|
The wand chooses the wizard and who are we to question the fate of others.
|
#
?
Dec 19, 2020 22:59
|
|
|
|
| # ? Apr 28, 2024 11:25 |
|
|
AlphaKeny1 posted:lol thanks, I have "research vue npm components" on my list of things to do for my team and it looks like I'll avoid that
|
|
#
?
Dec 21, 2020 14:41
|
|
|
Got a question. I'm not looking for all the answers, just some help. I want to change my career, from SEO to front end dev. I have worked on websites for 3 years and know the basics of HTML, CSS, and some JavaScript, plus a lot of WordPress stuff like configuring advanced custom fields plugin. My question is the best way to research front end web development bootcamps and choose an option. I am overwhelmed by all of the options.
|
|
#
?
Dec 23, 2020 17:33
|
|
|
blue squares posted:Got a question. I'm not looking for all the answers, just some help. i will never not shill https://www.freecodecamp.org/
|
|
#
?
Dec 23, 2020 18:15
|
|
|
Inacio posted:i will never not shill https://www.freecodecamp.org/ I went through FCC a few years ago and was very happy with the content. A+ would recommend to anyone.
|
|
#
?
Dec 23, 2020 18:40
|
|
|
Inacio posted:i will never not shill https://www.freecodecamp.org/ Thanks! I'm working through the HTML section now. I already know almost all of it, but its good to start at the beginning to make sure I don't miss anything.
|
|
#
?
Dec 23, 2020 18:47
|
|
|
For a NextJS app that is highly DB driven and doesn't require any SEO (a business tool where everything is behind a login) and doesn't need to handle huge scale, should I just use SSR (getServerSideProps) for everything to keep things simple, or are there other approaches to consider? For some data/pages I'm thinking doing static generation (getStaticProps) with a revalidate value set could work (pages that show the same data for every user with data that isn't likely to be updated that often), but still in rare cases could open the door to more sync related errors that would need to be handled (and/or provide a poor user experience), making the code more complex for what may be minimal benefit (as this won't be a high traffic site). Another approach I could see is to have some pages that need dynamic data to have skeletons statically generated server side, then fetch the data (or at least some of the data) client side via an API server. In some cases the client side code for fetching data from the API server may need to exist anyway, so this wouldn't necessarily be extra work/complexity in some cases, but it other cases it would be. Finally, is there a Next approach for not loading all dynamic data on a page at request time when doing SSR, or is that something that should just be handled via client side code? For example, an item detail page is going to have multiple tabs to show different sets of data related to that item. At first load of the detail page, only the initially visible tab really needs to have its data populated. Data for the non-visible tabs could be fetched in the background after the initial request. Off the top of my head I'm thinking the way to accomplish this in Next would be to have each tab be a separate route page, with the tab headers being 'next/link's, as I from the documentation I believe this will result in Next prefetching all the non-visible tab content. Will that accomplish what I'm expecting (with no hit to UX)? Might just end up fetching all needed data at request time at least initially, but I want to know what my options are if that ends up being too slow.
|
|
#
?
Dec 23, 2020 21:25
|
|
|
if you dont need seo, what do you need SSR for? (havent used Next much)
|
|
#
?
Dec 23, 2020 21:32
|
|
|
Inacio posted:if you dont need seo, what do you need SSR for? (havent used Next much) I don't need SSR, but it essentially comes free w/ Next (and if anything makes SSR less complex code wise than CSR) and I believe still has other benefits outside of SEO (e.g. less code sent to and run on the client), so my thought is why not use it.
|
|
#
?
Dec 23, 2020 22:14
|
|
|
There are performance benefits for SSR, sure, but there are some tradeoffs like larger pages (by a few KB) and a (more than likely unnoticeably) delayed TTFB. There are so many factors that go into site performance these days, I think it's a wash if you don't really need the SEO benefits of SSR, really.
|
|
#
?
