gecko-dev/parser/html/nsHtml5Highlighter.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

445 lines
11 KiB
C
Raw Normal View History

/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#ifndef nsHtml5Highlighter_h
#define nsHtml5Highlighter_h
#include "nsCOMPtr.h"
#include "nsHtml5TreeOperation.h"
#include "nsHtml5UTF16Buffer.h"
#include "nsHtml5TreeOperation.h"
#include "nsAHtml5TreeOpSink.h"
#define NS_HTML5_HIGHLIGHTER_HANDLE_ARRAY_LENGTH 512
/**
* A state machine for generating HTML for display in View Source based on
* the transitions the tokenizer makes on the source being viewed.
*/
class nsHtml5Highlighter {
public:
/**
* The constructor.
*
* @param aOpSink the sink for the tree ops generated by this highlighter
*/
explicit nsHtml5Highlighter(nsAHtml5TreeOpSink* aOpSink);
/**
* The destructor.
*/
~nsHtml5Highlighter();
Bug 1701828 - meta charset rewrite. r=smaug Implements https://github.com/whatwg/html/issues/6962 . Improves performance when <meta charset> occurs in head but after the first kilobyte and aligns behavior better with WebKit and Blink. The main change is to avoid reloads when meta appears within head but after the first kilobyte. Prior to this change, Gecko reloaded in that case (in compliance with the spec!) even though WebKit and Blink did not. Differences from WebKit and Blink: * WebKit and Blink honor <meta charset> in <noscript>. This implementation does not. * WebKit and Blink look for meta as if the tree builder was unaware of foreign content. This implementation is foreign content-aware. This makes a difference for CDATA sections that contain a > before the meta as well as style and script elements within foreign content. This could happen if the CDATA section that has mysteriously been introduced around a what looks like a meta tag also contains another prior tag-looking run of text. * This implementation processes rel=preload and speculative loads that are seen before <meta charset> has been seen. WebKit and Blink instead first look for the meta and rewind before starting speculative parsing. * Unlike WebKit, if there is neither an honored meta nor syntax resembling an XML declaration, detection from content takes place (as in Blink). * Unlike Blink, if there is neither an honored meta nor syntax resembling an XML declaration, the detection from content is not dependent of network buffer boundaries. * Unlike Blink, detection from content can trigger a reload at the end of the stream if the guess made at that point differs from the first guess. (See below for the definition of the input to the first guess.) Differences from the old spec and Gecko previously: * Meta inside script and RCDATA elements is no longer honored. * Late meta is now ignored and no longer triggers a reload. * Later meta counts as early enough meta: In addition to the previous meta within the first 1024 bytes, now a meta that started within the first 1024 bytes counts as early enough. Additionally, if by then there hasn't been a template start tag and head hasn't ended, meta occurring before the earlier of the end of the head or a template start tag counts as early enough. * Meta now counts as not-late even if the encoding label has numeric character reference escapes. * Syntax resembling an XML declaration longer than a kilobyte is honored if there is no honored meta. * If there is neither an honored meta nor syntax resembling an XML declaration, the initial chardetng scan is potentially longer than before: the first 1024 bytes, the token spanning the 1024-byte boundary if there is such a token, and, if by then head hasn't ended and there hasn't been a template start tag until the end of the template start tag or the end of the token that causes head to end, ever comes first. However, if the token implying the end of the head is a text token, bytes only to the end of the previous non-text token is considered. (This definition avoids depending on network buffer boundaries.) * XML View Source now uses the code for syntax resembling an XML declaration instead of expat for extracting the internal encoding label. Reftest are added as both WPT and Gecko reftests in order to test both http: and file: URL scenarios. The Gecko tests retain the WPT <link> tags in order to use the exact same bytes. An encoding declaration has been added to a number of old tests that didn't intend to test the new speculation behavior especially in the context of https://bugzilla.mozilla.org/show_bug.cgi?id=1727750 . Differential Revision: https://phabricator.services.mozilla.com/D125808
2021-12-08 11:34:20 +00:00
/**
* Set the op sink (for speculation).
