28339 2015-03-26 09:02:01 +0000 <meta refresh> Specify Refresh HTTP header 2017-08-09 09:48:40 +0000 1 1 1 Unclassified WHATWG HTML unspecified Other All RESOLVED MOVED https://html.spec.whatwg.org/#attr-meta-http-equiv-refresh P3 normal Unsorted 28563 1 contributor ian annevk d ian karl+w3c mike tyoshino zcorpan contributor oldest_to_newest 118947 0 contributor 2015-03-26 09:02:01 +0000 Specification: https://html.spec.whatwg.org/multipage/semantics.html Multipage: https://html.spec.whatwg.org/multipage/#attr-meta-http-equiv-refresh Complete: https://html.spec.whatwg.org/#attr-meta-http-equiv-refresh Referrer: https://html.spec.whatwg.org/multipage/ Comment: <meta refresh> Specify Refresh HTTP header Browsers support it, some Web content uses it. (36 pages in webdevdata have Refresh header but no <meta refresh>) Posted from: 90.230.218.37 by simonp@opera.com User agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/42.0.2311.50 Safari/537.36 OPR/29.0.1795.26 (Edition beta) 119104 1 ian 2015-03-30 23:20:46 +0000 What's this got to do with HTML? Isn't this an HTTP bug? 119161 2 zcorpan 2015-03-31 22:27:33 +0000 I don't know, it seems to me that navigating browsing contexts is HTML material. 119326 3 ian 2015-04-07 22:35:01 +0000 Yeah, we can add this to HTML. I'll add it to my list. 124921 4 karl+w3c 2016-02-08 23:41:59 +0000 See Also https://github.com/httpwg/http-extensions/issues/138 with some data for the patterns found on the Web and a couple of tests. http://www.otsukare.info/2015/03/26/refresh-http-header 128779 5 annevk 2017-07-24 13:20:50 +0000 This is what I have thus far: * Most of the algorithm at https://html.spec.whatwg.org/#attr-meta-http-equiv-refresh can be abstracted into a standalone algorithm (step 5 and onward looks like). * That new algorithm basically takes a string and a document. Though we'll also need to account somehow for the meta element insertion time if that's later than the document is created. Easiest would be to just pass in the meta element and null if it's from a header. And if it's null we don't apply the sandboxing as per bug 28563. * When navigation has created a document, we check if there's a Refresh header and then if there is decode the value of that header with "original ISO-8859-1" (map each byte to a code point that has the same value). Then pass that code point sequence and the created document to the new standalone algorithm. * Basic tests: https://github.com/w3c/web-platform-tests/pull/6606 The algorithm should probably also queue a task and navigate from that task. You can't really navigate out of thin air as it tries to do now. 128785 6 annevk 2017-07-25 09:22:52 +0000 Other things to take into account: * Multiple Refresh headers * Conflicts between headers and elements (probably a single flag on Document objects would do, need to test) * https://github.com/whatwg/html/pull/2865 128821 7 annevk 2017-08-03 13:01:40 +0000 It seems the processing model for multiple Refresh headers is that the values get combined (comma-separated, etc.). It's not clear to me this is good (Firefox doesn't do it for Location): https://bugzilla.mozilla.org/show_bug.cgi?id=1387032. 128822 8 d 2017-08-03 15:55:40 +0000 +tyoshino to see if we can get any Chrome network team thoughts on comment #7. 128823 9 annevk 2017-08-09 09:48:40 +0000 Defining Refresh in general: https://github.com/whatwg/html/pull/2892 Multiple headers: https://github.com/whatwg/html/issues/2900