Dec 23, 2020 22:30
|
|
|
If everything is behind a login and especially if it's a b2b app you probably don't need to be assed about squeezing the tiniest bits of performance out of your app. Like yeah if it starts to feel sluggish your users will complain and eventually the decision makers might switch to a competitor but it's not like large-scale public-facing stuff where every 25ms you add to page load times can increase your bounce rate by a significant number. If you're gonna have the kind of slowness that users actually perceive and are annoyed by it's probably going to come from database operations.
|
|
#
?
Dec 23, 2020 22:35
|
|
|
I’ve definitely seen people just use SSR for everything and it’s not a great experience if you have typical slow enterprise APIs. People expect rendered spinners now, not the browser to just “hang” after clicking a link for five or ten seconds.
|
|
#
?
Dec 23, 2020 23:26
|
|
|
prom candy posted:If everything is behind a login and especially if it's a b2b app you probably don't need to be assed about squeezing the tiniest bits of performance out of your app. Like yeah if it starts to feel sluggish your users will complain and eventually the decision makers might switch to a competitor but it's not like large-scale public-facing stuff where every 25ms you add to page load times can increase your bounce rate by a significant number. If you're gonna have the kind of slowness that users actually perceive and are annoyed by it's probably going to come from database operations. This is what I'm thinking and why I've been leaning toward just using getServerSideProps() (SSR) for everything initially (as it seems like this is the quickest approach development time wise) and only worrying about moving some fetching to the client if any issues actually come up in use. SWR (vercel's hook for client side fetching) was just brought to my attention, so I may look into that too as it seems like it takes care of much of the additional code that would come with fetching everything client side. smackfu posted:I’ve definitely seen people just use SSR for everything and it’s not a great experience if you have typical slow enterprise APIs. People expect rendered spinners now, not the browser to just “hang” after clicking a link for five or ten seconds. If we were talking long page loads like that I'd definitely be looking into a different approach, but that's not what I'm anticipating (we're building the API/DB as well). I should note that at least with NextJS, using SSR for everything doesn't prevent showing an app based loading indicator or progress bar while a new page is being rendered server side. I'm still obviously new to Next, but my understanding is that while it can render pages server side, it still has a client side component that is doing some magic to make routing and page transitions feel more like an SPA than say an old school backend MVC style webapp (e.g. I don't think it ends up re-rendering parts of the new page that are identical to the old page, but I could be wrong).
|
|
#
?
Dec 24, 2020 00:13
|
|
|
Something that used to work is a notification would come in from the connected socket and it'd trigger a notification, but this doesn't work anymore without user interaction? https://developer.mozilla.org/en-US/docs/Web/API/notification quote:Note: In the above example we spawn notifications in response to a user gesture (clicking a button). This is not only best practice — you should not be spamming users with notifications they didn't agree to — but going forward browsers will explicitly disallow notifications not triggered in response to a user gesture. Firefox is already doing this from version 72, for example. Notifications stopped showing up on mobile too. No, it's not a permissions thing. Is there any way to get this to work anymore, or are pressures that be making me switch to google's desktop push notifications thing.
|
|
#
?
Dec 24, 2020 14:50
|
|
|
The Merkinman posted:Well, the 4th time I tried it worked. Though I feel the import is a little odd. Hey, thanks for this! I saved it to look at after my deadlines for the first few weeks of Jan. Dunno if you already said it but may I ask what your use case is? For my team I'm trying to find ways to import our own components into other team's repos and still have access to their vuex store, ideally so that we don't have to follow their release schedule or anything. I'd prefer not to build our components directly into their code and deal with so much bureaucracy. From my limited understanding, npm component libraries is my answer. Am I on the right track?
|
|
#
?
Dec 26, 2020 08:47
|
|
|
AlphaKeny1 posted:Hey, thanks for this! I saved it to look at after my deadlines for the first few weeks of Jan. Yeah. We have a project that is the style guide that a number of other projects use. Originally this was just the CSS, but over time it involved functionality so, like you said rather than "build our components directly into their code", it's in a separate project. Of course you'd have to be aware of npm versioning to make sure the other teams are using the version you'd want.
|
|
#
?
