| 05:03 | <annevk> | JakeA: I guess… Seems like that might depend on the UA too if it actually reuses the connection and whether the push is bound to the connection lifetime |
| 08:50 | <annevk> | TabAtkins: sorry for the continued back-and-forth, but would still appreciate a reply to https://github.com/whatwg/dom/issues/185#issuecomment-240644834 |
| 08:50 | <annevk> | TabAtkins: ta! |
| 08:52 | <JakeA> | https://developer.mozilla.org/en-US/docs/Web/API/Service_Worker_API lists service worker as "experimental" although it's been in stable releases of two engines for some time now. Can this be removed? |
| 08:53 | <annevk> | JakeA: I suspect so |
| 08:54 | <annevk> | JakeA: I don't think they have consistent policies regarding that |
| 08:54 | <JakeA> | ta |
| 10:05 | <smaug____> | lovely, w3c has kicked me out from www-dom |
| 10:07 | <gsnedders> | smaug____: because of affiliation change? |
| 10:07 | <smaug____> | no |
| 10:07 | <gsnedders> | wait no |
| 10:07 | <annevk> | smaug____: hmm different email address? |
| 10:07 | <gsnedders> | I'm mixing you and Jonas Sicking up for some reason, depsite looking nothing alike. |
| 10:07 | <smaug____> | as far as I know I've been with Mozilla well over 10 years ;) |
| 10:08 | gsnedders | blames being ill and slightly feverish |
| 10:09 | <smaug____> | annevk: nope. I think this is the same nonsense what kicked me out from webapps mailing list last year |
| 10:09 | <smaug____> | annevk: so, I think I haven't got any www-dom emails since last autumn |
| 10:09 | <smaug____> | annevk: and I assume all the members of the WG were kicked out around that time |
| 10:09 | <smaug____> | MikeSmith: ^ |
| 10:10 | <smaug____> | er |
| 10:10 | <smaug____> | annevk: er, you're right |
| 10:10 | <smaug____> | aha, yes, mozilla.com email |
| 10:10 | <smaug____> | too many email accounts |
| 10:12 | <smaug____> | foolip: do you think you could add deprecation message about mutation events to blink? |
| 10:14 | <annevk> | I think Blink has a policy now that they only add those warnings if they can promise a release they'll remove the feature by |
| 10:14 | <annevk> | Presumably they no longer believe in long term deprecation |
| 10:18 | <smaug____> | sad |
| 10:19 | <smaug____> | and Edge folks just do whatever blink folks do |
| 10:44 | <MikeSmith> | smaug____: you need me to do something? |
| 10:45 | <smaug____> | MikeSmith: carefulness. I was being stupid. :) |
| 10:45 | <smaug____> | so, no |
| 10:45 | <MikeSmith> | ok |
| 10:46 | MikeSmith | sees the message smaug____ posted to www-dom |
| 10:46 | <smaug____> | MikeSmith: yeah, originally I used m.com email address |
| 10:46 | <MikeSmith> | ah |
| 10:47 | <MikeSmith> | I can add that to the accept list if you awnt |
| 10:49 | <smaug____> | MikeSmith: nah, I rarely use @m.com |
| 10:49 | <MikeSmith> | k |
| 11:34 | <foolip> | smaug____, annevk: there isn't really a policy, but I do think that *just* deprecating and waiting is unlikely to get us where we want |
| 11:34 | <smaug____> | foolip: it has worked before, so why wouldn't it work anymore? |
| 11:34 | <smaug____> | like, we have removed some old stuff in gecko after warnings |
| 11:35 | <foolip> | Interesting, what were those things? |
| 11:35 | <smaug____> | foolip: and how would you get usage of mutation events down if no one telling they are deprecated? |
| 11:35 | <smaug____> | foolip: some old netscapeisms in Event IIRC |
| 11:36 | <foolip> | I've never seen usage really move in response to deprecation, but that's been on timescales of months to a few years, never a decade-long thing like one could argue might work for say sync XHR |
