Page Speed Optimization Libraries
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Pages
Public Types | Public Member Functions | Static Public Member Functions | Static Public Attributes | Protected Member Functions | Friends | List of all members
net_instaweb::RewriteDriver Class Reference

#include "rewrite_driver.h"

Inheritance diagram for net_instaweb::RewriteDriver:

Public Types

enum  CssResolutionStatus { kWriteFailed, kNoResolutionNeeded, kSuccess }
 Status return-code for ResolveCssUrls.
enum  WaitMode { kNoWait, kWaitForCompletion, kWaitForCachedRender, kWaitForShutDown }
 Mode for BoundedWaitForCompletion. More...
enum  XhtmlStatus { kXhtmlUnknown, kIsXhtml, kIsNotXhtml }
enum  InlineAuthorizationPolicy { kInlineUnauthorizedResources, kInlineOnlyAuthorizedResources }
 See CreateInputResource.
enum  IntendedFor { kIntendedForInlining, kIntendedForGeneral }
 See CreateInputResource.

Public Member Functions

 RewriteDriver (MessageHandler *message_handler, FileSystem *file_system, UrlAsyncFetcher *url_async_fetcher)
virtual ~RewriteDriver ()
RewriteDriverClone ()
void Clear ()
void SetServerContext (ServerContext *server_context)
bool MayCacheExtendCss () const
bool MayCacheExtendImages () const
bool MayCacheExtendPdfs () const
bool MayCacheExtendScripts () const
const GoogleStringuser_agent () const
const RequestPropertiesrequest_properties () const
void ClearRequestProperties ()
 Reinitializes request_properties_, clearing any cached values.
bool write_property_cache_dom_cohort () const
void set_write_property_cache_dom_cohort (bool x)
void PropertyCacheSetupDone ()
RequestContextPtr request_context ()
void set_request_context (const RequestContextPtr &x)
RequestTracetrace_context ()
void TracePrintf (const char *fmt,...)
void TraceLiteral (const char *literal)
void TraceString (const GoogleString &s)
ResponseHeadersmutable_response_headers ()
const ResponseHeadersresponse_headers ()
void set_response_headers_ptr (ResponseHeaders *headers)
void SetRequestHeaders (const RequestHeaders &headers)
const RequestHeadersrequest_headers () const
UserAgentMatcheruser_agent_matcher () const
void AddFilters ()
void AddOwnedEarlyPreRenderFilter (HtmlFilter *filter)
void PrependOwnedPreRenderFilter (HtmlFilter *filter)
 Adds a filter to the beginning of the pre-render chain, taking ownership.
void AppendOwnedPreRenderFilter (HtmlFilter *filter)
 Adds a filter to the end of the pre-render chain, taking ownership.
void AppendUnownedPreRenderFilter (HtmlFilter *filter)
 Same, without taking ownership.
void AddOwnedPostRenderFilter (HtmlFilter *filter)
 Adds a filter to the end of the post-render chain, taking ownership.
void AddUnownedPostRenderFilter (HtmlFilter *filter)
 Same, without taking ownership.
void AppendRewriteFilter (RewriteFilter *filter)
void PrependRewriteFilter (RewriteFilter *filter)
void AddResourceUrlClaimant (ResourceUrlClaimant *claimant)
void SetWriter (Writer *writer)
Writerwriter () const
bool FetchResource (const StringPiece &url, AsyncFetch *fetch)
void FetchInPlaceResource (const GoogleUrl &gurl, bool proxy_mode, AsyncFetch *async_fetch)
bool FetchOutputResource (const OutputResourcePtr &output_resource, RewriteFilter *filter, AsyncFetch *async_fetch)
OutputResourcePtr DecodeOutputResource (const GoogleUrl &url, RewriteFilter **filter) const
bool DecodeOutputResourceName (const GoogleUrl &url, const RewriteOptions *options_to_use, const UrlNamer *url_namer, ResourceNamer *name_out, OutputResourceKind *kind_out, RewriteFilter **filter_out) const
bool LookupMetadataForOutputResource (StringPiece url, GoogleString *error_out, RewriteContext::CacheLookupResultCallback *callback)
bool DecodeUrl (const GoogleUrl &url, StringVector *decoded_urls) const
 Decodes the incoming pagespeed url to original url(s).
bool DecodeUrlGivenOptions (const GoogleUrl &url, const RewriteOptions *options, const UrlNamer *url_namer, StringVector *decoded_urls) const
FileSystemfile_system ()
UrlAsyncFetcherasync_fetcher ()
void SetSessionFetcher (UrlAsyncFetcher *f)
CacheUrlAsyncFetcherCreateCacheFetcher ()
CacheUrlAsyncFetcherCreateCacheOnlyFetcher ()
 Returns a cache fetcher that does not fall back to an actual fetcher.
ServerContextserver_context () const
Statisticsstatistics () const
void set_custom_options (RewriteOptions *options)
 Takes ownership of 'options'.
void set_options_for_pool (RewriteDriverPool *pool, RewriteOptions *options)
RewriteDriverPoolcontrolling_pool ()
 Pool in which this driver can be recycled. May be NULL.
const RewriteOptionsoptions () const
 Return the options used for this RewriteDriver.
virtual bool StartParseId (const StringPiece &url, const StringPiece &id, const ContentType &content_type)
virtual void FinishParse ()
void FinishParseAsync (Function *callback)
void InfoAt (const RewriteContext *context, const char *msg,...) INSTAWEB_PRINTF_FORMAT(3
void bool GenerateOutputResourceNameAndUrl (const UrlSegmentEncoder *encoder, const ResourceContext *data, const ResourcePtr &input_resource, GoogleString *name, GoogleUrl *mapped_gurl, GoogleString *failure_reason)
 Constructs name and URL for the specified input resource and encoder.
OutputResourcePtr CreateOutputResourceFromResource (const char *filter_id, const UrlSegmentEncoder *encoder, const ResourceContext *data, const ResourcePtr &input_resource, OutputResourceKind kind, GoogleString *failure_reason)
OutputResourcePtr CreateOutputResourceWithPath (const StringPiece &mapped_path, const StringPiece &unmapped_path, const StringPiece &base_url, const StringPiece &filter_id, const StringPiece &name, OutputResourceKind kind, GoogleString *failure_reason)
void PopulateResourceNamer (const StringPiece &filter_id, const StringPiece &name, ResourceNamer *full_name)
OutputResourcePtr CreateOutputResourceWithUnmappedUrl (const GoogleUrl &unmapped_gurl, const StringPiece &filter_id, const StringPiece &name, OutputResourceKind kind, GoogleString *failure_reason)
OutputResourcePtr CreateOutputResourceWithMappedPath (const StringPiece &mapped_path, const StringPiece &unmapped_path, const StringPiece &filter_id, const StringPiece &name, OutputResourceKind kind, GoogleString *failure_reason)
OutputResourcePtr CreateOutputResourceWithPath (const StringPiece &path, const StringPiece &filter_id, const StringPiece &name, OutputResourceKind kind, GoogleString *failure_reason)
ResourcePtr CreateInputResource (const GoogleUrl &input_url, bool *is_authorized)
ResourcePtr CreateInputResource (const GoogleUrl &input_url, InlineAuthorizationPolicy inline_authorization_policy, IntendedFor intended_for, bool *is_authorized)
ResourcePtr CreateInputResourceAbsoluteUncheckedForTestsOnly (const StringPiece &absolute_url)
bool IsResourceUrlClaimed (const GoogleUrl &url) const
bool MatchesBaseUrl (const GoogleUrl &input_url) const
bool MayRewriteUrl (const GoogleUrl &domain_url, const GoogleUrl &input_url, InlineAuthorizationPolicy inline_authorization_policy, IntendedFor intended_for, bool *is_authorized_domain) const
const GoogleUrlbase_url () const
StringPiece fetch_url () const
 The URL that was requested if FetchResource was called.
const GoogleUrldecoded_base_url () const
StringPiece decoded_base () const
bool IsHttps () const
 Quick way to tell if the document url is https (ie was fetched via https).
const UrlSegmentEncoderdefault_encoder () const
RewriteFilterFindFilter (const StringPiece &id) const
 Finds a filter with the given ID, or returns NULL if none found.
bool refs_before_base ()
 Returns refs_before_base.
void set_refs_before_base ()
StringPiece containing_charset ()
void set_containing_charset (const StringPiece charset)
HtmlResourceSlotPtr GetSlot (const ResourcePtr &resource, HtmlElement *elt, HtmlElement::Attribute *attr)
InlineResourceSlotPtr GetInlineSlot (const ResourcePtr &resource, HtmlCharactersNode *char_node)
InlineAttributeSlotPtr GetInlineAttributeSlot (const ResourcePtr &resource, HtmlElement *element, HtmlElement::Attribute *attribute)
SrcSetSlotCollectionPtr GetSrcSetSlotCollection (CommonFilter *filter, HtmlElement *element, HtmlElement::Attribute *attr)
bool InitiateRewrite (RewriteContext *rewrite_context) LOCKS_EXCLUDED(rewrite_mutex())
void InitiateFetch (RewriteContext *rewrite_context)
void RewriteComplete (RewriteContext *rewrite_context, bool permit_render)
void ReportSlowRewrites (int num)
void Cleanup ()
void AddUserReference ()
GoogleString ToString (bool show_detached_contexts) const LOCKS_EXCLUDED(rewrite_mutex())
 Debugging routines to print out data about the driver.
GoogleString ToStringLockHeld (bool show_detached_contexts) const EXCLUSIVE_LOCKS_REQUIRED(rewrite_mutex())
void PrintState (bool show_detached_contexts)
 For debugging.
void PrintStateToErrorLog (bool show_detached_contexts)
 For logs.
void WaitForCompletion ()
void WaitForShutDown ()
void BoundedWaitFor (WaitMode mode, int64 timeout_ms) LOCKS_EXCLUDED(rewrite_mutex())
void set_fully_rewrite_on_flush (bool x)
bool fully_rewrite_on_flush () const
 Returns if this response has a blocking rewrite or not.
void set_fast_blocking_rewrite (bool x)
bool fast_blocking_rewrite () const
void EnableBlockingRewrite (RequestHeaders *request_headers)
void set_externally_managed (bool x)
void DetachFetch ()
void DetachedFetchComplete ()
void FetchComplete ()
void DeleteRewriteContext (RewriteContext *rewrite_context)
int rewrite_deadline_ms ()
void set_max_page_processing_delay_ms (int x)
int max_page_processing_delay_ms ()
void set_device_type (UserAgentMatcher::DeviceType x)
 Sets the device type chosen for the current property_page.
UserAgentMatcher::DeviceType device_type () const
RewriteContextRegisterForPartitionKey (const GoogleString &partition_key, RewriteContext *candidate)
void DeregisterForPartitionKey (const GoogleString &partition_key, RewriteContext *candidate)
void RequestFlush ()
bool flush_requested () const
void ExecuteFlushIfRequested ()
void ExecuteFlushIfRequestedAsync (Function *callback)
virtual void Flush ()
