This was a race condition. In the repro case, LLViewerMediaImpl::seek() was getting called before the media source got a chance to be loaded (since the load is now delayed at least 1 frame due to the priority calculations), so the seek command was being dropped.
The fix is to save the seek time in mPreviousMediaTime in this case, and rearrange the code in LLViewerMediaImpl::loadURI() so that it gets used whenever it's non-zero. I've also added similar code in LLViewerMediaImpl::pause() and LLViewerMediaImpl::start() to set mPreviousMediaState. (If there are any cases where content wants parcel media to start out paused, they would have had a similar issue.)
Reviewed by Callum and Richard at http://codereview.lindenlab.com/2478002 .
monroe already did most of this, just moved dialog for confirming opening external browser to common code that calls out to external browser
so no way to skip notification unless the user opts out
Moved the processing of clicks on _external and _blank targeted links from LLMediaCtrl to LLViewerMediaImpl.
Removed LLMediaCtrl::setOpenInExternalBrowser() since that functionality is available through the use of the _external target attribute in web content.
Removed LLMediaCtrl:: setOpenInInternalBrowser() since it was unimplemented and not used.
Made the webkit media plugin set llqtwebkit's window open behavior to WOB_SIMULATE_BLANK_HREF_CLICK. This is #ifdefed out on Linux until we get a new Linux build of llqtwebkit.
Use LLMutexLock (stack-based locker/unlocker) for the straightforward cases instead of explicit lock()/unlock().
There are still a couple of cases (one overlapping lock lifetime and two loops that unlock the mutex to call another function inside the loop) where I'm leaving explicit lock/unlock calls.
Rename LLPluginProcessParent::sPollThread to sReadThread, for consistency.
Made the LLPluginProcessParent destructor hold mIncomingQueueMutex while removing the instance from the global list -- this should prevent a possible race condition in LLPluginProcessParent::poll().
Removed a redundant check when calling LLPluginProcessParent::setUseReadThread().
LLPluginProcessParent can now optionally use a separate thread for reading messages from plugin sockets. If this is enabled, it will spawn a single thread and use apr_pollset_poll to wake up the thread when incoming data arrives instead of polling all the descriptors round-robin every frame. This should be somewhat more efficient, and should also allow blocking requests from plugins to be serviced much more quickly (once we start using them). This is currently disabled by default, until it's had a bit more focused testing on multiple platforms.
Hooked up the switch to use the message read thread to the PluginUseReadThread debug setting and an item in the Advanced menu in the viewer, and to a checkbox in the UI in llmediaplugintest.
Updated some debug logging in the plugin system to have appropriate tags and not log dire-looking warnings during normal operation.
LLPluginProcessParent now once again explicitly kills plugin processes (instead of just closing their sockets and waiting for them to exit). The problem we were attempting to solve by not doing the kill (letting the webkit plugin write its cookie file on exit) has been solved another way.
LLPluginProcessParent::sendMessage() now attempts to write the outgoing message to the socket immediately instead of waiting for the next frame. This should reduce the latency of sending plugin messages.
Added a separate fast timer for LLViewerMedia::updateMedia().
Added "DefaultMimeType" string to strings.xml to make "none/none" mime type localizable.
Replaced all occurrences of "none/none" string with call to static method.
Reviewed by Vadim Savchuk - https://codereview.productengine.com/secondlife/r/265/
--HG--
branch : product-engine
The code was using the "authority" part of the URL returned by LLURL when constructing the cookie's domain, which incorrectly included the port number. This change makes sure any port number and username/password that might be in the authority are stripped from the string before handing it to the cookie code.
Reviewed by Gino at http://codereview.lindenlab.com/1297006
LLViewerMediaImpl::shouldShowBasedOnClass() was always returning true for focused media. This was incorrect, since it made click-focusing on media override the class-based disable preferences.
Reviewed by Callum at http://codereview.lindenlab.com/1178023
Extract openid_url and openid_token tokens from the login response in process_login_success_response() and send them to LLViewerMedia if they're present.
Added LLViewerMedia::openIDSetup() to receive openid_url and openid_token, and added code to LLViewerMedia to do a POST with LLHTTPClient, retrieve the resulting cookie, and push it into the central cookie store. Also made sure the OpenID cookie gets re-added when the cookie store is cleared.
Added LLPluginCookieStore::setCookiesFromHost() to properly add a cookie that may not have a domain set. Made LLPluginCookieStore::Cookie::parse() add missing domain and path fields to cookies as necessary.
Fixed an issue where carriage returns in the string passed to LLPluginCookieStore::setCookies() or LLPluginCookieStore::setCookiesFromHost() would cause a parse failure.
Reviewed by gino and callum at http://codereview.lindenlab.com/1254001
LLViewerMedia::clearAllCookies was sending "clear cookies" messages to all plugins and deleting cookie files, but it wasn't clearing the cookies in the central store. It now does all of the above.
Made LLViewerMedia::clearAllCookies() delete the new per-user cookie file (userdir/plugin_cookies.txt) as well as the old one.
Made LLViewerMedia::loadCookieFile() send the clear_cookies message to all loaded plugins.
Added LLPluginCookieStore, which manages the central list of cookies.
New Mac and Windows versions of llqtwebkit, built from the tip of the cookie-api branch on http://bitbucket.org/lindenlab/llqtwebkit/ (currently revision f35a5eab8c2f).
Added "set_cookies" and "cookie_set" messages to the media_browser message class in the plugin API, and made the webkit plugin use them appropriately.
Added methods to LLViewerMedia to read/write the cookie file and add/remove individual cookies.
Added hooks to read/write the cookie file (plugin_cookies.txt) in the same places as the location history (idle_startup() in llstartup.cpp and LLAppViewer::cleanup(), respectively).
Reviewed by Richard at http://codereview.lindenlab.com/1006003
Added static member functions to LLViewerMedia: clearAllCookies(), clearAllCaches(), setCookiesEnabled(), and setProxyConfig(). These iterate through all loaded instances and send appropriate messages.
Also added a hack to clearAllCookies() to delete cookie files for all accounts, which should serve until we centralize cookie storage.
Made LLViewerMedia set the proxy config properly when creating new instances. Callum is adding code to do something similar with the cookies enabled flag in a separate commit.
Removed get_web_media from llfloaterpreference.cpp and reworked everything that used it to use the new LLViewerMedia functions instead.
Reviewed by Callum at http://codereview.lindenlab.com/838003.
Made all existing plugins send their texture_params message from this init message instead of the LLPLUGIN_MESSAGE_CLASS_BASE "init" message. (This ensures that they won't start to receive 'size_change' messages until after the init has happened.)
Added "set_user_data_path" and "set_language_code" messages to LLPluginClassMedia.
Made webkit plugin deal with the new messages, when they're sent before it receives the media "init".
Removed the user_data_path and language_code arguments from the init function calls throughout the hierarchy.
Made LLViewerMediaImpl queue up the language code and user data path messages before initializing the media.
Reviewed by Callum at http://codereview.lindenlab.com/687006 .
(for B5) Fix for EXT-5314 "Inworld Browser blanks out at credit card entry"
Note: also includes an update to install.xml that points to a new version of LLQtWebKit that is required for these fixes
We don't actually have a repro for this bug, just a number of crash reports.
This potential fix makes an iterator loop safer in the face of the iterator changing during the loop. See the jira for further details.
Reviewed by Richard at http://codereview.lindenlab.com/331008
Monroe Linden