| 11:37 | <foolip> | So if I were to tackle mutation events, I'd first try to figure out *why* the usage is high and if there's anything that can be done about those causes. |
| 11:38 | <foolip> | If it looks doable and one makes an attempt, then a deprecation message would definitely be appropriate. |
| 11:39 | <foolip> | What I'm concerned about is "hopeful" deprecations, where deprecation is all we do and we remove if usage falls and otherwise standardize, because that's potentially wasting a lot of time for the people who take the messages seriously. |
| 11:39 | <smaug____> | foolip: sure, I agree that if we can get urls of the pages using mutation events, it might reveal something useful |
| 11:40 | <smaug____> | foolip: well, we spent already tons of time to design and implement the replacement for mutation events |
| 11:40 | <smaug____> | if browser vendors aren't after all too interested in to try to remove mutation events, it feels like a bit wasted time |
| 11:42 | <foolip> | That's true, and I definitely agree we need to bring this story to some kind of conclusion. |
| 11:43 | <foolip> | The "hopeful" kind of deprecation table isn't out of the question even, if we think mutation events are bad enough, and they are pretty bad. |
| 11:44 | <foolip> | But someone on Blink needs to invest the time in this for anything to happen, so let's hope that someone materializes. |
| 11:44 | <foolip> | (waiting on www-dom thread) |
| 11:48 | <foolip> | Overall I think we're pretty bad at dealing with these kind of stalemates. |
| 12:15 | <annevk> | Yes, or they get concluded by one vendor and then the others don't follow up |
| 12:15 | <annevk> | Sorry, concluded by all, implemented by just one |
| 12:40 | <annevk> | smaug____: in http://searchfox.org/mozilla-central/source/dom/base/nsINode.cpp#920 doesn't the chain of parents end up containing the Attr node twice? |
| 12:41 | <annevk> | smaug____: never mind |
| 13:13 | <zcorpan> | TabAtkins: trying to update bikeshed (git pull; bikeshed update) i get... |
| 13:13 | <zcorpan> | FATAL ERROR: This version of the test suite API is no longer supported. Please update Bikeshed. |
| 13:13 | <zcorpan> | ✘ Did not generate, due to fatal errors |
| 13:13 | <zcorpan> | hum. bikeshed update-d again and now it worked |
| 13:13 | <zcorpan> | very well |
| 14:19 | <ShakespeareFan00> | hi. |
| 14:20 | <ShakespeareFan00> | Where would be the best place to start a disscusion concerning side-titles? |
| 14:20 | <ShakespeareFan00> | https://en.wikisource.org/wiki/Page:Railways_Act_1921_(ukpga_19210055_en).pdf/39 being an example of content that has it.. |
| 14:25 | <ShakespeareFan00> | Currently I am using a rather hacked together approach of nested DIV's which isn't ideal |
| 14:30 | <ShakespeareFan00> | I was advised elsewhere that Aside was a possible soloution but apparently you can't put an aside inside a paragrph. |
| 14:32 | <ShakespeareFan00> | What I was therefore needing was a way to mark content within the flow of an element, that's associated with a point in the content of that element but isn't part of that element. |
| 14:32 | <ShakespeareFan00> | Such as a sidenote which is linked to a particular point in the paragraph of a document. |
| 14:34 | <ShakespeareFan00> | The user agent could on encountering such an item make an appropriate decision about how to handle it... be it as a footnote at the end, or as a 'styled' block which was moved to the left or right of a document page accordingly. |
| 17:40 | <TabAtkins> | annevk: Yeah, I've got that opened in a tab, I've just been overloaded with, well, everything for the past week or so, so I'm slow to reply to things that require architectural thought. Sorry. :( |
| 17:40 | <TabAtkins> | zcorpan: I occasionally see that server yield something funky. Updating again usually works, as you found. |