void FlushAsync (Function *done)
void AddRewriteTask (Function *task)
 Queues up a task to run on the (high-priority) rewrite thread.
void AddLowPriorityRewriteTask (Function *task)
QueuedWorkerPool::Sequencehtml_worker ()
Sequencerewrite_worker ()
Scheduler::Sequencescheduler_sequence ()
QueuedWorkerPool::Sequencelow_priority_rewrite_worker ()
void RunTasksOnRequestThread ()
void SwitchToQueuedWorkerPool () EXCLUSIVE_LOCKS_REQUIRED(rewrite_mutex())
Schedulerscheduler ()
DomainRewriteFilterdomain_rewriter ()
UrlLeftTrimFilterurl_trim_filter ()
CssResolutionStatus ResolveCssUrls (const GoogleUrl &input_css_base, const StringPiece &output_css_base, const StringPiece &contents, Writer *writer, MessageHandler *handler)
bool ShouldAbsolutifyUrl (const GoogleUrl &input_base, const GoogleUrl &output_base, bool *proxy_mode) const
void UpdatePropertyValueInDomCohort (AbstractPropertyPage *page, StringPiece property_name, StringPiece property_value)
PropertyPageproperty_page () const
FallbackPropertyPagefallback_property_page () const
PropertyPageorigin_property_page () const
void set_property_page (PropertyPage *page)
 Takes ownership of page.
void set_fallback_property_page (FallbackPropertyPage *page)
 Takes ownership of page.
void set_unowned_fallback_property_page (FallbackPropertyPage *page)
 Does not take the ownership of the page.
void set_origin_property_page (PropertyPage *page)
 Takes ownership of page.
bool is_critical_images_beacon_enabled ()
CriticalImagesInfocritical_images_info () const
 Used by ImageRewriteFilter for identifying critical images.
CriticalSelectorInfocritical_selector_info ()
void set_critical_selector_info (CriticalSelectorInfo *info)
void set_critical_images_info (CriticalImagesInfo *critical_images_info)
bool FlattenCssImportsEnabled () const
int num_inline_preview_images () const
void increment_num_inline_preview_images ()
 We expect to this method to be called on the HTML parser thread.
void IncrementAsyncEventsCount ()
void DecrementAsyncEventsCount ()
 Decrements a reference count bumped up by IncrementAsyncEventsCount()
void IncrementRenderBlockingAsyncEventsCount ()
void DecrementRenderBlockingAsyncEventsCount ()
XhtmlStatus MimeTypeXhtmlStatus ()
void set_is_lazyload_script_flushed (bool x)
bool is_lazyload_script_flushed () const
FlushEarlyInfo * flush_early_info ()
 This method is not thread-safe. Call it only from the html parser thread.
DependencyTrackerdependency_tracker () const
bool DebugMode () const
void InsertDebugComment (StringPiece unescaped_message, HtmlNode *node)
void InsertDebugComments (const protobuf::RepeatedPtrField< GoogleString > &unescaped_messages, HtmlElement *element)
void InsertUnauthorizedDomainDebugComment (StringPiece url, HtmlElement *element)
AbstractLogRecordlog_record ()
DomStatsFilterdom_stats_filter () const
bool can_rewrite_resources () const
bool is_nested () const
 Determine whether this driver is nested inside another.
bool Write (const ResourceVector &inputs, const StringPiece &contents, const ContentType *type, StringPiece charset, OutputResource *output)
void set_defer_instrumentation_script (bool x)
bool defer_instrumentation_script () const
void set_num_initiated_rewrites (int64 x)
int64 num_initiated_rewrites () const
void set_num_detached_rewrites (int64 x)
 Sets the num_detached_rewrites_. This should only be called from test code.
int64 num_detached_rewrites () const
void set_pagespeed_query_params (StringPiece x)
StringPiece pagespeed_query_params () const
void set_pagespeed_option_cookies (StringPiece x)
StringPiece pagespeed_option_cookies () const
const GoogleStringCacheFragment () const
bool SetOrClearPageSpeedOptionCookies (const GoogleUrl &gurl, ResponseHeaders *response_headers)
bool Decode (StringPiece leaf, ResourceNamer *resource_namer) const
bool filters_added () const
bool has_html_writer_filter () const
void SetIsAmpDocument (bool is_amp)
bool is_amp_document () const
- Public Member Functions inherited from net_instaweb::HtmlParse
 HtmlParse (MessageHandler *message_handler)
void AddFilter (HtmlFilter *filter)
 Application methods for parsing functions and adding filters. More...
bool StartParse (const StringPiece &url)
bool StartParseWithType (const StringPiece &url, const ContentType &content_type)
bool is_url_valid () const
 Returns whether the google_url() URL is valid.
void SetUrlForTesting (const StringPiece &url)
void ParseText (const char *content, int size)
void ParseText (const StringPiece &sp)
HtmlCdataNodeNewCdataNode (HtmlElement *parent, const StringPiece &contents)
 Utility methods for implementing filters. More...
HtmlCharactersNodeNewCharactersNode (HtmlElement *parent, const StringPiece &literal)
HtmlCommentNodeNewCommentNode (HtmlElement *parent, const StringPiece &contents)
HtmlDirectiveNodeNewDirectiveNode (HtmlElement *parent, const StringPiece &contents)
HtmlIEDirectiveNodeNewIEDirectiveNode (HtmlElement *parent, const StringPiece &contents)
void InsertScriptAfterCurrent (StringPiece text, bool external)
void InsertScriptBeforeCurrent (StringPiece text, bool external)
HtmlElementAppendAnchor (StringPiece link, StringPiece text, HtmlElement *parent)
void InsertNodeBeforeNode (const HtmlNode *existing_node, HtmlNode *new_node)
void InsertNodeAfterNode (const HtmlNode *existing_node, HtmlNode *new_node)
void InsertElementBeforeElement (const HtmlNode *existing_element, HtmlNode *new_element)
void InsertElementAfterElement (const HtmlNode *existing_element, HtmlNode *new_element)
void PrependChild (const HtmlElement *existing_parent, HtmlNode *new_child)
void AppendChild (const HtmlElement *existing_parent, HtmlNode *new_child)
void InsertNodeBeforeCurrent (HtmlNode *new_node)
void InsertNodeAfterCurrent (HtmlNode *new_node)
bool AddParentToSequence (HtmlNode *first, HtmlNode *last, HtmlElement *new_parent)
bool MoveCurrentInto (HtmlElement *new_parent)
bool MoveCurrentBefore (HtmlNode *existing_node)
bool DeleteNode (HtmlNode *node)
bool DeleteSavingChildren (HtmlElement *element)
bool MakeElementInvisible (HtmlElement *element)
bool HasChildrenInFlushWindow (HtmlElement *element)
bool ReplaceNode (HtmlNode *existing_node, HtmlNode *new_node)
HtmlElementCloneElement (HtmlElement *in_element)
HtmlElementNewElement (HtmlElement *parent, const StringPiece &str)
HtmlElementNewElement (HtmlElement *parent, HtmlName::Keyword keyword)
HtmlElementNewElement (HtmlElement *parent, const HtmlName &name)
void AddAttribute (HtmlElement *element, HtmlName::Keyword keyword, const StringPiece &value)
void AddAttribute (HtmlElement *element, StringPiece name, const StringPiece &value)
void AddEscapedAttribute (HtmlElement *element, HtmlName::Keyword keyword, const StringPiece &escaped_value)
void SetAttributeName (HtmlElement::Attribute *attribute, HtmlName::Keyword keyword)
HtmlName MakeName (const StringPiece &str)
HtmlName MakeName (HtmlName::Keyword keyword)
bool IsRewritable (const HtmlNode *node) const
bool CanAppendChild (const HtmlNode *node) const
void ClearElements ()
void DebugLogQueue ()
 Log the HtmlEvent queue_ to the message_handler_ for debugging.
void DebugPrintQueue ()
 Print the HtmlEvent queue_ to stdout for debugging.
bool IsImplicitlyClosedTag (HtmlName::Keyword keyword) const
bool IsOptionallyClosedTag (HtmlName::Keyword keyword) const
bool TagAllowsBriefTermination (HtmlName::Keyword keyword) const
 Determines whether a tag allows brief termination in HTML, e.g. <tag>
MessageHandlermessage_handler () const
const char * url () const
const GoogleUrlgoogle_url () const
 Gets a parsed GoogleUrl& corresponding to url().
const char * id () const
int line_number () const
GoogleString UrlLine () const
 Returns URL (or id) and line number as a string, to be used in messages.
const DocTypedoctype () const
void Info (const char *filename, int line, const char *msg,...) INSTAWEB_PRINTF_FORMAT(4
 Interface for any caller to report an error message via the message handler.
void void Warning (const char *filename, int line, const char *msg,...) INSTAWEB_PRINTF_FORMAT(4
void void void Error (const char *filename, int line, const char *msg,...) INSTAWEB_PRINTF_FORMAT(4
void void void void FatalError (const char *filename, int line, const char *msg,...) INSTAWEB_PRINTF_FORMAT(4
void void void void void InfoV (const char *file, int line, const char *msg, va_list args)
void WarningV (const char *file, int line, const char *msg, va_list args)
void ErrorV (const char *file, int line, const char *msg, va_list args)
void FatalErrorV (const char *file, int line, const char *msg, va_list args)
void InfoHere (const char *msg,...) INSTAWEB_PRINTF_FORMAT(2
 Report error message with current parsing filename and linenumber.
void void WarningHere (const char *msg,...) INSTAWEB_PRINTF_FORMAT(2
void void void ErrorHere (const char *msg,...) INSTAWEB_PRINTF_FORMAT(2
void void void void FatalErrorHere (const char *msg,...) INSTAWEB_PRINTF_FORMAT(2
void void void void void ShowProgress (const char *message)
void InfoHereV (const char *msg, va_list args)
void WarningHereV (const char *msg, va_list args)
void ErrorHereV (const char *msg, va_list args)
void FatalErrorHereV (const char *msg, va_list args)
void AddElement (HtmlElement *element, int line_number)
void CloseElement (HtmlElement *element, HtmlElement::Style style, int line_number)
void ApplyFilter (HtmlFilter *filter)
 Run a filter on the current queue of parse nodes.
void set_timer (Timer *timer)
Timertimer () const
void set_log_rewrite_timing (bool x)
void add_event_listener (HtmlFilter *listener)
bool InsertComment (StringPiece sp)
void set_size_limit (int64 x)
 Sets the limit on the maximum number of bytes that should be parsed.
bool size_limit_exceeded () const
 Returns whether we have exceeded the size limit.
void SetDynamicallyDisabledFilterList (StringVector *list)
void DeferCurrentNode ()
void RestoreDeferredNode (HtmlNode *deferred_node)
bool can_modify_urls ()
 Returns whether the filter pipeline can rewrite urls.

