#[fluffy]well I mean, any side effect that’s happening will have already happened, and the output you’re getting is by definition the same as the output that you got last time
[tw2113] joined the channel
#[snarfed]side effect idempotence doesn’t happen by default, you have to carefully engineer it
#[snarfed]but regardless, yes, 304 and etags are generally good things
#mblaneyI've seen pages with today's date in them and it updates every day... other than that they stay the same, so you can't rely on hashes if you want to know if the main content has changed.
#aaronpk^^ i have had that problem too when writing a feed reader
#[fluffy]Yeah but the hash value is based on the page contents. If the displayed date changed then the hash would have changed too
#[fluffy]So I mean yeah the hash changing doesn’t mean content has changed but
#[fluffy]The hash not changing means the content hasn’t changed
#[fluffy]So I fail to see a case where it’s a bad thing for there to be automatic assignment and 304ing based on the content hash. Unless you mean that this behavior also prevents you from setting your own ETag.
#[fluffy]But in theory you should be able to have more than one ETag based on whatever, given that the matching criterion is if-none-match
[Emma_Humphries], jonnybarnes, nickodd, geoffo, [asuh], KartikPrabhu, GWG, justache, [tantek], gRegorLove, [Ana_Rodrigues], [fluffy], moppy, dckc, [Murray], swentel, jeremych_, [James_Gallaghe], [Rose], leg, gxt and [KevinMarks] joined the channel
#GWG!tell [manton] Reached out to my podcast app and one of the WordPress plugins I use re supporting JSONFeed. One gave me a generic... we'll forward to the dev team, the WordPress plugin support told me that I needed to send my request somewhere and the developer would respond about integrating with the JSONFeed plugin
#ZegnatI guess this is especially relevant to client implementations, not so much endpoint implementors.
#ZegnatWant to drop a request-for-comment in here wrt https://github.com/indieweb/indieauth/pull/56. [dmitshur] raised a good point on whether the (confusing) redirect stuff is even needed if the client is supposed to refetch to establish `me` anyway.
#Loqi[Zegnat] #56 Remove requirement for same domain
swentel joined the channel
#[manton]@GWG By the way, unrelated… I’m rewriting some Micropub code and testing with WordPress, and getting a 403 from q=config in WordPress. Any tips for troubleshooting? Bearer token and URL look fine. (“/index.php?rest_route=/micropub/1.0/endpoint&q=config”)
#GWG[manton]: When was the last test you ran successfully?
#GWGI know I changed 401 to 403 in the last update, because it changed in the spec
nickodd joined the channel
#[manton]@GWG I’m not sure, because usually my code defaults to XML-RPC for WordPress. I disabled XML-RPC so I can use WordPress as a better place to test IndieAuth/Micropub.
geoffo joined the channel
#[manton]I wouldn’t expect either 401 or 403, though, just getting the config and passing a valid Auth header.
#GWG[manton]: Do you have the IndieAuth plugin installed?
#[manton]Checks in site health for IndieAuth all passed. Hmm.
jonnybarnes joined the channel
#sknebeldoes the route thing work like that with WP? or does the "&q=config" need to be encoded into the route for WP to accept it - which would be a WP bug?
#[manton]Good question. I can try encoding it like that, although that seems wrong and might not work with other Micropub implementations.
#sknebeltricky without being able to reproduce it on one of your systems
#[manton]@GWG Would it help if I created a WP user for you on this blog? It’s just for testing so doesn’t matter if anything breaks. https://test.manton.org/
[tw2113] joined the channel
#[tw2113]unreproducability is the bane of any support 😄
#GWG[manton]: If you could, I'll see what I can figure out.
#GWGI wonder if I should build in more information in the return
#GWGOauth response doesn't have an issue with extra parameters as long as it has the ones it expects
#[manton]As a test, I tried also passing the token in an access_token parameter and that didn’t fix it.
#[manton]So I don’t think it’s an HTTP header issue.
#GWG[manton]: That leaves what happens when it gets the token to look at
[chrisaldrich] and [tantek] joined the channel
#sknebelZegnat: that's an interesting one. I wonder if there is any case where the endpoint would want to have a URL along the path as a suggestion? Otherwise I don't think its needed anymore, and even that seems kinda contrived
#ZegnatThe client could store the entire redirect path, that way if the returned `me` value is anywhere within the redirect path you do not need to request again. But that is just the default redirect path and would not include any keeping track of types of redirects / "canonical profile URL"
#sknebelI specifically wondered if the auth endpoint ever wants to know "how did the client get redirected to me"
jonnybarnes joined the channel
#[tantek]petermolnar, in the past there has been a proposed "license" property for microformats that would apply to all the information in that specific microformat. e.g. in an h-entry, a u-license to a Creative Commons license URL would apply to all the other property values like summary, content, photo, video etc.
#[tantek]similarly it could be applied to an h-card and the photo there
#petermolnarthere's already a rel="licence" thing in the wild
#[tantek]I don't think there was much exploration beyond that
#Zegnatsknebel: probably not. All the auth endpoint cares about is what `me` the user is likely to have entered, incase of multi-url-users (e.g. the micro.blog case).
#[tb]So like a `<link>` in the header and/or as appropriate in like a copyright notice that links to the license?
#[tantek]as a simpler alternative to the ridiculously complex RDF gibberish that CC was encouraging people to embed in HTML comments
#[tantek]I'm like wait, you don't need all that, all you need is a rel=license attribute on the *existing a href link to the CC license that you're already encouraging*
#[tantek]hah! tl;dr on the Google proposal: they've picked up the "license" property (per object) *exactly* as it was previously discussed in microformats years ago
#[tantek]Zegnat, if you want to get ahead, go ahead and propose "u-license" as an experimental property for h-entry
#ZegnatI cannot get ahead of the work assignment, the work assignment will be to implement whatever Google says, because nobody cares about the meta data and everyone cares about what Google thinks of them *shrug*
#[tantek]you can do the right thing in microformats pre-emptively because it is literally what Google already copied
[Ana_Rodrigues], [Mat_Rosero], buswolley and [Murray] joined the channel
#[tantek]or were you hoping to save "IndieAir" for your own personal airline?
#sknebelI think one problem with weather is that its somewhat tricky to make data from random stations really usable for forecasts etc. although that doesn't explain lack of pure community projects like there are for other measurements, i.e. air quality, radioactivity, ...
#GWGI am looking for current conditions, not forecasts
#GWGI think offering observation would be a good community thing
[manton], KartikPrabhu and jonnybarnes joined the channel
#jamietanna[m]Ah, good point. I'll add a note to it to say maybe not to do it this way!
#[tantek]I still think most use-cases are handled by a simple "license" property that points to the license URL for the whole object
#sknebelthat could also be a use case for a h-media like wrapper object around photos/videos/..., since they're the most likely to have a different licenses