This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Regarding the meta element for an http-equiv Refresh value that includes a relative URL, a question has arisen on whether to begin it with a leading slash, since the path is not relative to the base element. See <http://stackoverflow.com/questions/30690448/does-http-equiv-refresh-meta-tag-url-get-a-leading-slash>. I recommend that HTML5, section 4.2.5.3, be clarified. (I don't know enough to propose specific wording.)
Change Component to "HTML5 spec". /paulc
The http-equiv attribute is used to associate metadata with the meaning defined for an HTTP header field of that name. Hence, it is defined by the specification that registered Refresh as a header field and, like all such header fields, the base URL for relative references is the request URI that was used to retrieve the document.
See comment 2 The relevant requirements for URL specified in the value of the meta content attribute here are no different that for other URLs specified in a document, and the spec already details what the requirements are, as the comments for that stackoverflow question point out.
Then the HTML5 spec needs rewording. "[T]he specification that registered Refresh as a header field" (comment 2) is HTML5, sec. 4.2.5.3. In that section, see Refresh State, step 23, which says, "Resolve the url value to an absolute URL, relative to the meta element." While the content can be an absolute URL, this step implies that the content can instead just be a relative URL, but one that is "relative to the meta element", rather than to the base element. If it is really the same as for other URLs that can be relative to the base element, then this section should say that or should omit the phrasing about "relative to the meta element". What is now in the HTML5 step is confusing, which I think is why the stackoverflow question was introduced in the first place.
(In reply to Nick Levinson from comment #4) > is HTML5, sec. 4.2.5.3. In that > section, see Refresh State, step 23, which says, "Resolve the url value to > an absolute URL, relative to the meta element." That's a conformance requirement for UA implementors, not a statement for authors. If you press the "Add developer views" button in the document, that entire section will disappear. > While the content can be an > absolute URL, this step implies that the content can instead just be a > relative URL, but one that is "relative to the meta element", rather than to > the base element. If it is really the same as for other URLs that can be > relative to the base element, then this section should say that or should > omit the phrasing about "relative to the meta element". What is now in the > HTML5 step is confusing, It's only confusing if you try to read it as being something that's intended for authors to need to understand, as opposed to UA implementors. > which I think is why the stackoverflow question was > introduced in the first place. I thought you were the same person who wrote that stackoverflow question.
There were two stackoverflow topics. I started the second one because I couldn't respond on the one begun by someone else, so in my opening post I linked the second topic to the other person's, which was <http://stackoverflow.com/questions/29645711/in-meta-tag-url-with-slash-vs-url-without-slash>. I'd been trying to figure out what that button was for, since I didn't see it cause anything elsewhere. Deleting is an unexpected result for the button label. Thanks.