Static Public Member Functions

static void InitStats (Statistics *statistics)
 Initialize statistics for all filters that need it.
static void Initialize ()
 Initialize statics. Initialize/Terminate calls must be paired.
static void Terminate ()
static GoogleString DeadlineExceededMessage (StringPiece filter_name)
 Formats a "deadline exceeded" message for a given filter.
static PropertyCache::CohortVector GetCohortList (const PropertyCache *pcache, const RewriteOptions *options, const ServerContext *server_context)
static GoogleString GenerateUnauthorizedDomainDebugComment (const GoogleUrl &gurl)
 Generates an unauthorized domain debug comment. Public for unit tests.
- Static Public Member Functions inherited from net_instaweb::HtmlParse
static bool IsLiteralTag (HtmlName::Keyword keyword)
static bool IsSometimesLiteralTag (HtmlName::Keyword keyword)

Static Public Attributes

static const char kDomCohort []
static const char kBeaconCohort []
 The cohort for properties that are written by the beacon handler.
static const char kDependenciesCohort []
static const char kLastRequestTimestamp []
static const char kParseSizeLimitExceeded []
 Tracks if we exceeded the maximum size limit of html which we should parse.
static const char kSubresourcesPropertyName []
 Flush Subresources Info associted with the HTML page.
static const char kStatusCodePropertyName []
 Status codes of previous responses.

