30432 2009-10-16 01:10:42 -0700 BASE tag is allowed more than once and outside the <head> section 2022-08-19 17:15:39 -0700 1 1 1 Unclassified WebKit DOM 528+ (Nightly build) PC All RESOLVED CONFIGURATION CHANGED http://zaz.sirdarckcat.net/zaz P2 Normal --- 1 sirdarckcat webkit-unassigned ahmad.saleem792 ap bfulgham cdumez rniwa oldest_to_newest 155039 0 sirdarckcat 2009-10-16 01:10:42 -0700 Reproducible: Always Steps to Reproduce: 1. Go to http://zaz.sirdarckcat.net/zaz/ (a script that alerts the location of the script being included). 2. See the alerts Actual Results: the alerts are: a/b.js a/b/c.js a/b/c/d.js a/b.js a/b/c.js a/b/c/d.js Expected Results: the alerts should be: a/b.js a/c.js a/d.js a/b.js a/c.js a/d.js I'm not very good making testcases, and I tried to look for the bug, but I couldn't find it.. I think this is a well known bug for developers, so I guess its a dupe, but I couldnt find the bug where this is being discussed. In any way, I want to propose this being fixed, or are there some details I'm missing? 155040 1 sirdarckcat 2009-10-16 01:12:53 -0700 I should have reported this here instead of chromium: http://code.google.com/p/chromium/issues/detail?id=25017 And I also submitted a bug report for firefox: https://bugzilla.mozilla.org/show_bug.cgi?id=522658 155044 2 sirdarckcat 2009-10-16 01:25:59 -0700 I want to add that this has security considerations, if an attacker is able to add a <base> tag in the middle of the source code, it may leak information to other domains. 155295 3 ap 2009-10-16 15:59:16 -0700 As implied by comment 1, we match Firefox here. We probably shouldn't change this unilaterally - but it's definitely something to keep an eye on. 156088 4 sirdarckcat 2009-10-20 00:08:17 -0700 FYI: https://bugzilla.mozilla.org/show_bug.cgi?id=515401 Apparently Gecko is fixing this 1890354 5 ahmad.saleem792 2022-08-10 11:02:22 -0700 WPT do have coverage for BASE tag multiple: https://wpt.fyi/results/html/semantics/document-metadata/the-base-element?label=master&label=experimental&aligned&view=subtest&q=base Safari TP 150 passes all tests, do we need to do something here? The test URL does not work and I tried to look into Wayback Archive but couldn't find as well. Appreciate if someone can comment and update the bug status accordingly or attached minimal test case (if needed). Thanks! 1892565 6 ap 2022-08-19 17:15:39 -0700 Not sure if WPT covers everything that this bug talks about, but in my testing, all browsers agree now. - When there are multiple BASE elements, only the first one is respected. - BASE elements in BODY _are_ respected. > I want to add that this has security considerations I think that you are correct, but at this point, it's up to websites to prevent base element injection.