https://htmx.org logo
Join Discord
Powered by
# htmx-general
  • m
    my default setup has always been: - domain logic executing on homogenous app servers, broken up by whatever modularization mechanism the platform supports - a simple job runner for async stuff - macro-services for things that stand on their own (e.g. email delivery)
  • m
    at one point leaddyno (a basic affiliate management system) was being used by a company trying to do youtube analytics because we were handling loads that google analytics barfed over. It was just a rails app talking to a postgres db on the app side and a sinatra app jamming stuff in a sidekiq work queue on the event ingress side. Aside from cloud misconfigs, it worked great
  • m
    Copy code
    [ app ] <-> [ job queue ]
  • m
    this is my political beliefs
  • c
    I heard about that in "Java-land". Basically means pretend you have hard borders between different parts of your application (modules). So what microservice-guru would put on different microservices you just put in different modules and make sure you expose a tiny API and hide all the implementation details
  • c
    it helps to hide large amounts of code and makes the system easier to reason with and avoids the network calls and other complexity of microservices. etc
  • c
    even database can be in scope here btw. to avoid multiple dbs just be strict with the schemas : each module has a schema, no joins or other funny business between schemas
  • c
    if a module really needs to it can be extracted and deployed as a service/microservice
  • c
    I haven't had the chance to work with this stuff but I tried some versions in pet projects. I think it can be challenging with getting a team on board with it and keeping it as it should be
  • m
    I have a question about hx-sync. I have a list of elements and I only want to send a request for the latest element to intersect the viewport. So I tried to put a delay on the intersection trigger and a sync of abort. It didn't seem to work... Does hx-sync abort a delayed request?
  • s
    Thanks, i'll try it out tonight and let you know.
  • w
    I've been thinking about a social network, where the "posts" are snippets of html. Here's a blog post I just wrote up with some interactive examples (https://www.sambutler.us/html-as-post) Is this worth exploring? Anybody interested in exploring it together?
  • g
    cohost.org has some html support i think
  • g
    are you planning on sanitizing the html or serving it on a separate origin?
  • m
    replace will replace an in-flight request
  • m
  • m
    Is there anyway to abort a request that is currently waiting on a delay via sync?
  • m
    no, they would all pile up
  • m
    it sounds like you might want to write a bit of custom JS and trigger your own event
  • r
    wouldn't "htmx:abort" event work?
  • m
    you could use 
    htmx:abort
    to kill a current request
  • m
    but the events are queued. Are you trying to debounce things?
  • r
    oh, got it
  • m
    let me think about it some more
  • m
    so you want only the last revealed element to trigger a request
  • m
    can you explain that more?
  • m
    seriously considering keeping IE support in htmx 2 out of sheer obstinacy (please don't argue w/ me it will only make things worse) https://twitter.com/htmx_org/status/1592543594249961474
  • m
    heres an introduction
  • m

    https://www.youtube.com/watch?v=5OjqD-ow8GE▾

  • t
    noice. watching
1...906907908...1146Latest