#dev 2020-04-26
2020-04-26 UTC
geoffo joined the channel
# GWG Only https://github.com/microformats/h-entry/issues/17 that I know of specifically.
# [tantek] hmm, I think it should at least be linked from /watch#Brainstorming with a brief description of the motivating use-case
[chrisaldrich] joined the channel
# [chrisaldrich] I would like to note the length of the watch, but haven't got examples of it yet.
# [chrisaldrich] Generally I post watches as I'm finishing watching them. I'll note if I didn't watch the entirety of the thing, so otherwise I presume the start time was the running time of the thing, which I could calculate or find.
# [chrisaldrich] Often when binge watching, I can guess the running times as the length of time between watches...
# [chrisaldrich] This isn't always the case with listens because I'm usually listening to most podcasts at 1.25-1.5x speed.
# [chrisaldrich] Tantek, I've saved some data in my database for it on some older posts, but haven't kept up with it since it's not something I'm displaying.
# Loqi [Tantek Çelik] ▶️ watched Arundhati Roy: “The Pandemic is a Portal” (95min) https://www.haymarketbooks.org/blogs/130-arundhati-roy-the-pandemic-is-a-portal published by @haymarketbooks hosted by @imaniperry.
Timely, highly recommended. #NoGoingBack #Pandemi...
geoffo joined the channel
geoffo, gRegorLove, [chrisaldrich], nickodd and [jeremycherfas] joined the channel
# [jeremycherfas] [LewisCowles] It took me a moment to realise what you intended by “chimping”. Digital photographers use the word differently.
[LewisCowles] joined the channel
# [LewisCowles] hmm... only online parser that seems to work with me is https://kartikprabhu.com/connection/mfparser
# [LewisCowles] xray really doesn't love a h-feed with h-entries with h-cards
# KartikPrabhu [LewisCowles]: mine is mostly mf2py the python one
# [LewisCowles] it's wonderful. I was beginning to doubt that I'd even marked things up in the right way
# [LewisCowles] I do have some p-x-{} in there and it even picked them out
# [LewisCowles] KartikPrabhu++
# [LewisCowles] mf2py++
# KartikPrabhu I'm not really keeping up to speed on mf2 land anymore so you might want to try https://microformats.io/
[KevinMarks] and [chrisaldrich] joined the channel
# KartikPrabhu [LewisCowles]: the *-x-* properties are valid mf2 so parsers should get them, but consumers might not use them at all
# [LewisCowles] oh thats more that the spec didn't have (or I didn't find) properties I wanted to have machine readable
# KartikPrabhu so you should not use my site as an indication of what feed consuming applications will parse
# [LewisCowles] good practice even if useless
oodani joined the channel
# [LewisCowles] TBH anyone parsing what I'm testing is probably going to use something I'll have to write.
# [LewisCowles] with the aim being "please don't call, email or contact"
[tantek] and [jgmac1106] joined the channel
# [jgmac1106] [cleverdevil] has the best watch posts, they are tied directly into his plex and automatically post
leg and jeremych_ joined the channel
# [LewisCowles] either a bug in browser, or bug in me
# [LewisCowles] ``` const changeTagName = (element, newTagName) => {
# [LewisCowles] // Create a replacement tag of the desired type
# [LewisCowles] var replacement = document.createElement(newTagName);
# [LewisCowles] Array.from(element.attributes).forEach(attribute => {
# [LewisCowles] replacement.setAttribute(attribute.nodeName, attribute.nodeValue);
# [LewisCowles] });
# [LewisCowles] replacement.innerHTML = element.innerHTML;
# [LewisCowles] element.parentNode.replaceChild(replacement, element);
# [LewisCowles] }```
# [LewisCowles] the innerHTML I could change to addChild which is a fragment of the innerHTML text.
# [LewisCowles] this is resulting what was an anchor with many children; having the children added alongside a <span>
# [LewisCowles] when I try to re-create in browser using the function... Nothing
# [LewisCowles] crickets
jeremych_ joined the channel
# [LewisCowles] I think the transposing of innerHTML might not be helping me a great deal
# [LewisCowles] but I need the elements children (try's not to sound evil)
# Zegnat [LewisCowles]: https://jsfiddle.net/0npjh2me/
# [LewisCowles] yeah, I have a codepen with the same. I Guess I just have to step through a debugger to see what part of it's interactions mess things up
# [LewisCowles] its resulting in something akin to `<span class="I was a parent of should be child"></span><div class="should be child"><!-- blah blah blah --></div>`
# [LewisCowles] I thought it might be the browser engine at first
# [LewisCowles] I think it's block elements within inlines too, but I'm merely changing the inline tag from HTML someone else is producing
# [LewisCowles] TBH I don't think there will be a pay-off from this adding microformats to kanban exports
[mapkyca] and omz13 joined the channel
# [LewisCowles] the software already exports JSON
# [LewisCowles] useless JSON for consumers who are not me
jamietanna joined the channel
# [LewisCowles] the bug is mine / how `document.createRange().createContextualFragment` works
# [LewisCowles] I can't give it a whole HTML document
# [LewisCowles] `let fragment = document.createRange().createContextualFragment(
# [LewisCowles] document.querySelector('html').outerHTML
# [LewisCowles] );`
# [LewisCowles] works just like a C integer overflow in that there is no error at all, but it does surprising things
# [LewisCowles] I was experimenting with fragments as a way to not mutate visible DOM and need to reload
# [LewisCowles] > The HTML fragment parsing algorithm is used if the range belongs to a `Document` whose HTMLness bit is set. In the HTML case, if the context node would be `html`, for historical reasons the fragment parsing algorithm is invoked with `body` as the context instead.