*/
void SetOpSink(nsAHtml5TreeOpSink* aOpSink);
/**
* Reset state to after generated head but before processing any of the input
* stream.
*/
void Rewind();
/**
* Starts the generated document.
*/
void Start(const nsAutoString& aTitle);
/**
* Updates the charset source via the op queue.
*/
void UpdateCharsetSource(nsCharsetSource aCharsetSource);
/**
* Report a tokenizer state transition.
*
* @param aState the state being transitioned to
* @param aReconsume whether this is a reconsuming transition
* @param aPos the tokenizer's current position into the buffer
*/
int32_t Transition(int32_t aState, bool aReconsume, int32_t aPos);
/**
* Report end of file.
*
* Returns `true` normally and `false` on OOM.
*/
[[nodiscard]] bool End();
/**
* Set the current buffer being tokenized
*/
void SetBuffer(nsHtml5UTF16Buffer* aBuffer);
/**
* Let go of the buffer being tokenized but first, flush text from it.
*
* @param aPos the first UTF-16 code unit not to flush
*/
void DropBuffer(int32_t aPos);
/**
* Query whether there are some many ops in the queue
* that they should be flushed now.
*
* @return true if FlushOps() should be called now
*/
bool ShouldFlushOps();
/**
* Flush the tree ops into the sink.
*
* @return Ok(true) if there were ops to flush, Ok(false)
* if there were no ops to flush and Err() on OOM.
*/
mozilla::Result<bool, nsresult> FlushOps();
/**
* Linkify the current attribute value if the attribute name is one of
* known URL attributes. (When executing tree ops, javascript: URLs will
* not be linkified, though.)
*
* @param aName the name of the attribute
* @param aValue the value of the attribute
*/
void MaybeLinkifyAttributeValue(nsHtml5AttributeName* aName,
nsHtml5String aValue);
/**
* Inform the highlighter that the tokenizer successfully completed a
* named character reference.
*/
void CompletedNamedCharacterReference();
/**
* Adds an error annotation to the node that's currently on top of
* mStack.
*
* @param aMsgId the id of the message in the property file
*/
void AddErrorToCurrentNode(const char* aMsgId);
/**
* Adds an error annotation to the node that corresponds to the most
* recently opened markup declaration/tag span, character reference or
* run of text.
*
* @param aMsgId the id of the message in the property file
*/
void AddErrorToCurrentRun(const char* aMsgId);
/**
* Adds an error annotation to the node that corresponds to the most
* recently opened markup declaration/tag span, character reference or
* run of text with one atom to use when formatting the message.
*
* @param aMsgId the id of the message in the property file
* @param aName the atom
*/
void AddErrorToCurrentRun(const char* aMsgId, nsAtom* aName);
/**
* Adds an error annotation to the node that corresponds to the most
* recently opened markup declaration/tag span, character reference or
* run of text with two atoms to use when formatting the message.
*
* @param aMsgId the id of the message in the property file
* @param aName the first atom
* @param aOther the second atom
*/
void AddErrorToCurrentRun(const char* aMsgId, nsAtom* aName, nsAtom* aOther);
/**
* Adds an error annotation to the node that corresponds to the most
* recent potentially character reference-starting ampersand.
*
* @param aMsgId the id of the message in the property file
*/
void AddErrorToCurrentAmpersand(const char* aMsgId);
/**
* Adds an error annotation to the node that corresponds to the most
* recent potentially self-closing slash.