Protected Member Functions

virtual void DetermineFiltersBehaviorImpl ()
- Protected Member Functions inherited from net_instaweb::HtmlParse
void BeginFinishParse ()
void EndFinishParse ()
void Clear ()
size_t GetEventQueueSize ()
 Returns the number of events on the event queue.
void DetermineFiltersBehavior ()
 Calls DetermineFiltersBehaviorImpl in an idempotent way.
void DetermineFilterListBehavior (const FilterList &list)
void CheckFilterBehavior (HtmlFilter *filter)
void set_buffer_events (bool x)
void DisableFiltersInjectingScripts ()
void DisableFiltersInjectingScripts (const FilterList &filters)
 Same, but over a passed-in list of filters.


class RewriteContext
class RewriteDriverTest
class RewriteTestBase
class ServerContextTest
class ScanFilter
class CategorizedRefcount< RewriteDriver, RefCategory >

Additional Inherited Members

- Protected Types inherited from net_instaweb::HtmlParse
typedef std::vector< HtmlFilter * > FilterVector
typedef std::list< HtmlFilter * > FilterList
typedef std::pair< HtmlNode
*, HtmlEventList * > 
typedef std::map< const
HtmlNode *, HtmlEventList * > 
typedef std::map< HtmlFilter
*, DeferredNode > 
typedef std::set< const
HtmlNode * > 

Detailed Description

This extends class HtmlParse (which should renamed HtmlContext) by providing context for rewriting resources (css, js, images).

Member Enumeration Documentation

Mode for BoundedWaitForCompletion.


Used internally. Do not pass in.


wait for everything to complete (up to deadline)


wait for at least cached rewrites to complete, and anything else that finishes within deadline.


Makes sure that all work, including any that's being done in background, finishes.

Indicates document's mimetype as XHTML, HTML, or is not known/something else. Note that in Apache we might not know the correct mimetype because a downstream module might change it. It's not clear how likely this is, since mod_rewrite and mod_mime run upstream of mod_pagespeed. However if anyone sets mimetype via "Header Add", it would affect the Browser's view of the document's mimetype (which is what determines the parsing) but mod_pagespeed would not know.

Note that we also have doctype().IsXhtml() but that indicates quirks-mode for CSS, and does not control how the parser parses the document.

Constructor & Destructor Documentation

virtual net_instaweb::RewriteDriver::~RewriteDriver ( )

Need explicit destructors to allow destruction of scoped_ptr-controlled instances without propagating the include files.

Member Function Documentation

void net_instaweb::RewriteDriver::AddFilters ( )

Adds the filters from the options, specified by name in enabled_filters. This must be called explicitly after object construction to provide an opportunity to programatically add custom filters beyond those defined in RewriteOptions, via AddFilter(HtmlFilter* filter) (below).

void net_instaweb::RewriteDriver::AddLowPriorityRewriteTask ( Function task)

Queues up a task to run on the low-priority rewrite thread. Such tasks are expected to be safely cancelable.

void net_instaweb::RewriteDriver::AddOwnedEarlyPreRenderFilter ( HtmlFilter filter)

Adds a filter to the very beginning of the pre-render chain, taking ownership. This should only be used for filters that must run before any filter added via PrependOwnedPreRenderFilter.

void net_instaweb::RewriteDriver::AddResourceUrlClaimant ( ResourceUrlClaimant claimant)

Tells RewriteDriver that a certain portion of URL namespace should not be handled via usual (HTTP proxy semantics) means. It's up to the filters to actually arrange for that to do something. Takes ownership of the claimant object. Note that it's important for the claims to be disjoint, since the RewriteContext framework needs to be able to assign compatible Resource objects for same URLs/slots among all filters that deal with them.

void net_instaweb::RewriteDriver::AddUserReference ( )

Adds an extra external reference to the object. You should not normally need to call it (NewRewriteDriver does it initially), unless for some reason you want to pin the object (e.g. in tests). Matches up with Cleanup.

void net_instaweb::RewriteDriver::AppendRewriteFilter ( RewriteFilter filter)

Add a RewriteFilter to the end of the pre-render chain and take ownership of the filter. This differs from AppendOwnedPreRenderFilter in that it adds the filter's ID into a dispatch table for serving rewritten resources. E.g. if your filter->id == "xy" and FetchResource("NAME.pagespeed.xy.HASH.EXT"...) is called, then RewriteDriver will dispatch to filter->Fetch().

This is used when the filter being added is not part of the core set built into RewriteDriver and RewriteOptions, such as platform-specific or server-specific filters, or filters invented for unit-testing the framework.

const GoogleUrl& net_instaweb::RewriteDriver::base_url ( ) const

Returns the appropriate base gurl to be used for resolving hrefs in the document. Note that HtmlParse::google_url() is the URL for the HTML file and is used for printing html syntax errors.

void net_instaweb::RewriteDriver::BoundedWaitFor ( WaitMode  mode,
int64  timeout_ms 

As above, but with a time bound, and taking a mode parameter to decide between WaitForCompletion or WaitForShutDown behavior. If timeout_ms <= 0, no time bound will be used.

const GoogleString& net_instaweb::RewriteDriver::CacheFragment ( ) const

We fragment the cache based on the hostname we got from the request, unless that was overridden in the options with a cache_fragment.

bool net_instaweb::RewriteDriver::can_rewrite_resources ( ) const

Determines whether the system is healthy enough to rewrite resources. Currently, systems get sick based on the health of the metadata cache.

void net_instaweb::RewriteDriver::Cleanup ( )

If there are not outstanding references to this RewriteDriver, delete it or recycle it to a free pool in the ServerContext. If this is a fetch, calling this also signals to the system that you are no longer interested in its results.

void net_instaweb::RewriteDriver::Clear ( )

Clears the current request cache of resources and base URL. The filter-chain is left intact so that a new request can be issued. Deletes all RewriteContexts.

WaitForCompletion must be called prior to Clear().

RewriteDriver* net_instaweb::RewriteDriver::Clone ( )

Returns a fresh instance using the same options we do, using the same log record. Drivers should only be cloned within the same request.

Clones share the same request_context, which contains bits derived from the request headers, so request_headers_ is also cloned (or shared if we make them shareable).

You must call SetRequestHeaders before calling Clone.

StringPiece net_instaweb::RewriteDriver::containing_charset ( )

Get/set the charset of the containing HTML page. See for an explanation of how this is determined, but NOTE that the determined charset can change as more of the HTML is seen, in particular after a meta tag.

CacheUrlAsyncFetcher* net_instaweb::RewriteDriver::CreateCacheFetcher ( )

Creates a cache fetcher that uses the driver's fetcher and its options. Note: this means the driver's fetcher must survive as long as this does.

ResourcePtr net_instaweb::RewriteDriver::CreateInputResource ( const GoogleUrl input_url,
bool *  is_authorized 

Creates an input resource based on input_url. Returns NULL if the input resource url isn't valid or is a data url, or can't legally be rewritten in the context of this page, in which case *is_authorized will be false. Assumes that resources from unauthorized domains may not be rewritten and that the resource is not intended exclusively for inlining.

ResourcePtr net_instaweb::RewriteDriver::CreateInputResource ( const GoogleUrl input_url,
InlineAuthorizationPolicy  inline_authorization_policy,
IntendedFor  intended_for,
bool *  is_authorized 

Creates an input resource. Returns NULL if the input resource url isn't valid or is a data url, or can't legally be rewritten in the context of this page (which could mean that it was a resource from an unauthorized domain being processed by a filter that does not allow unauthorized resources, in which case *is_authorized will be false).

