|
View:
New views
5 Messages
—
Rating Filter:
Alert me
|
|
|
Content-Location as Base URIHi all, I have a question about the Content-Location HTTP Header. In Section 14.14 of RFC 2616, it says: The value of Content-Location also defines the base URI for the entity. However, a lot of user agents do not support this feature: http://jigsaw.w3.org/HTTP/CL/ https://bugzilla.mozilla.org/show_bug.cgi?id=109553 And there are other alternatives which are much reliable, such as the HTML "base" element and XML Base. So, should the base URI feature be marked as obsolete or deprecated? Regards, Franklin Tse |
|
|
Re: Content-Location as Base URIOn mån, 2007-12-10 at 13:17 +0800, Franklin Tse wrote:
> So, should the base URI feature be marked as obsolete or deprecated? Not entirely sure what the conclusion was the last time this was discussed some months ago.. It's a nice feature, helping unifying the URL-namespace on the server. But browser vendors do not dare implementing it due to there being a handful number of noticeably broken servers out there sending invalid/nonworking Content-Location headers. (mostly caused by rewriting reverse-proxies not noticing the existence of Content-Location) Opera tried, but backed out again due to customer complains blaiming Opera.. And no, there isn't really any good alternatives. It's true that HTML and XML has alternatives, but that's within the content itself. For example you can't add a HTML base href to a PDF document. Regards Henrik |
|
|
Re: Content-Location as Base URIOn Monday 2007-12-10 06:34 +0100, Henrik Nordstrom wrote: > But browser vendors do not dare implementing it due to there being a > handful number of noticeably broken servers out there sending It's not just broken servers. With the list of Mozilla's Content-Location bugs: https://bugzilla.mozilla.org/showdependencytree.cgi?id=238654&maxdepth=1&hide_resolved=0 I found one bug that's not about a broken server: https://bugzilla.mozilla.org/show_bug.cgi?id=241981 It's about Apache serving a content-negotiated document at a URL of the form http://example.com/doc . For browsers supporting XHTML, it serves doc.xhtml; for other browsers, it serves an equivalent doc.html. The document contains links to anchors within itself (e.g., <a href="#intro">Introduction</a>); links in the document are relative to Content-Location. Thus, clicking one of those links takes the browser to http://example.com/doc.xhtml#intro, which is not a URL intended to be linked to or exchanged. -David -- L. David Baron http://dbaron.org/ Mozilla Corporation http://www.mozilla.com/ |
|
|
Re: Content-Location as Base URIOn mån, 2007-12-10 at 00:41 -0800, L. David Baron wrote:
> On Monday 2007-12-10 06:34 +0100, Henrik Nordstrom wrote: > > But browser vendors do not dare implementing it due to there being a > > handful number of noticeably broken servers out there sending > > It's not just broken servers. With the list of Mozilla's > Content-Location bugs: > https://bugzilla.mozilla.org/showdependencytree.cgi?id=238654&maxdepth=1&hide_resolved=0 > I found one bug that's not about a broken server: > https://bugzilla.mozilla.org/show_bug.cgi?id=241981 > > It's about Apache serving a content-negotiated document at a URL of > the form http://example.com/doc . For browsers supporting XHTML, it > serves doc.xhtml; for other browsers, it serves an equivalent > doc.html. The document contains links to anchors within itself > (e.g., <a href="#intro">Introduction</a>); links in the document are > relative to Content-Location. Thus, clicking one of those links > takes the browser to http://example.com/doc.xhtml#intro, which is > not a URL intended to be linked to or exchanged. empty relative reference is within the current document. Content-Location, base href or any other means of specifying the intended location of the current document should not change that. rfc3986 setion 4.4. Content-Location (and base href) does not change the location of the object, only the Base-URI. Regards Henrik |
|
|
Re: Content-Location as Base URIOn Mon, 10 Dec 2007 09:41:38 +0100, L. David Baron <dbaron@...> wrote: > It's about Apache serving a content-negotiated document at a URL of > the form http://example.com/doc . For browsers supporting XHTML, it > serves doc.xhtml; for other browsers, it serves an equivalent > doc.html. The document contains links to anchors within itself > (e.g., <a href="#intro">Introduction</a>); links in the document are > relative to Content-Location. Thus, clicking one of those links > takes the browser to http://example.com/doc.xhtml#intro, which is > not a URL intended to be linked to or exchanged. That is actually bug https://bugzilla.mozilla.org/show_bug.cgi?id=275689 although there's no real conclusion there either it seems. -- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/> |
| Free embeddable forum powered by Nabble | Forum Help |