*
* @param aMsgId the id of the message in the property file
*/
void AddErrorToCurrentSlash(const char* aMsgId);
/**
* Enqueues a tree op for adding base to the urls with the view-source:
*
* @param aValue the base URL to add
*/
void AddBase(nsHtml5String aValue);
Bug 1701828 - meta charset rewrite. r=smaug Implements https://github.com/whatwg/html/issues/6962 . Improves performance when <meta charset> occurs in head but after the first kilobyte and aligns behavior better with WebKit and Blink. The main change is to avoid reloads when meta appears within head but after the first kilobyte. Prior to this change, Gecko reloaded in that case (in compliance with the spec!) even though WebKit and Blink did not. Differences from WebKit and Blink: * WebKit and Blink honor <meta charset> in <noscript>. This implementation does not. * WebKit and Blink look for meta as if the tree builder was unaware of foreign content. This implementation is foreign content-aware. This makes a difference for CDATA sections that contain a > before the meta as well as style and script elements within foreign content. This could happen if the CDATA section that has mysteriously been introduced around a what looks like a meta tag also contains another prior tag-looking run of text. * This implementation processes rel=preload and speculative loads that are seen before <meta charset> has been seen. WebKit and Blink instead first look for the meta and rewind before starting speculative parsing. * Unlike WebKit, if there is neither an honored meta nor syntax resembling an XML declaration, detection from content takes place (as in Blink). * Unlike Blink, if there is neither an honored meta nor syntax resembling an XML declaration, the detection from content is not dependent of network buffer boundaries. * Unlike Blink, detection from content can trigger a reload at the end of the stream if the guess made at that point differs from the first guess. (See below for the definition of the input to the first guess.) Differences from the old spec and Gecko previously: * Meta inside script and RCDATA elements is no longer honored. * Late meta is now ignored and no longer triggers a reload. * Later meta counts as early enough meta: In addition to the previous meta within the first 1024 bytes, now a meta that started within the first 1024 bytes counts as early enough. Additionally, if by then there hasn't been a template start tag and head hasn't ended, meta occurring before the earlier of the end of the head or a template start tag counts as early enough. * Meta now counts as not-late even if the encoding label has numeric character reference escapes. * Syntax resembling an XML declaration longer than a kilobyte is honored if there is no honored meta. * If there is neither an honored meta nor syntax resembling an XML declaration, the initial chardetng scan is potentially longer than before: the first 1024 bytes, the token spanning the 1024-byte boundary if there is such a token, and, if by then head hasn't ended and there hasn't been a template start tag until the end of the template start tag or the end of the token that causes head to end, ever comes first. However, if the token implying the end of the head is a text token, bytes only to the end of the previous non-text token is considered. (This definition avoids depending on network buffer boundaries.) * XML View Source now uses the code for syntax resembling an XML declaration instead of expat for extracting the internal encoding label. Reftest are added as both WPT and Gecko reftests in order to test both http: and file: URL scenarios. The Gecko tests retain the WPT <link> tags in order to use the exact same bytes. An encoding declaration has been added to a number of old tests that didn't intend to test the new speculation behavior especially in the context of https://bugzilla.mozilla.org/show_bug.cgi?id=1727750 . Differential Revision: https://phabricator.services.mozilla.com/D125808
2021-12-08 11:34:20 +00:00
/**
* Starts a wrapper around a run of characters.
*/
void StartCharacters();
private:
/**
* Starts a span with no class.
*/
void StartSpan();
/**
* Starts a <span> and sets the class attribute on it.
*
* @param aClass the class to set (MUST be a static string that does not
* need to be released!)
*/
void StartSpan(const char16_t* aClass);
/**
* End the current <span> or <a> in the highlighter output.
*/
void EndSpanOrA();
/**
* Ends a wrapper around a run of characters.
*/
void EndCharactersAndStartMarkupRun();
/**
* Starts an <a>.
*/
void StartA();
/**
* Flushes characters up to but not including the current one.
*/
void FlushChars();
/**
* Flushes characters up to and including the current one.
*/
void FlushCurrent();
/**
* Finishes highlighting a tag in the input data by closing the open
* <span> and <a> elements in the highlighter output and then starts
* another <span> for potentially highlighting characters potentially
* appearing next.