There are two "special" options, and if you don't care about them you should just call CreateInputResource(input_url, is_authorized) to use their defaults:

  • If resources from unauthorized domains may be inlined, set inline_authorization_policy to kInlineUnauthorizedResources, otherwise set it to kInlineOnlyAuthorizedResources.
  • If this resource will be inlined after fetching, then set intended_for to kIntendedForInlining, otherwise use kIntendedForGeneral. This is to support AllowWhenInlining.
ResourcePtr net_instaweb::RewriteDriver::CreateInputResourceAbsoluteUncheckedForTestsOnly ( const StringPiece &  absolute_url)

Creates an input resource from the given absolute url. Requires that the provided url has been checked, and can legally be rewritten in the current page context. Only for use by unit tests.

OutputResourcePtr net_instaweb::RewriteDriver::CreateOutputResourceFromResource ( const char *  filter_id,
const UrlSegmentEncoder encoder,
const ResourceContext *  data,
const ResourcePtr input_resource,
OutputResourceKind  kind,
GoogleString failure_reason 

Creates a reference-counted pointer to a new OutputResource object.

The content type is taken from the input_resource, but can be modified with SetType later if that is not correct (e.g. due to image transcoding). Constructs an output resource corresponding to the specified input resource and encoded using the provided encoder. Assumes permissions checking occurred when the input resource was constructed, and does not do it again. To avoid if-chains, tolerates a NULL input_resource (by returning NULL).

TODO(jmaessen, jmarantz): Do we want to permit NULL input_resources here? jmarantz has evinced a distaste.
OutputResourcePtr net_instaweb::RewriteDriver::CreateOutputResourceWithMappedPath ( const StringPiece &  mapped_path,
const StringPiece &  unmapped_path,
const StringPiece &  filter_id,
const StringPiece &  name,
OutputResourceKind  kind,
GoogleString failure_reason 

Version of CreateOutputResourceWithPath where the unmapped and mapped paths are different and the base_url is this driver's base_url.

OutputResourcePtr net_instaweb::RewriteDriver::CreateOutputResourceWithPath ( const StringPiece &  mapped_path,
const StringPiece &  unmapped_path,
const StringPiece &  base_url,
const StringPiece &  filter_id,
const StringPiece &  name,
OutputResourceKind  kind,
GoogleString failure_reason 

Creates an output resource where the name is provided. The intent is to be able to derive the content from the name, for example, by encoding URLs and metadata.

This method succeeds unless the filename is too long.

This name is prepended with path for writing hrefs, and the resulting url is encoded and stored at file_prefix when working with the file system. So hrefs are: /usr/local/google/home/cheesy/bin:/usr/lib/google-golang/bin:/usr/local/buildtools/java/jdk/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin/.pagespeed[.$EXPERIMENT].. .

EXPERIMENT is set only when there is an active experiment_spec.

Could be private since you should use one of the versions below but put here with the rest like it and for documentation clarity.

OutputResourcePtr net_instaweb::RewriteDriver::CreateOutputResourceWithPath ( const StringPiece &  path,
const StringPiece &  filter_id,
const StringPiece &  name,
OutputResourceKind  kind,
GoogleString failure_reason 

Version of CreateOutputResourceWithPath where the unmapped and mapped paths and the base url are all the same. FOR TESTS ONLY.

OutputResourcePtr net_instaweb::RewriteDriver::CreateOutputResourceWithUnmappedUrl ( const GoogleUrl unmapped_gurl,
const StringPiece &  filter_id,
const StringPiece &  name,
OutputResourceKind  kind,
GoogleString failure_reason 

Version of CreateOutputResourceWithPath which first takes only the unmapped path and finds the mapped path using the DomainLawyer and the base_url is this driver's base_url.

CriticalSelectorInfo* net_instaweb::RewriteDriver::critical_selector_info ( )

This should only be called by the CriticalSelectorFinder. Normal users should call CriticalSelectorFinder::IsCriticalImage.

TODO(jud): Remove when the finders reside in RewriteDriver and manage their own state.
bool net_instaweb::RewriteDriver::DebugMode ( ) const

Determines whether we are currently in Debug mode; meaning that the site owner or user has enabled filter kDebug.

bool net_instaweb::RewriteDriver::Decode ( StringPiece  leaf,
ResourceNamer resource_namer 
) const

Calls the provided ResourceNamer's Decode() function, passing the hash and signature lengths from this RewriteDriver.

const GoogleUrl& net_instaweb::RewriteDriver::decoded_base_url ( ) const

Returns the decoded version of base_gurl() in case it was encoded by a non-default UrlNamer (for the default UrlNamer this returns the same value as base_url()). Required when fetching a resource by its encoded name.

OutputResourcePtr net_instaweb::RewriteDriver::DecodeOutputResource ( const GoogleUrl url,
RewriteFilter **  filter 
) const

Attempts to decode an output resource based on the URL pattern without actually rewriting it. No permission checks are performed on the url, though it is parsed to see if it looks like the url of a generated resource (which should mean checking the hash to ensure we generated it ourselves).

TODO(jmaessen): add url hash & check thereof.
bool net_instaweb::RewriteDriver::DecodeOutputResourceName ( const GoogleUrl url,
const RewriteOptions options_to_use,
const UrlNamer url_namer,
ResourceNamer name_out,
OutputResourceKind kind_out,
RewriteFilter **  filter_out 
) const

As above, but does not actually create a resource object, and instead outputs the decoded information into the various out parameters. Returns whether decoding successful or not. Uses options_to_use rather than this->options() to determine which drivers are forbidden from applying, etc.

bool net_instaweb::RewriteDriver::DecodeUrlGivenOptions ( const GoogleUrl url,
const RewriteOptions options,
const UrlNamer url_namer,
StringVector *  decoded_urls 
) const

As above, but lets one specify the options and URL namer to use. Meant for use with the decoding_driver.

void net_instaweb::RewriteDriver::DecrementRenderBlockingAsyncEventsCount ( )

Decrements a reference count bumped up by IncrementRenderBlockingAsyncEventsCount()

void net_instaweb::RewriteDriver::DeleteRewriteContext ( RewriteContext rewrite_context)

Deletes the specified RewriteContext. If this is the last RewriteContext active on this Driver, and there is no other outstanding activity, then the RewriteDriver itself can be recycled, and WaitForCompletion can return.

We expect to this method to be called on the Rewrite thread.

DependencyTracker* net_instaweb::RewriteDriver::dependency_tracker ( ) const

dependency_tracker()->RegisterDependencyCandidate and ReportDependencyCandidate can be called from any thread.

void net_instaweb::RewriteDriver::DeregisterForPartitionKey ( const GoogleString partition_key,
RewriteContext candidate 

Must be called after all other rewrites that are currently relying on this one have had their RepeatedSuccess or RepeatedFailure methods called.

Must only be called from rewrite thread.

void net_instaweb::RewriteDriver::DetachedFetchComplete ( )

Called by RewriteContext when a detached async fetch is complete, allowing the RewriteDriver to be recycled if FetchComplete() got invoked as well.

void net_instaweb::RewriteDriver::DetachFetch ( )

Called by RewriteContext to let RewriteDriver know it will be continuing on the fetch in background, and so it should defer doing full cleanup sequences until DetachedFetchComplete() is called.

virtual void net_instaweb::RewriteDriver::DetermineFiltersBehaviorImpl ( )

Call DetermineEnabled() on each filter. Should be called after the property cache lookup has finished since some filters depend on pcache results in their DetermineEnabled implementation. If a subclass has filters that the base HtmlParse doesn't know about, it should override this function and call DetermineEnabled on each of its filters, along with calling the base DetermineEnabledFiltersImpl. For all enabled filters the CanModifyUrl() flag will be aggregated (or'ed) and can be queried on the can_modify_url function.