| 17:40 | <tantek> | lol TabAtkins when *haven't* you been overloaded in the past (few years) or so? ;) |
| 17:41 | <TabAtkins> | tantek: There are *degrees*, man. |
| 17:41 | <TabAtkins> | In particular, I'm moving right now. |
| 17:42 | <tantek> | TabAtkins: to SF to join the rest of us? ;) |
| 17:42 | <TabAtkins> | Oh god no. |
| 17:42 | <TabAtkins> | Just moving back to Sunnyvale, because I bought a house. ^_^ |
| 17:46 | <ShakespeareFan00> | Good Afternoon |
| 17:46 | <ShakespeareFan00> | I posted earlier, but as I didn't get a response I assumed you were all busy |
| 18:15 | <annevk> | TabAtkins: no rush at all |
| 18:15 | <annevk> | TabAtkins: unless SimonSapin_ says otherwise 😊 |
| 18:16 | <SimonSapin_> | annevk: what is? |
| 18:17 | <TabAtkins> | https://github.com/whatwg/dom/issues/185#issuecomment-240644834 |
| 18:18 | <SimonSapin> | yeah, no hurry. I think we all agree on the intended behavior, we just need to make the specs consistent |
| 18:53 | <ShakespeareFan00> | TabAtkins: Good evening |
| 18:54 | <TabAtkins> | Footnotes (of which sidenotes are a display variant) have a long and terrible history of attempting to be supported on the web. |
| 18:55 | <TabAtkins> | They're not compatible in a simple way with the existing 'display' values - they always need some additional coordination between containers, in a way that can't be expressed via a simple generic relationship. |
| 18:56 | <TabAtkins> | It does suck that <aside> can't be used inside of <p>, tho. |
| 18:57 | <ShakespeareFan00> | The problem I have is how to tell a browser that an "Element" inside an Element is Associated content, not part of the flow.. |
| 18:57 | <ShakespeareFan00> | Mediawiki markup has pusdeo tags... and a lot of PHP code i don't fully understand to implement them :( |
| 18:58 | <TabAtkins> | Hard to do generically, unfortunately. Can't really use any of the appropriate elements, as they all auto-close <p>. |
| 18:58 | <TabAtkins> | But that information wouldn't help you visually anyway; as I said, there's no generic way to accommodate *displaying* sidenotes or footnotes. At best it would help screenreaders. |
| 18:59 | <ShakespeareFan00> | TabAtkins: As HTML is a 'structural' language, the display side would be an issue for user agents... |
| 19:00 | <TabAtkins> | The display side is *never* something to ignore. Silent metadata that has no visible effect is worthless metadata that will inevitably bitrot and become useless/harmful. |
| 19:00 | <ShakespeareFan00> | Okay |
| 19:00 | <TabAtkins> | And "it's something for user agents" doesn't excuse us from dealing with that, since we write the spec *for* user agents. |
| 19:01 | <ShakespeareFan00> | I have the template I partially wrote for doing a very hacky version of Sidenotes... |
| 19:02 | <ShakespeareFan00> | https://en.wikisource.org/wiki/Template:Cl-act-paragraph - This isn't clean HTML though, as it's HTML+some mediawiki markup for passing params... |
| 19:02 | <TabAtkins> | Yeah, you can handle it kinda-sort on your own, by explicitly coordinating several elements. At that point, just use <span class=sidenote>. It's not ideal, but it's the best you'll reasonably have. |
| 19:02 | <ShakespeareFan00> | TabAtkins: I wasn't sure you could do "clear:left" with spans... |
| 19:03 | <ShakespeareFan00> | Or define a span to behave like a block |
| 19:03 | <ShakespeareFan00> | In any case I don't think redefininf what a span is was a good idea |
| 19:05 | <ShakespeareFan00> | Another example for a template for sidenotes in Mediawiki was this https://en.wikisource.org/wiki/Template:Right_sidenote |
| 19:05 | <ShakespeareFan00> | Which uses a span |