# [LewisCowles] This probably meant something to the author, although I at least see it's removing invalid DOM when parsing with body context
# [LewisCowles] This is quite brutal, but I'll pivot back to modifying the DOM inline and refreshing after initiating download
# [LewisCowles] all of the quirks should be serialized and exposed somewhere. It would help overcome oddities like this
# [LewisCowles] CPU's have CPUFlags, which you can get at via inline ASM in C. Ultimately checking things seems to slow down the break-neck speed, but really charging around like a bull in a china shop is as limiting due to all the need to step through these monsters we've made
# [LewisCowles] DOM Mutation events?
# [LewisCowles] ooh
# Zegnat But if you find yourself in a place where you care about how that algorithm is running inside the browser, you have gotten yourself in a place where you no longer can work inside the browser’s JavaScript VM. As far as I know, it does not expose HTML parsing (as HTML parsing it not part of ECMAScript)
# [LewisCowles] yeah it'll be some extension API possibly not fully exposed to JS VM
# [LewisCowles] In this case I may just serve text
# Zegnat The parse errors I linked there almost all have specced way to silently recover build into the spec. Because browsers never hard exit on HTML parsing these days. (Unlike how they used to stop parsing XHTML previously on the slightest syntax error). So whenever you are seeing inconsistencies between input string and output DOM it is almost always because some error got handled. And these will be consistently handled, they are part of
# [LewisCowles] the string is coming from DOM
# [LewisCowles] that is the most frustrating part. I'm feeding the serpent it's own tail
# [LewisCowles] Zegnat++
# [LewisCowles] you are a lot more patient, forgiving and understanding than I feel right now
# [LewisCowles] sknebel I'm literally taking the whole document and trying to dump it into something I can use the same API's on without visual effects so html node outerHTML, into a new parser
# [LewisCowles] it fails to re-construct itself silently
# [LewisCowles] but I'm setting fire to that work now. I'll just mutate the DOM and refresh
# [LewisCowles] it at least is simple and easy to reason about
# [LewisCowles] the artifact of that has less moving parts so may be more forgiving
# [LewisCowles] really a frontend hack is all this has been that I've spent way too long hacking / exploring
# [LewisCowles] I know I could I frame and postMessage
# [LewisCowles] Zegant the answer to that was not locating `document.documentElement.cloneNode(true)`
# [LewisCowles] ha TBF I've never come across document.documentElement before either
# [LewisCowles] Oh I'm never updating the document itself btw
# [LewisCowles] the point of cloning is to only need the outerHTML when writing into a zip file
# Zegnat Hmm, you may also want to look at https://github.com/WebMemex/freeze-dry as that does a lot of live-website-to-static-string work
# [LewisCowles] I solved the doctype in an evil way
# [LewisCowles] I concatenate a const
# [LewisCowles] allocation hell
# [LewisCowles] but so is the way I'm cleaning filenames
# [LewisCowles] I'll look to iterate using freeze-dry next
# Zegnat I run it inside a headless Chromium instance for any URL I throw at it, thus being able to run my own web archiver of links. Code here, though I offer no support running it https://github.com/Zegnat/node-beanstalkd-web-archiver
