https://htmx.org logo
Join Discord
Powered by
# htmx-general
  • m
    šŸ˜‘
  • e
    Are you using 
    package-lock.json
    or 
    yarn.lock
    to pin the version?
  • w
    I'm not super familiar with npm, but I do know that 
    package-lock.json
    is being used
  • e
    Okay. In there, you should find 
    htmx.org
    somewhere, and it should specify a version. Is it 1.8.0 or 1.7.0? I'd think 1.8.0, but šŸ¤·ā€ā™‚ļø
  • e
    Assuming it's 1.8.0, the next thing I'd check is whether your code has been recompiled on production, and if the server has been restarted since that.
  • w
    It's listed twice. But doing 
    npm list
    shows 1.8
  • e
    Seeing it twice makes sense. The first 
    ^1.7.0
    says "I want any version above 1.7, but below 2.0" so, 1.7.0 and 1.8.0 are both allowed. (more about that versioning syntax here: https://docs.npmjs.com/about-semantic-versioning) The second screenshot is where it's pinned to a specific version (1.8.0). So when you run 
    npm install
    , it should be using 1.8.0
  • m
  • m
    head merge extension landed in dev just now
  • w
    Yeah, hmm. I'll try running 
    npm intall
    on the prod server. Thanks for your help
  • m
    will cut a 1.8.3 with it soon /cc@square-oyster-36295
  • e
    @mysterious-toddler-20573 any thoughts on this?
  • e
    I can submit a PR if you agree that's an okay change
  • m
    I kinda don't want to move that logic around, I can see it breaking things in strange ways for people
  • e
    Hmm, okay
  • e
    I just feel like "polling + redirect" is a common pattern. eg I have a delayed job running on my server, I want to check if it's done, and when it's done, I want to send the user to a new page that they now have access to. I could solve it with 
    HX-Trigger
    like you said, but I agree that feels hacky. I could also solve it with load polling, but imo that obfuscates the polling a bit, making it more complicated to understand. 
    every 1s
    makes it pretty clear it's polling, whereas 
    load delay:1s
    doesn't read as polling to me
  • e
    I agree it could break for people though. So a minor version bump would probably be required. I doubt a major one is needed though, personally
  • b
    I have been tripped up by nginx silently dropping non-standard headers once... not sure if that's your case...
  • h
    My head hurts, but I think there was more to the Safari back button issue than I thought: https://github.com/bigskysoftware/htmx/issues/1076#issuecomment-1278206624
  • m
    The javascript history api will be the death of me
  • m
    Sometimes I worry for the future of software and software development
  • m
    you should
  • b
    Hey, is it possible to swap CSS or JS using HTMX? Like can you swap the or tag?
  • m
    in a head tag?
  • b
    Maybe. I guess Iā€™m trying to think through how far the concept can be extended. Like if the server-side could drive not only HTML swaps, but also hot swap CSS and JS, what that might enable.
  • m
    I am currently working on an extension to merge head tags: https://discord.com/channels/725789699527933952/725789747212976259/1030151411079585892
  • m
    I'm gonna try to cut a release next week w/ it
  • b
    Oh, I should have just scrolled up šŸ˜Š
  • f
    Honestly I looked at htmx and I wouldn't have thought "swap out CSS/JS/Head with it" would be a use case and I don't intend to do it. I see how ppl have use for it though.
  • f
    Didn't even cross my mind
1...854855856...1146Latest