I've seen shoelace before, it looks good

hey, is there an event that htmx emits on all errors? i have kind of generic error-displaying function that i want to run if any htmx req fails. is there one single one that i can catch, or do i need to catch individually htmx:sendError, htmx:responseError, etc?

does discord have a slapping-own-forehead emoji? thanks. (didn't see it on https://htmx.org/events/ though, fwiw.)

@mysterious-toddler-20573 Does this strategy sound reasonable to you or am I missing something simpler?

unfortunately if you want to avoid the spinner flash, that's probably the best way

if you log a bug we can think about not removing the spinner when we get a redirect

Thanks. There might be an even better approach, but one improvement I was thinking about is if there was something like a

htmx:redirectReceived

event, then I could handle that by keeping the overlay. This has the advantage of not spreading the logic across both server and client for this particular feature (ie, server would be able to redirect as normal, not worrying about the client experience with trigger headers, and all the logic is on the client, albeit in the form of both an hx-indicator property and an event handler.

I think you could achieve this w/ a custom event

i guess the problem is that the redirect is conditional

so if redirect do not remove indicator, otherwise do

kinda hard w/o some deep integration w/ htmx

A more systemic solution would be adding a config so you could write something like (this syntax isn't right but you get the idea):

hx-indicator="#overlay keepOnRedirect

. I'd guess the pattern is probably reasonably common that it could be justified, but I don't know.

Anyone tried https://barba.js.org/ with hmtx? Just saw it on hn

Here it goes. I ended putting it near the end with things like security, as it is more a "yeah, everything works out of the box according to HTTP standards" rather than something htmx specific. https://github.com/bigskysoftware/htmx/pull/1049

thank you @mysterious-toddler-20573! Good advice, I was wondering where it's acceptable to post stuff related to htmx :). Will do so.

How can I trigger a GET with 2 request parameters that are coming from 2 inputs? I tried this:

<div hx-swap="none">
            <input id="date-range-from"
                   type="date"
                   name="fromDate"
                   th:value="${fromDate}"
                   hx:get="@{/batches-table}"
                   hx-params="fromDate,toDate"
            >
            <input id="date-range-to"
                   type="date"
                   name="toDate"
                   th:value="${toDate}">
        </div>

But that does a request with only the

fromDate

, not the

toDate

from the other input.

hx-include="closest div"

on the first

input

?

Perfect, thanks!

Is it possible to set the encoding with

htmx.ajax

?

it doesn't look like it currently, although you should be able to catch

htmx:configRequest

and override the

Content-Type

header

Okay, I'll try that

In my codebase, that proved to be tricky. Would you accept a PR to allow for changing this? https://github.com/bigskysoftware/htmx/pull/1050

Scratch that. Here's the PR: https://github.com/bigskysoftware/htmx/pull/1051

I just noticed I am getting a favicon request for every request and is bothering me

and the responses to these requests are fragments that do not have a favicon link on them

is there a favicon in the default location? I think it is /favicon.ico. I have a tiny one there, and all the others hidden in a folder

I would guess once your browsers gets that favicon then it shouldn't ask all the time

I add

<link rel="icon" href="data:,">

to get rid of favicon requests

Has anybody already given some thoughts to the concept of using using fragments to implement an automatic/declarativ lazy-loading strategy? I am thinking like a template where you have multiple tabs, every tabs content is marked as a fragment and every tab has some htmx-get on click that can load the according content fragment. Then, for the full-page rendering, there would be an option to exclude all fragments (or some of them, via a selector). On initial page-load only the current active tab content gets fetched. I am thinking about use cases where we load a really fat objects and render a really fat UI with many tabs. Am I overthinking things or would that be an appropriate use case for fragments?