*/
void FinishTag();
/**
* Adds a class attribute to the current node.
*
* @param aClass the class to set (MUST be a static string that does not
* need to be released!)
*/
void AddClass(const char16_t* aClass);
/**
* Allocates a handle for an element.
*
* See the documentation for nsHtml5TreeBuilder::AllocateContentHandle()
* in nsHtml5TreeBuilderHSupplement.h.
*
* @return the handle
*/
nsIContent** AllocateContentHandle();
/**
* Enqueues an element creation tree operation.
*
* @param aName the name of the element
* @param aAttributes the attribute holder (ownership will be taken) or
* nullptr for no attributes
* @param aIntendedParent the intended parent node for the created element
* @param aCreator the content creator function
* @return the handle for the element that will be created
*/
nsIContent** CreateElement(nsAtom* aName, nsHtml5HtmlAttributes* aAttributes,
nsIContent** aIntendedParent,
mozilla::dom::HTMLContentCreatorFunction aCreator);
/**
* Gets the handle for the current node. May be called only after the
* root element has been set.
*
* @return the handle for the current node
*/
nsIContent** CurrentNode();
/**
* Create an element and push it (its handle) on the stack.
*
* @param aName the name of the element
* @param aAttributes the attribute holder (ownership will be taken) or
* nullptr for no attributes
* @param aCreator the content creator function
*/
void Push(nsAtom* aName, nsHtml5HtmlAttributes* aAttributes,
mozilla::dom::HTMLContentCreatorFunction aCreator);
/**
* Pops the current node off the stack.
*/
void Pop();
/**
* Appends text content to the current node.
*
* @param aBuffer the buffer to copy from
* @param aStart the index of the first code unit to copy
* @param aLength the number of code units to copy
*/
void AppendCharacters(const char16_t* aBuffer, int32_t aStart,
int32_t aLength);
/**
* Enqueues a tree op for adding an href attribute with the view-source:
* URL scheme to the current node.
*
* @param aValue the (potentially relative) URL to link to
*/
void AddViewSourceHref(nsHtml5String aValue);
/**
* The state we are transitioning away from.
*/
int32_t mState;
/**
* The index of the first UTF-16 code unit in mBuffer that hasn't been
* flushed yet.
*/
int32_t mCStart;
/**
* The position of the code unit in mBuffer that caused the current
* transition.
*/
int32_t mPos;
/**
* The current line number.
*/
int32_t mLineNumber;
/**
* The number of inline elements open inside the <pre> excluding the
* span potentially wrapping a run of characters.
*/
int32_t mInlinesOpen;
/**
* Whether there's a span wrapping a run of characters (excluding CDATA
* section) open.
*/
bool mInCharacters;
/**
* The current buffer being tokenized.
*/
nsHtml5UTF16Buffer* mBuffer;
/**
* The outgoing tree op queue.