Dec 27, 2020 21:42
|
|
|
Excellent, thanks!
|
|
#
?
Dec 28, 2020 00:49
|
|
|
Does anyone here use a nice headless CMS? I'm looking for something that is not wordpress-like, but insteawd allows me to create my own schema via GUI. I've tried Strapi, but the Digital Ocean instance crashed and burned while I was playing with content types. What is similar, but maybe more stable? SaaS solutions work too. I just need a fine grained control over the input fields and data relations, and Strapi seemed to be doing that fine. About to investigate: - agilitycms - prismic - umbraco - buttercms Any pointers would be greatly appreciated.
|
|
#
?
Dec 28, 2020 02:33
|
|
|
gbut posted:Does anyone here use a nice headless CMS? My employer has had some big sass success with Contentful, but our applications are only reading from it, with the content team using Contentful’s interface for authoring.
|
|
#
?
Dec 28, 2020 04:37
|
|
|
I mentioned earlier that I am looking at switching from SEO to Web Development. I'm sure a lot of you work in some capacity with SEOs. Any wisdom to impart? Things you wish you knew when you were getting started? Good websites to read to learn more about a career in web dev? I'm doing Free Code Camp, and as I encounter terms I am not familiar with I am doing supplementary research and taking notes on paper blue squares fucked around with this message at 13:33 on Dec 28, 2020 |
|
#
?
Dec 28, 2020 12:53
|
|
|
I learned everything I know trying to turn my dumb ideas into reality.
|
|
#
?
Dec 28, 2020 14:34
|
|
|
If you have already worked in the industry, you are already way more competent than any college new hire.
|
|
#
?
Dec 28, 2020 14:38
|
|
|
Free Code Camp is excellent, but I did have friends go through it recently and felt like parts of the curriculum was either a bit dated or not particularly relevant to my day-to-day. Don't break your brain trying to master recursion in JavaScript or floats in CSS. Once you have the foundations from FCC, I generally recommend consuming Wes Bos-related things, particularly his (free) JavaScript 30 class and the back catalog of his Syntax podcast. I think his teaching style is pretty universally liked, but there are thousands of other online video instructors you can also try to discover via YouTube, Udemy, Lynda, or more premium sites like Level Up Tutorials, Front End Masters, and egghead. Finally, one of the bigger obstacles I encountered before I formally broke into the industry was trying to learn sone of the supporting technology. Learning "just" HTML, CSS, and JavaScript is probably inadequate for most jobs, unfortunately. Make sure to take some time to learn about terminal commands, Git, npm, bundlers, and at least a high level overview of a JS framework like React.
|
|
#
?
Dec 28, 2020 15:49
|
|
|
Summit posted:I learned everything I know trying to turn my dumb ideas into reality. Yes, don't spend too long in "learning" mode, you need a healthy mix of that along with just trying to make poo poo and figuring out why it doesn't work.
|
|
#
?
Dec 29, 2020 06:23
|
|
|
Is there any fast command you can run that checks that the node_modules directory is in sync with the package-lock file? Working in a team, it’s easy to pull down a package-lock update and not realize it, and then be running with old package versions accidentally. npm ci fixes it but is a bit brute force and takes 30-60s.
|
|
#
?
Dec 30, 2020 18:04
|
|
|
smackfu posted:Is there any fast command you can run that checks that the node_modules directory is in sync with the package-lock file? rm -rf ./node_modules && npm install  Bonus: you get a free coffee break every time you run it!
|
|
#
?
Dec 30, 2020 22:58
|
|
|
This. Npm is trash, and you should never trust it. Nor yarn. E: or force the team to use nix packages
|
|
#
?
Dec 31, 2020 00:03
|
|
|
gbut posted:This. I don’t see how this is an npm problem
|
|
#
?