| 19:06 | <ShakespeareFan00> | However if you use a span vs a div (with clear:left) and they can overlap... |
| 19:06 | <ShakespeareFan00> | *then the two spans can end up overlapping which looks ugly |
| 19:08 | <ShakespeareFan00> | Like here - https://en.wikisource.org/w/index.php?title=Page:Historical_Works_of_Venerable_Bede_vol._2.djvu/121&action=submit |
| 19:08 | <ShakespeareFan00> | And whilst mediawiki's approach is nominally browse independent, I have my doubts |
| 19:08 | <ShakespeareFan00> | Hence my earlier comments about needing to find a long term soloution. |
| 19:10 | <ShakespeareFan00> | I'd love to be able to replace my current nested divs's with a single <sidenote> or <aside> and have the browser do the hard work... |
| 19:11 | <ShakespeareFan00> | And of course being able to embed Aside inside other elements would be useuful to this.... |
| 19:13 | <ShakespeareFan00> | Example <p> Pursant to the Act of 14. Wales c. 21 It was rehearsed that <aside class="sidetitle" break=flow>Rehersal of the Act 14. Wales c. 21</aside> (1) No measures previously enacted would be ....</p> |
| 19:13 | <ShakespeareFan00> | Naturally you still shouldn't be able to nest <p> directly inside <p> ffor what should be obvious structural reasons.... |
| 19:14 | <ShakespeareFan00> | but bweing able to do <p> foo-<aside><p></p></aside>bar</p> to form a side-note with the main flow reading "for-bar" would be really useful to my use case. :) |
| 19:15 | <ShakespeareFan00> | If you read the source code for cl-act-paragraph I also had to do some convolouted things to get it to work over multiple pages |
| 19:15 | <ShakespeareFan00> | hence the split param. |
| 19:15 | <ShakespeareFan00> | And apologies for the wall of text |
| 19:16 | <ShakespeareFan00> | Perhaps my use case is beyond what HTML is capable of? |
| 19:38 | <avivr> | hey, is this the place to ask questions regarding the web notifications standard? |
| 19:38 | <avivr> | or is there a more specific channel? |
| 19:39 | <ShakespeareFan00> | TabAtkins: I hope my comments will be considered in the next round of spec revisions |
| 19:39 | <Domenic> | avivr: we could probably help, yeah |
| 19:40 | <Domenic> | ShakespeareFan00: why is it important to you to use a built-in tag name? <article> is not treated any differently by browsers or other technology than <span class="sidenote"> |
| 19:41 | <avivr> | i wonder if there is a way to specify the screen in which the notification will be shown at? i'm using chrome/mac, and the notifications always shown on the primary screen, even if i have multiple monitors and the browser window is at screen #2 or #3.. |
| 19:42 | <ShakespeareFan00> | Domenic: Did you take a look at the link I gave? |
| 20:23 | <avivr> | anyone can help? should i open this question as an issue on the github repo? |
| 20:24 | <ShakespeareFan00> | Domenic: I could use a span , but in the example I linked, they overlapped... |
| 20:24 | <ShakespeareFan00> | https://en.wikisource.org/w/index.php?title=Page:Historical_Works_of_Venerable_Bede_vol._2.djvu/121&action=submit being the example |
| 20:25 | <ShakespeareFan00> | Using Aside's would be next choice, but you can't put those inside a <p></p> apparently |
| 20:25 | <Domenic> | ShakespeareFan00: right, you need to use two spans then. |
| 20:25 | <ShakespeareFan00> | They are two spans |
| 20:26 | <Domenic> | ShakespeareFan00: then you're good! |
| 20:26 | <ShakespeareFan00> | I'm not sure why they overlpa though |
| 20:26 | <ShakespeareFan00> | Structurally I don't see annything wrong.. |
| 20:26 | <ShakespeareFan00> | MAybe I need to go back and ask #css again? |