*/
nsTArray<nsHtml5TreeOperation> mOpQueue;
/**
Bug 1701828 - meta charset rewrite. r=smaug Implements https://github.com/whatwg/html/issues/6962 . Improves performance when <meta charset> occurs in head but after the first kilobyte and aligns behavior better with WebKit and Blink. The main change is to avoid reloads when meta appears within head but after the first kilobyte. Prior to this change, Gecko reloaded in that case (in compliance with the spec!) even though WebKit and Blink did not. Differences from WebKit and Blink: * WebKit and Blink honor <meta charset> in <noscript>. This implementation does not. * WebKit and Blink look for meta as if the tree builder was unaware of foreign content. This implementation is foreign content-aware. This makes a difference for CDATA sections that contain a > before the meta as well as style and script elements within foreign content. This could happen if the CDATA section that has mysteriously been introduced around a what looks like a meta tag also contains another prior tag-looking run of text. * This implementation processes rel=preload and speculative loads that are seen before <meta charset> has been seen. WebKit and Blink instead first look for the meta and rewind before starting speculative parsing. * Unlike WebKit, if there is neither an honored meta nor syntax resembling an XML declaration, detection from content takes place (as in Blink). * Unlike Blink, if there is neither an honored meta nor syntax resembling an XML declaration, the detection from content is not dependent of network buffer boundaries. * Unlike Blink, detection from content can trigger a reload at the end of the stream if the guess made at that point differs from the first guess. (See below for the definition of the input to the first guess.) Differences from the old spec and Gecko previously: * Meta inside script and RCDATA elements is no longer honored. * Late meta is now ignored and no longer triggers a reload. * Later meta counts as early enough meta: In addition to the previous meta within the first 1024 bytes, now a meta that started within the first 1024 bytes counts as early enough. Additionally, if by then there hasn't been a template start tag and head hasn't ended, meta occurring before the earlier of the end of the head or a template start tag counts as early enough. * Meta now counts as not-late even if the encoding label has numeric character reference escapes. * Syntax resembling an XML declaration longer than a kilobyte is honored if there is no honored meta. * If there is neither an honored meta nor syntax resembling an XML declaration, the initial chardetng scan is potentially longer than before: the first 1024 bytes, the token spanning the 1024-byte boundary if there is such a token, and, if by then head hasn't ended and there hasn't been a template start tag until the end of the template start tag or the end of the token that causes head to end, ever comes first. However, if the token implying the end of the head is a text token, bytes only to the end of the previous non-text token is considered. (This definition avoids depending on network buffer boundaries.) * XML View Source now uses the code for syntax resembling an XML declaration instead of expat for extracting the internal encoding label. Reftest are added as both WPT and Gecko reftests in order to test both http: and file: URL scenarios. The Gecko tests retain the WPT <link> tags in order to use the exact same bytes. An encoding declaration has been added to a number of old tests that didn't intend to test the new speculation behavior especially in the context of https://bugzilla.mozilla.org/show_bug.cgi?id=1727750 . Differential Revision: https://phabricator.services.mozilla.com/D125808
2021-12-08 11:34:20 +00:00
* The tree op stage for the tree op executor or a speculation when looking
* for meta charset.
*
* The op sink is owned by the nsHtml5TreeOpExecutor, which outlives this
* object, because this object is owned by the nsHtml5Tokenizer instance that
* is owned by the nsHtml5StreamParser, which keeps the executor alive via
* nsHtml5Streamparser::mExecutorFlusher.
*/
nsAHtml5TreeOpSink* mOpSink;
/**
* The most recently opened markup declaration/tag or run of characters.
*/
nsIContent** mCurrentRun;
/**
* The most recent ampersand in a place where character references were
* allowed.
*/
nsIContent** mAmpersand;
/**
* The most recent slash that might become a self-closing slash.
*/
nsIContent** mSlash;
/**
* Memory for element handles.
*/
mozilla::UniquePtr<nsIContent*[]> mHandles;
/**
* Number of handles used in mHandles
*/
int32_t mHandlesUsed;
/**
* A holder for old contents of mHandles
*/
nsTArray<mozilla::UniquePtr<nsIContent*[]>> mOldHandles;
/**
* The element stack.
*/
nsTArray<nsIContent**> mStack;
/**
* The string "comment"
*/
static char16_t sComment[];
/**
* The string "cdata"
*/
static char16_t sCdata[];
/**
* The string "start-tag"
*/
static char16_t sStartTag[];
/**
* The string "attribute-name"
*/
static char16_t sAttributeName[];
/**
* The string "attribute-value"
*/
static char16_t sAttributeValue[];
/**
* The string "end-tag"
*/
static char16_t sEndTag[];
/**
* The string "doctype"
*/
static char16_t sDoctype[];
/**
* The string "entity"
*/
static char16_t sEntity[];
/**
* The string "pi"
*/
static char16_t sPi[];
/**
* Whether base is already visited once.
*/
bool mSeenBase;
};
#endif // nsHtml5Highlighter_h