https://htmx.org logo
Powered by
# htmx-general
  • m
    ah, good ol' inheritance strikes again
  • m
    I wonder if we should log where attributes come from in debug mode
  • m
    I've seen this before
  • m
    it is an example of spooky-action-at-a-distance
  • g
    Is there a way to troubleshoot flickring during a swap? I've got an interaction similar to the edit form example, but my page flickers when the new element is swapped in.
  • w
    Yeah, I really gotta get my inheritance down. I discovered that in an attempt to reduce the number of hx- attributes I use (not for any reason other than less code)
  • g
    Gah...nevermind. I had an ID attached to the wrong div on the incoming template.
  • w
    welcome to the "nevermind" club, it's good to have company 🙂
  • l
    At my first programming job (mainframe-based COBOL 1998-2007), we had a running joke that "you're not a real programmer until you've made the mainframe reboot." When I left, there had been a handful of us. 🙂 Mine maybe shouldn't have counted, but it did; I triggered a known-and-fixed issue, but the operators hadn't loaded the patch because "no one uses that"... LOL
  • l
    Most of the time, it's your code. On occasion, you find something in "theirs". The vast majority of the time, though, it's... "Oh...." (shuffles off quietly)
  • g
    It always happens as soon as you type the question. I need a duck on my desk.
  • b
    I did CICS/DB2 for a couple of years around 2001. My interface design skills have not improved much.
  • l
    80x24 FTW!
  • b
    The trendy team was the web team - they wrote Java that screen scraped the CICS and rendered it as html. 😂
  • l
    Unisys had a similar thing for their green-screens. One of my "whoa, you can do that?" moments was when I created a transaction that returned XML, then loaded an XSLT with it. They ended up implementing that technique for lookups (what we would call autocomplete now, except you had to hit a button to get results) throughout the system. As of a couple of years ago, they were still using that interface! (They're going to make us move this to #geezer-chat, aren't they? LOL)
  • l
    The interesting thing is that htmx would work really well in that environment! Return HTML fragments, and let the fun begin!
  • b
    It looks like you can create a github gist that has the html, js, css and a manifest file that loads external libs. Doesn't look like there's a way to do mock responses though - you can echo things back from special URLs but that won't really help us
  • b
    I had similar issue before. My issue was not implementing 
    Access-Control-Expose-Headers
    on the response in my server. https://htmx.org/docs/#cors
  • b
    So confusing that the you can see the header in the Network tab in dev tools, but JS can't see the header value
  • f
    Thanks – saw you bringing it up when I was browsing through chat logs for possible solutions. My issue was the event not propagating as expected when using a modal.
  • f
    Any best practices or ideas for polling the backend for new data? It feels like my idea is overly complex: Register an event listener on the parent div 
    hx-trigger="myEvent"
    and then poll an endpoint from an (invisible) child element using 
    hx-trigger="load delay:1s"
    . The endpoint just returns 
    HX-Trigger = myEvent
    (and nothing else) when new data is available so the div refreshes. 🤔
  • f
    Could also use hyperscript's 
    fetch
    to allow for more flexibility and raise the event manually when the header is set, but that ultimately seems like a reimplementation of the 
    HX-Trigger
    header.
  • b
    That sounds similar to load-polling from the docs: https://htmx.org/docs/#load_polling
  • b
    Your other option is to use something like SSE (if your backend supports it)
  • f
    Ah seems like I left out one critical detail. The content I want to refresh includes forms, so I cannot just keep on polling when nothing changed.
  • b
    Ah - so users could have entered info that you don't want to be cleared on the refresh?
  • f
    It could become frustrating from a UX-perspective if the forms are refreshed every 1s-5s – effectively preventing the user from making any changes. When new data is available however I think it's acceptable that the forms are refreshed and possible input lost in favor of consistent data.
  • f
    What would be cool is if htmx could only refresh the diff 🤔
  • b
    what are we? unpoly?
  • b
    (I believe that diffs whole pages and only updates what's needed)
1...547548549...1146Latest