| 02:55 | <xialvjun> | Can I issue a proposal? When the people open the browser at first time, the browser will ask people 'create or import a private key'. Then websites will call a js function to use this private key and it's own things(like origin or some static string) to generate a hash string which can be used as an user_id. Thus, the login and register flow may be left out in many websites. If a website need a more strict login flow, it can use it's own login flow |
| 02:55 | <xialvjun> | like login twice(receiveing a sms verify code). |
| 08:13 | <smaug____> | kochi: do you know some Shadow DOM (v1) heavy websites? |
| 08:16 | <kochi> | ideally any Polymer 2+ web sites should turn on native shadow DOM... |
| 08:18 | <smaug____> | kochi: but Youtube is still using the old Polymer, I think |
| 08:18 | <kochi> | Google's santa tracker, chromestatus.com, etc.? |
| 08:18 | smaug____ | wonders if there are big non-Google polymer sites |
| 08:18 | <kochi> | yeah, YouTube and Google Earth use old Polymer |
| 08:19 | <kochi> | I'm not sure commercial adopters like ING or EA uses which version of Polymer |
| 08:20 | <kochi> | https://twitter.com/AJStacy06/status/857628546507968512 |
| 08:22 | <kochi> | giving "?dom=shadow" forces Polymer to use native shadow dom instead of polyfilled one, just FYI |
| 08:23 | <kochi> | e.g. https://gaming.youtube.com/ and https://gaming.youtube.com/?dom=shadow |
| 08:25 | <kochi> | EA (for me, it redirects the top page to https://www.ea.com/ja-jp ) seems to use native shadow DOM on Chrome. |
| 22:29 | <domfarolino> | Domenic: Would a <script> tag attribute be a viable approach to function tostring censorship, or is that a bit reckless? I noticed this option didn't make an appearance under https://github.com/domenic/proposal-function-prototype-tostring-censorship#an-external-to-javascript-switch & was slightly curious about it. |
| 22:31 | <domfarolino> | And I assume for module scripts, this would mean any other script loaded via the module would fall under the same censorship defined by the attribute on the top-level script |
| 22:38 | <bradleymeck> | domfarolino: if there are 2 competing graphs 1 with the flag and one w/o it that might get weird |
| 22:42 | <domfarolino> | bradleymeck: wouldn't the HTTP-header idea suffer the same sort of thing? If a module script is served with `Script-ToString: Off`, which loads other scripts not served with this header I feel the same competition ensues |
| 22:43 | <bradleymeck> | domfarolino: how? each individual resource would declare its own mode with that |
| 22:43 | <bradleymeck> | you couldn't have a collision since the fetch of the resource would only ever have 1 result |
| 22:45 | <Domenic> | domfarolino: seems plausible, indeed with potential interactions, but yeah |
| 22:46 | <bradleymeck> | Domenic: i know in the past that vms were using source text to regenerate IR, I'm guessing thats no longer the case? |
| 22:48 | <Domenic> | bradleymeck: nah, aklein pointed that out, so yeah it wouldn't allow immediate memory savings in all architectures for all functions |
| 22:51 | <aklein> | bradleymeck: v8 no longer re-generates from source, but we do keep source around for functions that have not yet been compiled to bytecode (either because they weren't called, or because no other heuristic caused compilation) |
| 22:52 | <aklein> | I expect other engines are similar |
| 22:52 | <bradleymeck> | as long as we can keep the fn position offsets for heap snapshots ;p |
| 23:36 | <domfarolino> | bradleymeck: (no longer here): Ok yeah that makes sense. |
| 23:37 | <domfarolino> | Domenic: potential interactions, as in touching multiple things? HTML would definitely need touched |
| 23:37 | <Domenic> | domfarolino: I meant among modules in multiple graphs, as bradleymeck says. |