| 20:27 | <Domenic> | avivr: ah, that's not part of the API. It might be worth asking on the GitHub repo, but I kind of doubt we want to give that kind of control... do you know if native apps give that control? |
| 20:28 | <ShakespeareFan00> | Domenic: Sadly I;m having to conclude on the basis of what people have said that my use case is beyond HTML's capabiloity (other than by hack soloutuons) |
| 20:28 | <ShakespeareFan00> | :( |
| 20:28 | <ShakespeareFan00> | Back to paying out for Adobe products it is. |
| 20:28 | <ShakespeareFan00> | :* |
| 20:28 | <ShakespeareFan00> | :( |
| 20:29 | ShakespeareFan00 | out |
| 20:29 | <Domenic> | ShakespeareFan00: it's kind of hard to figure out what you're asking, but maybe stackoverflow will be able to help. This channel is generally not about solving peoples web pages. |
| 20:29 | <ShakespeareFan00> | I camein here to ask about sidenote support in HTML |
| 20:29 | <ShakespeareFan00> | I was using the linked pages as examples of how it's currently done |
| 20:29 | <Domenic> | Yeah. I am sure HTML can make it happen with <span>s, but it's hard to tell what you're asking and this isn't the right group to solve your problems. |
| 20:30 | <ShakespeareFan00> | Thank you |
| 20:30 | <ShakespeareFan00> | Demenic: Thank you making me feel stupid |
| 20:30 | <ShakespeareFan00> | Domenic: :( |
| 20:30 | <Domenic> | :( that certainly wasn't the intent. I'm just trying to tell you a good place to ask for help. |
| 20:31 | <ShakespeareFan00> | I know it wasn't, but I still feel that way when I try to ask for <sidenote> support and can't explain what i mean |
| 20:32 | <boogyman> | ShakespeareFan00: maybe create a drawing and upload it, then ask again back over in #css |
| 20:32 | <ShakespeareFan00> | Possibly but someone on #css said to ask in here based on my comments there |
| 20:32 | <ShakespeareFan00> | (maybe they were in error) |
| 20:33 | <boogyman> | I recommended that you mention your use-case here |
| 20:33 | <ShakespeareFan00> | I did that |
| 20:33 | <Domenic> | You never did explain why it's important to you that you use a built-in HTML tag instead of a <span class="sidenote"> |
| 20:33 | <ShakespeareFan00> | Domenic: Because in the linked example , the span cuse overlapped... |
| 20:33 | <ShakespeareFan00> | Anyway I'm too tired to continue arguing |
| 20:34 | <Domenic> | ShakespeareFan00: right. But creating a new tag, or using <aside>, does not solve that. |
| 20:34 | <ShakespeareFan00> | or trying to explain |
| 20:34 | <Domenic> | Tag names have nothing to do with overlapping or not. |
| 20:34 | <Domenic> | No tag in HTML is able to overlap |
| 20:34 | <ShakespeareFan00> | Spans don't behave as blocks |
| 20:34 | <Domenic> | They do if you do style="display: block" |
| 20:34 | <ShakespeareFan00> | Time to examine what Mediwiki thinks it's doing with the span class then |
| 20:35 | <ShakespeareFan00> | And sorry if I was bit brusque earlier... |
| 20:35 | <ShakespeareFan00> | And apologies for wasting your time on something that seems not to be an issue you can actually handle... |
| 20:35 | ShakespeareFan00 | out |
| 20:38 | <avivr> | not sure about native applications, will try to figure out.. the use-case of dual+monitors is quite common, and sometimes the notification can be very far from your working area (website window), so they can be easliy missed. |
| 20:39 | <avivr> | maybe the solution is using a sound with the notification, but i always found that to be too anoyying :\ |
| 20:40 | <avivr> | if I'll find that this is possible with native apps I'll try discussing it on the github repo |
| 20:40 | <avivr> | thanks a lot for the help Domenic! |
| 22:36 | <TabAtkins> | jyasskin: Okay, that "global object" definition will work correctly now. Whitespace is now normalized in anchors. |