Reimplemented from net_instaweb::HtmlParse.

DomainRewriteFilter* net_instaweb::RewriteDriver::domain_rewriter ( )

Used by CacheExtender, CssCombineFilter, etc. for rewriting domains of sub-resources in CSS.

void net_instaweb::RewriteDriver::EnableBlockingRewrite ( RequestHeaders request_headers)

If the value of X-PSA-Blocking-Rewrite request header matches the blocking rewrite key, set fully_rewrite_on_flush flag.

void net_instaweb::RewriteDriver::ExecuteFlushIfRequested ( )

Executes an Flush() if RequestFlush() was called, e.g. from the Listener Filter (see set_event_listener below). Consider an HTML parse driven by a UrlAsyncFetcher. When the UrlAsyncFetcher temporarily runs out of bytes to read, it calls response_writer->Flush(). When that happens, we may want to consider flushing the outstanding HTML events through the system so that the browser can start fetching subresources and rendering. The event_listener (see set_event_listener below) helps determine whether enough "interesting" events have passed in the current flush window so that we should take this incoming network pause as an opportunity.

void net_instaweb::RewriteDriver::ExecuteFlushIfRequestedAsync ( Function callback)

Asynchronous version of the above. Note that you should not attempt to write out any data until the callback is invoked. (If a flush is not needed, the callback will be invoked immediately).

FallbackPropertyPage* net_instaweb::RewriteDriver::fallback_property_page ( ) const

Returns the property page which contains the cached properties associated with the current URL and fallback URL (i.e. without query params). This should be used where a property is interested in fallback values if actual values are not present.

void net_instaweb::RewriteDriver::FetchComplete ( )

Cleans up the driver and any fetch rewrite contexts, unless the fetch rewrite got detached by a call to DetachFetch(), in which case a call to DetachedFetchComplete() must also be performed.

void net_instaweb::RewriteDriver::FetchInPlaceResource ( const GoogleUrl gurl,
bool  proxy_mode,
AsyncFetch async_fetch 

Initiates an In-Place Resource Optimization (IPRO) fetch (A resource which is served under the original URL, but is still able to be rewritten).

proxy_mode indicates whether we are running as a proxy where users depend on us to send contents. When set true, we will perform HTTP fetches to get contents if not in cache and will ignore kRecentFetchNotCacheable and kRecentFetchFailed since we'll have to fetch the resource for users anyway. Origin implementations (like mod_pagespeed) should set this to false and let the serve serve the resource if it's not in cache.

If proxy_mode is false and the resource could not be found in HTTP cache, async_fetch->Done(false) will be called and async_fetch->status_code() will be CacheUrlAsyncFetcher::kNotInCacheStatus (to distinguish this from a different reason for failure, like kRecentFetchNotCacheable).

Note that if the request headers have not yet been set on the driver then they'll be taken from the fetch.

bool net_instaweb::RewriteDriver::FetchOutputResource ( const OutputResourcePtr output_resource,
RewriteFilter filter,
AsyncFetch async_fetch 

See FetchResource. There are two differences:

  1. It takes an OutputResource instead of a URL.
  2. It returns whether a fetch was queued or not. This is safe to ignore because in either case the callback will be called.
  3. If 'filter' is NULL then the request only checks cache and (if enabled) the file system.
bool net_instaweb::RewriteDriver::FetchResource ( const StringPiece &  url,
AsyncFetch fetch 

Initiates an async fetch for a rewritten resource with the specified name. If url matches the pattern of what the driver is authorized to serve, then true is returned and the caller must listen on the callback for the completion of the request.

If the driver is not authorized to serve the resource for any of the following reasons, false is returned and the callback will -not- be called - the request should be passed to another handler.

  • The URL is invalid or it does not match the general pagespeed pattern.
  • The filter id in the URL does not map to a known filter.
  • The filter for the id in the URL doesn't recognize the format of the URL.
  • The filter for the id in the URL is forbidden.

In other words there are three outcomes for this routine:

  1. the request was handled immediately and the callback called before the method returns. true is returned.
  2. the request looks good but was queued because some other resource fetch is needed to satisfy it. true is returned.
  3. the request does not look like it belongs to Instaweb. The callback will not be called, and false will be returned.

In even other words, if this routine returns 'false' then the callback will not be called. If the callback -is- called, then this should be the 'final word' on this request, whether it was called with success=true or success=false.

Note that if the request headers have not yet been set on the driver then they'll be taken from the fetch.

virtual void net_instaweb::RewriteDriver::FinishParse ( )

Override HtmlParse's FinishParse to ensure that the request-scoped cache is cleared immediately.

Note that the RewriteDriver can delete itself in this method, if it's not externally managed, and if all RewriteContexts have been completed.

Reimplemented from net_instaweb::HtmlParse.

void net_instaweb::RewriteDriver::FinishParseAsync ( Function callback)

As above, but asynchronous. Note that the RewriteDriver may already be deleted at the point the callback is invoked. The scheduler lock will not be held when the callback is run.

bool net_instaweb::RewriteDriver::FlattenCssImportsEnabled ( ) const

Return true if we must flatten css imports, either because the filter is enabled explicitly or because it is enabled by PrioritizeCriticalCss.

virtual void net_instaweb::RewriteDriver::Flush ( )

Overrides HtmlParse::Flush so that it can happen in two phases:

  1. Pre-render chain runs, resulting in async rewrite activity
  2. async rewrite activity ends, calling callback, and post-render filters run. This API is used for unit-tests & Apache (which lacks a useful event model) and results in blocking behavior.

FlushAsync is prefered for event-driven servers.

Reimplemented from net_instaweb::HtmlParse.

void net_instaweb::RewriteDriver::FlushAsync ( Function done)

Initiates an asynchronous Flush. done->Run() will be called when the flush is complete. Further calls to ParseText should be deferred until the callback is called. Scheduler mutex is not held while done is called.

static PropertyCache::CohortVector net_instaweb::RewriteDriver::GetCohortList ( const PropertyCache pcache,
const RewriteOptions options,
const ServerContext server_context 

Returns the list of cohorts that should be read in based on our options.

InlineAttributeSlotPtr net_instaweb::RewriteDriver::GetInlineAttributeSlot ( const ResourcePtr resource,
HtmlElement element,
HtmlElement::Attribute attribute 

Creates and registers an inline attribute resource slot for rewriting. If this is the first time called for this position, a new slot will be returned. On subsequent calls, the original slot will be returned so that rewrites are propagated between filters.

InlineResourceSlotPtr net_instaweb::RewriteDriver::GetInlineSlot ( const ResourcePtr resource,
HtmlCharactersNode char_node 

Creates and registers an inline resource slot for rewriting. If this is the first time called for this position, a new slot will be returned. On subsequent calls, the original slot will be returned so that rewrites are propagated between filters.

HtmlResourceSlotPtr net_instaweb::RewriteDriver::GetSlot ( const ResourcePtr resource,
HtmlElement elt,
HtmlElement::Attribute attr 

Creates and registers a HtmlElement slot for rewriting. If this is the first time called for this position, a new slot will be returned. On subsequent calls, the original slot will be returned so that rewrites are propagated between filters.

SrcSetSlotCollectionPtr net_instaweb::RewriteDriver::GetSrcSetSlotCollection ( CommonFilter filter,
HtmlElement element,
HtmlElement::Attribute attr 

Create and and registers a source set slot collection for rewriting all the images in the srcset attribute of an . Also creates the neccessary resources using the provided filter's policy.

