on the HTTP requests for inventory. We'll benchmark with that
and see how it goes. Document some of the history of the
background fetcher for future devs. Suggest some future
projects to make things faster. Pointers on using LLSD with
the llcorehttp library in the readme. And restructured
the LLSD onCompleted() processing phases using do{}while(false)
which produced a code flow that is fairly attractive.
accept headers in mesh and textures. For texture metrics
reporting, use the AP_INVENTORY policy class which is
non-pipelined and pointing (usually) in the right direction.
Use a do-while(false) structure to manage common exit path
code in onCompleted() methods. Identical to a 'goto' but
might amuse the pedantic. Tuning on background fetch to
have it cycle faster. This is experimental. I suspect
with HTTP balancing in llcorehttp, we can do away with the
timers here.
First, introduced some LLSD-based interfaces to the llcorehttp code
using utils classes (in llcorehttputil). I've kept LLSD out of
the llcorehttp library up to now and will continue to do that.
Functions provide a requestPost based on LLSD body and conversion
utils for HttpResponse-to-LLSD and HttpResponse-to-string
conversions. Inventory fetch operations now do more thorough
error checking including 200-with-error status checking. Still
do retry forever on folders though I don't like that.
connections. Convert background and foreground fetches, both
items and folders/inventory and library, to use new HTTP.
Non-fetch inventory operations continue to use LLHTTPClient
(at least for now). Error handling and retry on fetches wasn't
100% previously and that's still the case. I'll rip through
this again to clean that up. Cleaned up logging in much of
the inventory code with consistent labels on logging events
and correct macros (removed deprecation warnings).
This started as an attempt to get libcurl to do pipelining
on POSTs and PUTs. Discovered that this is going to be
very difficult to support in general in libcurl. May
look at that again in the future.
* Additional error checking in http handlers.
* Uniform log spam for http errors.
* Switch to using constants for http heads and status codes.
* Fixed bugs in incorrectly checking if parsing LLSD xml resulted in an error.
* Reduced spam regarding LLSD parsing errors in the default completedRaw http handler. It should not longer be necessary to short-circuit completedRaw to avoid spam.
* Ported over a few bug fixes from the server code.
* Switch mode http status codes to use S32 instead of U32.
* Ported LLSD::asStringRef from server code; avoids copying strings all over the place.
* Ported server change to LLSD::asBinary; this always returns a reference now instead of copying the entire binary blob.
* Ported server pretty notation format (and pretty binary format) to llsd serialization.
* The new LLCurl::Responder API no longer has two error handlers to choose from. Overriding the following methods have been deprecated:
** error - use httpFailure
** errorWithContent - use httpFailure
** result - use httpSuccess
** completed - use httpCompleted
** completedHeader - no longer necessary; call getResponseHeaders() from a completion method to obtain these headers.
* In order to 'catch' a completed http request, override one of these methods:
** httpSuccess - Called for any 2xx status code.
** httpFailure - Called for any non-2xx status code.
** httpComplete - Called for all status codes. Default implementation is to call either httpSuccess or httpFailure.
* It is recommended to keep these methods protected/private in order to avoid triggering of these methods without using a 'push' method (see below).
* Uniform error handling should followed whenever possible by calling a variant of this during httpFailure:
** llwarns << dumpResponse() << llendl;
* Be sure to include LOG_CLASS(your_class_name) in your class in order for the log entry to give more context.
* In order to 'push' a result into the responder, you should no longer call error, errorWithContent, result, or completed.
* Nor should you directly call httpSuccess/Failure/Completed (unless passing a message up to a parent class).
* Instead, you can set the internal content of a responder and trigger a corresponding method using the following methods:
** successResult - Sets results and calls httpSuccess
** failureResult - Sets results and calls httpFailure
** completedResult - Sets results and calls httpCompleted
* To obtain information about a the response from a reponder method, use the following getters:
** getStatus - HTTP status code
** getReason - Reason string
** getContent - Content (Parsed body LLSD)
** getResponseHeaders - Response Headers (LLSD map)
** getHTTPMethod - HTTP method of the request
** getURL - URL of the request
* It is still possible to override completeRaw if you want to manipulate data directly out of LLPumpIO.
* See indra/llmessage/llcurl.h for more information.
background fetch always proceeds when fetching a specific folder
this allows us to flag folders that are received asynchronously as complete by
performing a background fetch when opening them
The problem was caused by an outdated message name stored in LLInventoryObserver::mMessageName and not updated properly in LLInventoryModel::notifyObservers().
The message name used in LLInventoryAddedObserver::changed() was the name of the message most recently passed by LLInventoryModel::notifyObservers(), instead of the name of the latest actually received message. Using the most recent message name in this case fixed the problem.
Going through and cleaning up any todos that have "Seraph" attached to it. In this case, did some header file cleanup of llinventorymodelbackgroundfetch and moved some functions protected/const/etc.
* Moved initialization (with start fetch) of the LLFilteredWearableListManager into LLPanelOutfitEdit::onOpen
* LLFilteredWearableListManager::changed now ignores non-related inventory changes (CALLING_CARD, GESTURE, SORT)
Tried to fix an issue with empty Gesture list on startup with clean inventory cache.
Reason: logic is based on count of "inventory changed" events. In case of there was too many requests requested items can be removed from queue by mistake.
* Increased a number of the "change()" method calls to wait fetched items.
Unfortunately this only works if My Inventory category does not have too many children. An does not work if it has 2000+ items
Logic to remove item from the incompleted list should be based on timer and number of attempts.
Also add some debug information about fetching inventory category and its content to log.
Reviewed by Brad Payne at https://codereview.productengine.com/secondlife/r/456/
--HG--
branch : product-engine
andreykproductengine