisted on an nda before they could send their data to us as xml, rather than rt-loving-f
|
|
#
?
Nov 2, 2017 07:34
|
|