| 03:12 | <crankharder> | it seems that if I create an indexedDB on a page that is https |
| 03:12 | <crankharder> | i cannot access it from a page that is http |
| 03:12 | <crankharder> | is that expected, and if so - is there a workaround? |
| 03:13 | <crankharder> | https://developer.mozilla.org/en/IndexedDB/Basic_Concepts_Behind_IndexedDB "IndexedDB adheres to a same-origin policy" |
| 03:13 | <crankharder> | bah |
| 03:15 | <crankharder> | so i'm trying to download data (json) over SSL since it's somewhat sensitive, and store it in idexedDB to be used by the application cache |
| 03:15 | <Hixie> | so... is there a spec anywhere that actually describes the precise steps for parsing http or websocket requests? or am i supposed to guess at things like error handling |
| 03:15 | <crankharder> | my application cache needs to be HTTP since the homepage is HTTP. nobody will visit https to reach my homepage |
| 03:15 | <zewt> | so redirect to https and stop using http |
| 03:16 | <Hixie> | crankharder: if your app is server over http, then it doesn't matter what you encrypt, since someone can just inject some arbitrary JS and take over your site. |
| 03:16 | <Hixie> | served, rather |
| 03:16 | <crankharder> | zewt: that redirect won't happen when offline |
| 03:16 | <zewt> | sure it will, if it always happens first thing so the url they bookmark is https |
| 03:17 | <crankharder> | if they bookmark it |
| 03:17 | <crankharder> | what if they just type in "domain.com" |
| 03:17 | <crankharder> | the app cache needs to be nonsecure |
| 03:17 | <Hixie> | their browser will autocomplete to wherever they had gone |
| 03:17 | <zewt> | then their browser should autocomplete the url |
| 03:18 | <crankharder> | nah, too error prone for dumb users |
| 03:19 | <zewt> | shrug, you'll have the same problem with people typing url.com vs www.url.com |
| 03:22 | <crankharder> | hrm, maybe i can subdomain this whole feature at offline.domain.com or some such, then all of that can be https, instead of the entire website |
| 03:22 | <zewt> | sounds like it's bypassing the whole design of appcache (use the same url) |
| 03:23 | <crankharder> | bah, yea :/ |
| 03:23 | <zewt> | and if people are inconsistently entering urls by hand, that's hardly going to decrease the error rate |
| 03:26 | <zewt> | note that browsers *may* cache permanent redirects, so loading foo.com with a cached redirect to https://www.foo.com may still work even offline (don't know how consistent that is and you need to get the cache headers right) |
| 03:26 | <crankharder> | maybe i could make a second http app cache that simply javascript redirects to https |
| 03:27 | <crankharder> | that creates all kinds of problems though... can't cache the homepage since it's dynamic |
| 05:09 | <crankharder> | my /offline has <html manifest='/cache.manifest'>. my / does not. How is it that when I visit / while offline my fallback clause still invoked? |
| 05:09 | <crankharder> | manifest file: http://pastie.org/3273979 |
| 05:09 | <crankharder> | is it because the same original has a saved cached? |
| 05:09 | <crankharder> | *origin |
| 08:15 | <Hixie> | crankharder: if you're offline, then it'll fail to get /, and will then look in all its caches, and will find the fallback. |
| 11:01 | <Ms2ger> | The spec isn't being updated? |
| 14:06 | <Ms2ger> | Philip`, thanks for your tests :) |
| 14:16 | <Philip`> | Ms2ger: Is that a precursor to asking "but when you are going to stop being so lazy and get around to responding to all the bugs?" :-) |
| 14:16 | Philip` | needs to stop putting that off... |
| 14:16 | <Ms2ger> | You do |
| 14:17 | <Ms2ger> | But it was just because I found a bug in my code with them :) |
| 19:58 | <Hixie> | Ms2ger: which spec? |
| 20:00 | <Ms2ger> | HTML |
| 20:01 | <Hixie> | modulo a problem at dreamhost as we speak, i'm not aware of any non-updating going on (or rather, not going on) |
| 20:01 | <Hixie> | so if something isn't updating, it's a bug somewhere |
| 20:03 | <Ms2ger> | It claimed to be last updated on the 24th earlier |
| 20:03 | <Hixie> | "it" being which copy, exactly? |
| 20:03 | <Ms2ger> | http://www.whatwg.org/specs/web-apps/current-work/multipage/ |
| 20:03 | <Hixie> | ah, multipage |
| 20:03 | <Hixie> | yeah that might be broken |
| 20:03 | <Ms2ger> | Hmm, didn't check singlepage |
| 20:04 | <Hixie> | i'll investigate when i get access to the server again |
| 20:42 | <crankharder> | I've got a page w/ this: <html manifest='/offline_app.manifest'> --- when I visit it I'm getting a cache error event: "resource fetch failed" |
| 20:42 | <crankharder> | I can access the file directly just fine though. thoughts? |
| 20:44 | <crankharder> | just curled it, Content-Type: text/cache-manifest |
| 21:00 | <AryehGregor> | It's always nice to be reminded how many Gmail accounts I have when Google sends an e-mail to each and every one without noticing that they're set up to forward everything. |
| 21:00 | <AryehGregor> | Apparently: nine. |
| 21:00 | <AryehGregor> | At least. |