[whatwg] Fragments included in Application Cache master entries

[Andrew Grieve]

The problem I'm seeing is not with URLs listed in the manifest, but rather
for URLs added to the Application Cache because they were navigated to and
pointed to an existing manifest. I think the following excerpt is the
relevant part of the spec:

 Master entries
> Documents that were added to the cache because a browsing context<http://www.whatwg.org/specs/web-apps/current-work/#browsing-context>was
> navigated <http://www.whatwg.org/specs/web-apps/current-work/#navigate> to
> that document and the document indicated that this was its cache, using the
> manifest<http://www.whatwg.org/specs/web-apps/current-work/#attr-html-manifest>attribute.


Andrew

On Mon, Jul 6, 2009 at 7:13 PM, Maria Khomenko <maria.khomenko at gmail.com>wrote:

> Actually, I believe the spec does address the question in the following
> passage (this is in the manifest parsing algorithm):
>
>
>> If mode is "explicit"
>
> Resolve the first item in tokens, relative to base URL; ignore the rest.
>
>
>> If this fails, then jump back to the step labeled "start of line".
>
>
>> If the resulting absolute URL has a different <scheme> component than the
>> manifest's URL (compared in an ASCII case-insensitive manner), then jump
>> back to the step labeled "start of line".
>
>
>> *Drop the <fragment> component of the resulting absolute URL, if it has
>> one.*
>
>
>> Add the resulting absolute URL to the explicit URLs.
>
>
> So it must be an implementation issue.
>
> Cheers,
> Maria
>
> On Mon, Jul 6, 2009 at 3:37 PM, Andrew Grieve <agrieve at google.com> wrote:
> >
> > The current behavior in Webkit is for URL fragments to be stored in the
> > URLs for master entries. I believe this to be a bug in Webkit, but cannot
> > determine from the spec if this is or not.
> >
> > Example:
> >
> > 1. Navigate to: http://www.thecssninja.com/demo/offline_webapp/#foo
> > 2. Go offline
> > 3. Do a browser refresh and ensure the page refreshes from AppCache.
> > 4. Change the URL hash to #bar
> > 5. Do a browser refresh and notice that it fails to load.
> >
> > I've filed a bug on webkit.org (
> https://bugs.webkit.org/show_bug.cgi?id=26925)
> > regarding this, but realized that the spec is unclear about what is
> > expected here. Since fragments are not sent to servers, I can't see
> > why they would be included in the master entry URLs as they make no
> > difference in the content that is served.
> >
> > Anyone know if the spec does in fact address this issue?
> >
> > Andrew
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.whatwg.org/pipermail/whatwg-whatwg.org/attachments/20090706/8ccd3a96/attachment-0002.htm>