# [LewisCowles] this isn't a straight "archiving" task though
# [LewisCowles] I won't accept data URI's either
# [LewisCowles] it's why I went the zip route
# [LewisCowles] I can build a whole filesystem
# [LewisCowles] I may use it to see about a whole site export. That is a pain point although I doubt I'll ever use it
# [LewisCowles] some of the push-back on data URI's is that there are size limits. Is it still 64kb?
# [LewisCowles] My page weight is circa 1MB
# [LewisCowles] as I iterate, I hope to chop more off of that as it's frankly a little disturbing
# [LewisCowles] if it works for you it's very cool in any case
# [LewisCowles] that 1MB is for a single card kanban with a cover
# [LewisCowles] a png cover
# [LewisCowles] I think if I use cover images, I'd like to use gif's
# [LewisCowles] or some on-hover gif play thing
swentel and nickodd joined the channel
# [LewisCowles] scammers--
indier joined the channel
# [LewisCowles] I think they may be able to waive the cost
# [LewisCowles] > 4p
# [LewisCowles] > per transaction In addition to payment processing fees.
# [LewisCowles] > *Waived* for accounts with standard 1.4% + 20p pricing
[jgmac1106] joined the channel
[tantek] joined the channel
GWG joined the channel
# [LewisCowles] RE: Instagram from chat. The filters were maybe innovative when they came out, but I'm not sure that is true any more
[KevinMarks] joined the channel
# [KevinMarks] [LewisCowles] have you looked at web bundles? https://web.dev/web-bundles/
# [LewisCowles] [KevinMarks] I've not, but I want to upload mine over FTP
# [LewisCowles] well by filesystem. I should stop saying FTP
# [LewisCowles] RE: Instagram again, I think Snapchat filters are way more cutting edge
# [KevinMarks] I think web bundles work by filesystem too, Bluetooth is just an example
# [LewisCowles] okay I just tried it [KevinMarks] it's not the same thing either
# [LewisCowles] learning about CBOR was a whole thing
# aaronpk not a lot of stuff here http://microformats.org/wiki/payment
# [LewisCowles] it is still faster than having meteorjs rebuild every single time I want to test changes
# [LewisCowles] this is going to be like webassembly, xml and javascript. It's going to grow in popularity and I'm going to hate it
# [LewisCowles] victory is mine. One teensy edit to do less and I have what I want / need
# [LewisCowles] • removed useless app-chrome & page elements
# [LewisCowles] • insert non-CSS dependent content instead of relying on title="" attributes
# [LewisCowles] • bundle to zip
# [LewisCowles] • be able to read in css disabled / text-mode browser
# [LewisCowles] I did have to use relative URL's and write weird DOM hacking scripts to patch HTML & CSS files to get CSS background's to become HTML entities
# [LewisCowles] • upload zip file
# [LewisCowles] I still have to fix some CSS so I'm not using `display: none;`
# [LewisCowles] I have no way to mark up all the h-cards I'm embedding in h-entry
# [LewisCowles] h-cards are not for authors, but assignees on tasks
# [LewisCowles] for now I just mark them up as h-card
# [LewisCowles] https://gist.github.com/ffoodd/000b59f431e3e64e4ce1a24d5bb36034 is a nice assistive text shim
# [LewisCowles] I need to check it'll work for what I'm doing, but in any case it'll hide text in a less sucky way
# Zegnat accessibility < [https://gist.github.com/ffoodd/000b59f431e3e64e4ce1a24d5bb36034 CSS for marking something screen reader only] including optional code for making it hidden but focusable
# Zegnat accessibility << [https://gist.github.com/ffoodd/000b59f431e3e64e4ce1a24d5bb36034 CSS for marking something screen reader only] including optional code for making it hidden but focusable
# Loqi ok, I added "[https://gist.github.com/ffoodd/000b59f431e3e64e4ce1a24d5bb36034 CSS for marking something screen reader only] including optional code for making it hidden but focusable" to the "See Also" section of /accessibility https://indieweb.org/wiki/index.php?diff=69687&oldid=68968
# [LewisCowles] hopefully upstream wants some of these changes
# aaronpk well that was a fun exercise https://indieweb.org/payment#Payment_Request_Post_Type
[mapkyca] and [grantcodes] joined the channel
# [grantcodes] I have 2 bank accounts that accept payments via specials urls. One generates random looking urls and another uses my username + value in the url so it's easy to generate
# [grantcodes] One generates a random link that is specific. But the other (that I use on my site) is generic.
# [grantcodes] But because it's directly to the bank, they handle card transactions and there are no fees
gRegorLove joined the channel; nickodd left the channel
# [KevinMarks] Uh oh, I do?
# [KevinMarks] Ah, I don't tend to receive imessages. Technically they're on my mac, but I just ignore the OS notifications there
# [KevinMarks] So, yes, let's do a test case.
# [KevinMarks] Also I was using a throwaway sim at xoxo, so if it did map to that phone # it may not have worked either
# aaronpk first draft of what these posts are going to look like in a list https://media.aaronpk.com/2020/04/file-20200426111239-2746.png
GWG joined the channel
# [KevinMarks] No https for me
OlyaAbram joined the channel
KartikPrabhu, geoffo and [mapkyca] joined the channel
leg and [KevinMarks] joined the channel
# [KevinMarks] I like how you have a beer in your photo, as opposed to a was of money
leg joined the channel
# [KevinMarks] So what post type do I make to say I sent it?
manificattion10, mitchell, geoffo, [jgmac1106], gRegorLove_ and tomasparks joined the channel