If this is the first time called for this element + attr, a new collection will be returned. On subsequent calls, the original collection will be returned so that rewrites are propagated between filters. All filters using this call are expected to have the same values for AllowUnauthorizedDomain() and IntendedForInlining().

void net_instaweb::RewriteDriver::IncrementAsyncEventsCount ( )

Increment reference count for misc. async ops that need the RewriteDriver kept alive.

void net_instaweb::RewriteDriver::IncrementRenderBlockingAsyncEventsCount ( )

Increment reference count for misc async ops that should be waited for before doing rendering for current flush window.

void net_instaweb::RewriteDriver::InfoAt ( const RewriteContext context,
const char *  msg,

Report error message with description of context's location (such as filenames and line numbers). context may be NULL, in which case the current parse position will be used.

bool net_instaweb::RewriteDriver::InitiateRewrite ( RewriteContext rewrite_context)

Method to start a resource rewrite. This is called by a filter during parsing, although the Rewrite might continue after deadlines expire and the rewritten HTML must be flushed. Returns InitiateRewrite returns false if the system is not healthy enough to support resource rewrites.

void net_instaweb::RewriteDriver::InsertDebugComment ( StringPiece  unescaped_message,
HtmlNode node 

Log the given debug message(s) as HTML comments after the given element, if not NULL, it has not been flushed, and if debug is enabled. The form that takes a repeated field is intended for use by CachedResult, e.g: InsertDebugComment(cached_result.debug_message(), element); Messages are HTML-escaped before being written out to the DOM.

bool net_instaweb::RewriteDriver::is_critical_images_beacon_enabled ( )

The JS to detect above-the-fold images should only be enabled if one of the filters that uses critical image information is enabled, the property cache is enabled (since the critical image information is stored in the property cache), and it is not explicitly disabled through options.

bool net_instaweb::RewriteDriver::IsResourceUrlClaimed ( const GoogleUrl url) const

Returns true if some ResourceUrlClaimant has staked a claim on given URL. If this returns true, CreateInputResource will fail, but it's probably not worth logging any debug filter hints about that.

AbstractLogRecord* net_instaweb::RewriteDriver::log_record ( )

log_record() always returns a pointer to a valid AbstractLogRecord, owned by the rewrite_driver's request context.

bool net_instaweb::RewriteDriver::LookupMetadataForOutputResource ( StringPiece  url,
GoogleString error_out,
RewriteContext::CacheLookupResultCallback callback 

Attempts to lookup the metadata cache info that would be used for the output resource at url with the RewriteOptions set on this driver.

If there is a problem with the URL, returns false, and *error_out will contain an error message.

If it can determine the metadata cache key successfully, returns true, and eventually callback will be invoked with the metadata cache key and the decoding results.

After calling this method, the driver should not be used for anything else.

bool net_instaweb::RewriteDriver::MatchesBaseUrl ( const GoogleUrl input_url) const

Checks to see if the input_url has the same origin as and the base url, to make sure we're not fetching from another server. Does not consult the domain lawyer, and is not affected by AddDomain(). Precondition: input_url.IsWebValid()

bool net_instaweb::RewriteDriver::MayCacheExtendCss ( ) const

Returns true if we may cache extend Css, Images, PDFs, or Scripts respectively.

bool net_instaweb::RewriteDriver::MayRewriteUrl ( const GoogleUrl domain_url,
const GoogleUrl input_url,
InlineAuthorizationPolicy  inline_authorization_policy,
IntendedFor  intended_for,
bool *  is_authorized_domain 
) const

Checks to see if we can write the input_url resource in the domain_url taking into account domain authorization, wildcard allow/disallow from RewriteOptions, and the intended use of the url's resource. After the function is executed, is_authorized_domain will indicate whether input_url was found to belong to an authorized domain or not.

XhtmlStatus net_instaweb::RewriteDriver::MimeTypeXhtmlStatus ( )

Determines whether the document's Content-Type has a mimetype indicating that browsers should parse it as XHTML.

ResponseHeaders* net_instaweb::RewriteDriver::mutable_response_headers ( )

Return a mutable pointer to the response headers that filters can update before the first flush. Returns NULL after Flush has occurred.

int net_instaweb::RewriteDriver::num_inline_preview_images ( ) const

We expect to this method to be called on the HTML parser thread. Returns the number of images whose low quality images are inlined in the html page.

PropertyPage* net_instaweb::RewriteDriver::origin_property_page ( ) const

Returns property page which contains cached properties associated with the current origin (host/port/protocol). May be NULL.

void net_instaweb::RewriteDriver::PopulateResourceNamer ( const StringPiece &  filter_id,
const StringPiece &  name,
ResourceNamer full_name 

Fills in the resource namer based on the give filter_id, name and options stored in the driver.

void net_instaweb::RewriteDriver::PrependRewriteFilter ( RewriteFilter filter)

Like AppendRewriteFilter, but adds the filter to the beginning of the pre-render chain.

PropertyPage* net_instaweb::RewriteDriver::property_page ( ) const

Returns the property page which contains the cached properties associated with the current URL.

void net_instaweb::RewriteDriver::PropertyCacheSetupDone ( )

Should be called once everything in the property cache has been read, and the pages set on the object.

RewriteContext* net_instaweb::RewriteDriver::RegisterForPartitionKey ( const GoogleString partition_key,
RewriteContext candidate 

Tries to register the given rewrite context as working on its partition key. If this context is the first one to try to handle it, returns NULL. Otherwise returns the previous such context.

Must only be called from rewrite thread.

void net_instaweb::RewriteDriver::ReportSlowRewrites ( int  num)

Provides a mechanism for a RewriteContext to notify a RewriteDriver that a certain number of rewrites have been discovered to need to take the slow path.

void net_instaweb::RewriteDriver::RequestFlush ( )

Indicates that a Flush through the HTML parser chain should happen soon, e.g. once the network pauses its incoming byte stream.

CssResolutionStatus net_instaweb::RewriteDriver::ResolveCssUrls ( const GoogleUrl input_css_base,
const StringPiece &  output_css_base,
const StringPiece &  contents,
Writer writer,
MessageHandler handler 

Rewrites CSS content to absolutify any relative embedded URLs, streaming the results to the writer. Returns 'false' if the writer returns false or if the content was not rewritten because the domains of the gurl and resolved_base match.

input_css_base contains the path where the CSS text came from. output_css_base contains the path where the CSS will be written.

const ResponseHeaders* net_instaweb::RewriteDriver::response_headers ( )

Returns a const version of the ResponseHeaders*, indepdendent of whether Flush has occurred. Note that ResponseHeaders* may still be NULL if no one has called set_response_headers_ptr.

TODO(jmarantz): Change API to require response_headers in StartParse so we can guarantee this is non-null.
void net_instaweb::RewriteDriver::RewriteComplete ( RewriteContext rewrite_context,
bool  permit_render 

Provides a mechanism for a RewriteContext to notify a RewriteDriver that it is complete, to allow the RewriteDriver to delete itself or return it back to a free pool in the ServerContext.

This will also call back into RewriteContext::Propagate, letting it know whether the context is still attached to the HTML DOM (and hence safe to render), and to do other bookkeeping.

If 'permit_render' is false, no rendering will be asked for even if the context is still attached.

void net_instaweb::RewriteDriver::RunTasksOnRequestThread ( )

Make the rewrite_worker tasks run on the request thread. This must be called immediately after initializing the driver, before it starts processing the request.

void net_instaweb::RewriteDriver::set_critical_images_info ( CriticalImagesInfo critical_images_info)

Inserts the critical images present on the requested html page. It takes ownership of critical_images_info. This should only be called by the CriticalImagesFinder, normal users should just be using the automatic management of critical_images_info that CriticalImagesFinder provides.

void net_instaweb::RewriteDriver::set_critical_selector_info ( CriticalSelectorInfo info)

This should only be called by the CriticalSelectorFinder.