Dec 31, 2020 16:28
|
|
|
Npm, while improving, still feels unfinished. It took ages to add some sort of reproducible builds, and the whole ecosystem is a bit of dumpster fire--remember the whole leftpad fiasco? Having it be my main tool for the job makes me a bit jaded, so I tend to hate on it. And there's that relatively recent npm inc poop-soup and that left a bitter taste in my mouth, metaphorically speaking.
|
|
#
?
Dec 31, 2020 21:45
|
|
|
The primary reason I'm excited about https://deno.land/ is to get away from npm. I've never had any project I've worked on go smoothly with npm, not even small projects. Big projects forget about it, have it pop up in nightmares instead.
|
|
#
?
Jan 1, 2021 13:20
|
|
|
Yarn solved the reproducible build thing for my team, and yeah npm itself (the tool, not the ecosystem) has a number of flaws. There are for sure issues with node's package management but I don't think any package manager has a perfect solution yet. Rust's solution is quite similar to node's, and deno's bigger standard library is nice but you'll still eventually need to import external modules, at which point you have the same problem as npm users.
|
|
#
?
Jan 1, 2021 17:30
|
|
|
Other build platforms like Gradle seem to do a decent job of making sure your dependencies match what you specify before running so doesn’t seem like any technical reason npm couldn’t do it. Come to think of it, Webstorm does tell you if the package.json doesn’t match what is installed but no obvious way to automate that.
|
|
#
?
Jan 1, 2021 17:36
|
|
|
Sure, you could do that by making a prebuild npm script that rimrafs node_modules and installs deps. Npm isn’t a build system so I’m not sure why you would expect that out of the box.
|
|
#
?
Jan 1, 2021 18:28
|
|
|
I mentioned “npm ci” in my original question, which is a one-step faster version of what you describe. It’s just slow, and I only want to do it if necessary.
|
|
#
?
Jan 1, 2021 18:47
|
|
|
On another topic, does anyone have any good YouTube subscriptions for front-end stuff? Subscriptions so I can just see new videos come up rather than searching for one off videos.
|
|
#
?
Jan 3, 2021 20:04
|
|
|
I've got a page with a persistent websocket connection to the server. On mobile if I navigate away from my browser it'll disconnect eventually. When I come back, since it's configured to, it'll eventually reconnect which is all fine. Is there a way to detect that I've come back, and have it try to reconnect immediately?
|
|
#
?
Jan 4, 2021 17:09
|
|
|
In Cordova land the Internets say to use the “pause” and “resume” events. Normally you could probably try “visibilitychange” per MDN and test for visibilityState becoming visible and hope that it is reliable.
|
|
#
?
Jan 4, 2021 20:05
|
|
|
Ok, so say I have a context in React, and I'm using TypeScript. Issue I've got is. If I type the context like this:code:code:code:code:And this is static, mainly primitive, read-only properties. When I have functions held in context it generally gets substantially more complicated and annoying. This is driving me slowly insane, I use context quite a bit and most of the time I use it I spend a half hour or so googling solutions to above issue. Is there a sensible way to do this in TS that I'm just missing here
|
|
#
?
Jan 7, 2021 22:20
|
|
|
Can you mark the FooStuff interface to indicate which fields are optional? Or can you just use default values that conform to the type but are otherwise nonsense (i.e. if it expects a number just pass in -1)? Alternately can you make a custom hook to use your context that coerces the result into a different type? i.e. you initialize the context with FooStuff | null but your custom hook casts the result as just FooStuff? This is just one of those annoying cases where you have to tell typescript "no I know better than you." Normally that's accomplished with the bang operator but obviously you don't want that littering your code. edit: vvv yeah that prom candy fucked around with this message at 23:08 on Jan 7, 2021 |
|
#
?
Jan 7, 2021 22:51
|
|
|
|
| # ? Apr 28, 2024 11:25 |
|
|
RobertKerans posted:Ok, so say I have a context in React, and I'm using TypeScript. Issue I've got is. If I type the context like this: Assuming you're using the useContext hook to get to the context later, I wrap the hook to "fix" the TS declaration. code:
|
|
#
?
Jan 7, 2021 22:52
|
|
