Previously in Browser, when we navigate back from a page that has an
icon to a page that does not have an icon, the icon does not update and
the old icon is displayed because FrameLoader does not set the default
favicon when the favicon cannot be loaded. This patch ensures that
Browser receives a new icon bitmap every time a load takes place.
Change all the places that were including the deprecated parser, to
include the new one instead, and then delete the old parser code.
`ParentNode::query_selector[_all]()` now treat their input as a
comma-separated list of selectors, instead of just one, and return
elements that match any of the selectors in that list. This is according
to these specs:
- querySelector/querySelectorAll:
https://dom.spec.whatwg.org/#ref-for-dom-parentnode-queryselector%E2%91%A0
- selector matching algorithm:
https://www.w3.org/TR/selectors-4/#match-against-tree
To transparently support multi-frame images, all decoder plugins have
already been updated to return their only bitmap for frame(0).
This patch completes the remaining cleanup work by removing the
ImageDecoder::bitmap() API and having all clients call frame() instead.
Previously, ImageDecoder::create() would return a NonnullRefPtr and
could not "fail", although the returned decoder may be "invalid" which
you then had to check anyway.
The new interface looks like this:
static RefPtr<Gfx::ImageDecoder> try_create(ReadonlyBytes);
This simplifies ImageDecoder since it no longer has to worry about its
validity. Client code gets slightly clearer as well.
Making a bitmap non-volatile after being volatile may fail to allocate
physical pages after the kernel stole the old pages in a purge.
This is different from the pages being purged, but reallocated. In that
case, they are simply replaced with zero-fill-on-demand pages as if
they were freshly allocated.
The new one is the same as the old one, just in the new Parser's
source files. This isn't the most elegant solution but it seemed
like the best option. And it's all temporary, after all.
The LexicalPath instance methods dirname(), basename(), title() and
extension() will be changed to return StringView const& in a further
commit. Due to this, users creating temporary LexicalPath objects just
to call one of those getters will recieve a StringView const& pointing
to a possible freed buffer.
To avoid this, static methods for those APIs have been added, which will
return a String by value to avoid those problems. All cases where
temporary LexicalPath objects have been used as described above haven
been changed to use the static APIs.
This changes the RequestClient::start_request() method to take a URL
object instead of a URL string as argument. All callers of the method
already had a URL object anyway, and start_request() in turn parses the
URL string back into a URL object. This removes this unnecessary
conversion.
Our "frame" concept very closely matches what the web specs call a
"browsing context", so let's rename it to that. :^)
The "main frame" becomes the "top-level browsing context",
and "sub-frames" are now "nested browsing contexts".
Surprisingly this is not used by the browser's page reload functionality
but only JS's location.reload() - that's probably why this hasn't been
noticed yet. Make sure we notify the page client about the load start in
that case as well. :^)
Otherwise we would sometimes (dependent on the load time, I believe) end
up setting the document and eventually calling title change callbacks
before communicating that the page started loading.
We had two functions for doing mostly the same thing. Combine both
of them into String::find() and use that everywhere.
Also add some tests to cover basic behavior.
This patch implements the HTML specification's "encoding sniffing
algorithm", which is used when no encoding can be obtained from the
Content-Type header (either because it doesn't contain a charset=...)
value or the file has not been opened via HTTP (as with local files).
It also modifies the creator of the HTMLDocumentParser to use the new
HTMLDocumentParser::create_with_uncertain_encoding static method, which
runs the encoding sniffing algorithm before instantiating the parser.
This now allows us to load local HTML pages (or remote pages without a
charset specified in the 'Content-Type' header) with a non-UTF-8
encoding such as 'windows-1252'. This would previously crash the
browser. :^)
This modifies the Document class to use Optional<String> for the
encoding. If the encoding is unknown, the Optional will not have a
value. It also implements the has_encoding() and encoding_or_default()
instance methods, the latter of which will return "UTF-8" as a fallback
if no encoding is present.
The usage of Optional<String> instead of the null string is part of an
effort to explicitly indicate that a string could not have a value.
This also modifies the former callers of encoding() to use
encoding_or_default(). Furthermore, the encoding will now only be set if
it is actually known, rather than just guessed by earlier code.
This modifies the Resource class to use Optional<String> for the
encoding. If the encoding is unknown, the Optional will not have a
value (instead of using the null state of the String class). It also
implements a has_encoding() instance method and modifies the callers
of Resource::encoding() to use the new API.
This patch changes the encoding_from_content_type function to only
return an encoding if it actually finds one, and leave it up to the
caller to decided on a default to use.
It also modifies the caller to expect an Optional<String> (instead of
relying on the null state of the String class) as a return value and
separates the encoding and MIME type determination. This will be built
upon in a further commit.
This prevents the browser from crashing when trying to load an infinite
redirects loop. The chosen limit is based on the fetch specification:
"If request's redirect count is twenty, return a network error."
A Frame now knows about its nesting-level.
The FrameLoader checks whether the recursion level of the current
frame allows it to be displayed and if not doesn't even load the
requested resource.
The nesting-check is done on a per-URL-basis, so there can be many many
nested Frames as long as they have different URLs.
If there are however Frames with the same URL nested inside each other
we only allow this to happen 3 times.
This mitigates infinetely recursing <iframe>s in an HTML-document
crashing the browser with an OOM.
The current ProtocolServer was really only used for requests, and with
the recent introduction of the WebSocket service, long-lasting
connections with another server are not part of it. To better reflect
this, this commit renames it to RequestServer.
This commit also changes the existing 'protocol' portal to 'request',
the existing 'protocol' user and group to 'request', and most mentions
of the 'download' aspect of the request to 'request' when relevant, to
make everything consistent across the system.
Note that LibProtocol still exists as-is, but the more generic Client
class and the more specific Download class have both been renamed to a
more accurate RequestClient and Request to match the new names.
This commit only change names, not behaviors.
SPDX License Identifiers are a more compact / standardized
way of representing file license information.
See: https://spdx.dev/resources/use/#identifiers
This was done with the `ambr` search and replace tool.
ambr --no-parent-ignore --key-from-file --rep-from-file key.txt rep.txt *
This error page template is slightly hilarious and should probably
be replaced with AK::SourceGenerator or some such, but for now let's
just get rid of the call to String::format().
A website with a 4xx status code is still a valid website, we should not
artificially hide it. In fact, many websites implement custom 404 error
pages for example, often containing search functionality or links back
to the homepage.
This might have implications regarding the loading of stylesheets where
the request 404s, but since we were not handling 5xx status codes as
errors either, I think that's fine for now (but might need additional
work later). Worst case, the parser rejects to load some error page HTML
as CSS :^)
This required changing the load_sync API to take a LoadRequest instead
of just a URL. Since HTMLScriptElement was the only (non-test) user of
this API, it didn't seem useful to instead add an overload of load_sync
for this.
Instead of all interested parties needing to write out the code to get
the cookie value for a load request, add a static helper to do it in
one location.
To implement the HttpOnly attribute, the CookieJar needs to know where a
request originated from. Namely, it needs to distinguish between HTTP /
non-HTTP (i.e. JavaScript) requests. When the HttpOnly attribute is set,
requests from JavaScript are to be blocked.