TODO(jud): Remove when the finders reside in RewriteDriver and manage their own state.
void net_instaweb::RewriteDriver::set_externally_managed ( bool  x)

Indicate that this RewriteDriver will be explicitly deleted, and thus should not be auto-deleted at the end of the parse. This is primarily for tests.

TODO(jmarantz): Consider phasing this out to make tests behave more like servers.
void net_instaweb::RewriteDriver::set_fast_blocking_rewrite ( bool  x)

This is relevant only when fully_rewrite_on_flush is true. When this is set to true, Flush of HTML will not wait for async events while it does wait when it is set to false.

void net_instaweb::RewriteDriver::set_fully_rewrite_on_flush ( bool  x)

If this is set to true, during a Flush of HTML the system will wait for results of all rewrites rather than just waiting for cache lookups and a small deadline. Note, however, that in very rare circumstances some rewrites may still be dropped due to excessive load.

Note: reset every time the driver is recycled.

void net_instaweb::RewriteDriver::set_max_page_processing_delay_ms ( int  x)

Sets a maximum amount of time to process a page across all flush windows; i.e., the entire lifecycle of this driver during a given pageload. A negative value indicates no limit. Setting fully_rewrite_on_flush() overrides this.

void net_instaweb::RewriteDriver::set_num_initiated_rewrites ( int64  x)

Sets the num_initiated_rewrites_. This should only be called from test code.

void net_instaweb::RewriteDriver::set_options_for_pool ( RewriteDriverPool pool,
RewriteOptions options 

Takes ownership of 'options'. pool denotes the pool of rewrite drivers that use these options. May be NULL if using custom options.

void net_instaweb::RewriteDriver::set_refs_before_base ( )

Sets whether or not there were references to urls before the base tag (if there is a base tag). This variable has document-level scope. It is reset at the beginning of every document by ScanFilter.

void net_instaweb::RewriteDriver::set_response_headers_ptr ( ResponseHeaders headers)

Set the pointer to the response headers that filters can update before the first flush. RewriteDriver does NOT take ownership of this memory.

void net_instaweb::RewriteDriver::SetIsAmpDocument ( bool  is_amp)

Declares whether the current document is AMP or not. Prior to calling this, all HTML events are buffered, to avoid waking up filters that inject scripts.

bool net_instaweb::RewriteDriver::SetOrClearPageSpeedOptionCookies ( const GoogleUrl gurl,
ResponseHeaders response_headers 

Utility function to set/clear cookies for PageSpeed options. gurl is the URL of the request from which the host is extracted for a cookie attribute.

TODO(matterbury): Get the URL from 'this' which we can't do now because it isn't set until we've decided that the content of requested URL is HTML. Returns true if any Set-Cookie headers are added, in which case ComputeCaching has been called on response_headers.
void net_instaweb::RewriteDriver::SetRequestHeaders ( const RequestHeaders headers)

Reinitializes request_headers_ (a scoped ptr) with a copy of the original request headers. Note that the fetches associated with the driver could be using a modified version of the original request headers. There MUST be exactly 1 call to this method after a rewrite driver object has been constructed or recycled, before the RewriteDriver is used for request processing.

This method also sets up the user-agent and device properties.

void net_instaweb::RewriteDriver::SetServerContext ( ServerContext server_context)

Sets a server context enabling the rewriting of resources. This will replace any previous server context.

void net_instaweb::RewriteDriver::SetSessionFetcher ( UrlAsyncFetcher f)

Set a fetcher that will be used by RewriteDriver for current request only (that is, until Clear()). RewriteDriver will take ownership of this fetcher, and will keep it around until Clear(), even if further calls to this method are made.

void net_instaweb::RewriteDriver::SetWriter ( Writer writer)

Controls how HTML output is written. Be sure to call this last, after all other filters have been established.

TODO(jmarantz): fix this in the implementation so that the caller can install filters in any order and the writer will always be last.
bool net_instaweb::RewriteDriver::ShouldAbsolutifyUrl ( const GoogleUrl input_base,
const GoogleUrl output_base,
bool *  proxy_mode 
) const

Determines if an URL relative to the given input_base needs to be absolutified given that it will end up under output_base:

  • If we are proxying and input_base isn't proxy encoded, then yes.
  • If we aren't proxying and input_base != output_base, then yes.
  • If we aren't proxying and the domain lawyer will shard or rewrite input_base, then yes. If not NULL also set *proxy_mode to whether proxy mode is active or not.
virtual bool net_instaweb::RewriteDriver::StartParseId ( const StringPiece &  url,
const StringPiece &  id,
const ContentType content_type 

Override HtmlParse's StartParseId to propagate any required options. Note that if this (or other variants) returns true you should use FinishParse(), otherwise Cleanup().

Reimplemented from net_instaweb::HtmlParse.

void net_instaweb::RewriteDriver::SwitchToQueuedWorkerPool ( )

Switches the driver back to running rewrite_worker tasks using the QueuedWorkerPool. This can be called when we are retiring a server-request on behalf of the client (e.g. after a deadline was exceeded), but want background optimization to continue. It can no longer continue on the request thread.

RequestTrace* net_instaweb::RewriteDriver::trace_context ( )

Convenience method to return the trace context from the request_context() if both are configured and NULL otherwise.

void net_instaweb::RewriteDriver::TracePrintf ( const char *  fmt,

Convenience methods to issue a trace annotation if tracing is enabled. If tracing is disabled, these methods are no-ops.

void net_instaweb::RewriteDriver::UpdatePropertyValueInDomCohort ( AbstractPropertyPage page,
StringPiece  property_name,
StringPiece  property_value 

Update the PropertyValue named 'property_name' in dom cohort with the value 'property_value'. It is the responsibility of the client to ensure that property cache and dom cohort are enabled when this function is called. It is a programming error to call this function when property cache or dom cohort is not available, more so since the value payload has to be serialised before calling this function. Hence this function will DFATAL if property cache or dom cohort is not available.

void net_instaweb::RewriteDriver::WaitForCompletion ( )

Wait for outstanding Rewrite to complete. Once the rewrites are complete they can be rendered.

void net_instaweb::RewriteDriver::WaitForShutDown ( )

Wait for outstanding rewrite to complete, including any background work that may be ongoing even after results were reported.

Note: while this guarantees that the result of the computation is known, the thread that performed it may still be running for a little bit and accessing the driver.

bool net_instaweb::RewriteDriver::Write ( const ResourceVector &  inputs,
const StringPiece &  contents,
const ContentType type,
StringPiece  charset,
OutputResource output 

Writes the specified contents into the output resource, and marks it as optimized. 'inputs' described the input resources that were used to construct the output, and is used to determine whether the result can be safely cache extended and be marked publicly cacheable. 'content_type' and 'charset' specify the mimetype and encoding of the contents, and will help form the Content-Type header. 'charset' may be empty when not specified.

Note that this does not escape charset.

Callers should take care that dangerous types like 'text/html' do not sneak into content_type.

Member Data Documentation

const char net_instaweb::RewriteDriver::kDependenciesCohort[]

Cohort for dependency information. This is written at different time than kDomCohort, and might not be in use for some requests, depending on settings.

const char net_instaweb::RewriteDriver::kDomCohort[]

This string identifies, for the PropertyCache, a group of properties that are computed from the DOM, and thus can, if desired, be rewritten on every HTML request.

const char net_instaweb::RewriteDriver::kLastRequestTimestamp[]

Property Names in DomCohort. Tracks the timestamp when we last received a request for this url.

The documentation for this class was generated from the following file: