SDL to 1.2.15, c-ares to latest 1.10.0 build, Boost to 1.55.0
with coroutine updates/fixes, curl to 7.34.0, libpng to 1.6.8,
openssl to 1.0.1e, zlib to latest 1.2.8 build, llqtwebkit
built from 4.7.1 sources refactored and tested in 3p-llqtwebkit2
repository.
Windows is functional with a good number of warning messages
at runtime from libpng and KDU. MoaP/slplugin functioning.
If a marker file indicates that an installer has been downloaded and verified
and is ready to install, but that installer file does not exist, log it and
revert to checking for update.
Also add logging to every code path that can remove the installer explicitly,
in case the missing installer is actually being killed off by some weird
sequence of state-machine transitions.
* 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.
In autobuild.xml, specify today's build of the Boost package that includes the
Boost.Context library, and whose boost::dcoroutines library uses Boost.Context
exclusively instead of its previous context-switching underpinnings (source of
the ucontext.h dependency).
Add BOOST_CONTEXT_LIBRARY to Boost.cmake and Copy3rdPartyLibs.cmake. Link it
with the viewer and with the lllogin.cpp test executable.
Track new Boost package convention that our (early, unofficial) Boost.Coroutine
library is now accessed as boost/dcoroutine/etc.h and boost::dcoroutines::etc.
Remove #include <boost/coroutine/coroutine.hpp> from
llviewerprecompiledheaders.h and lllogin.cpp: old rule that Boost.Coroutine
header must be #included before anything else that might use ucontext.h is
gone now that we no longer depend on ucontext.h. In fact remove
-D_XOPEN_SOURCE in 00-Common.cmake because that was inserted specifically to
work around a known problem with the ucontext.h facilities.
Capture actual error output from mkdir and mv; display it to user.
Introduce mysudo function used by sudo_mv function for graphical sudo command.
Since update_install actually only displays a single status message, just use
zenity --info instead of a zenity progress box: need not update its message.
Borrow semantics for clear_message and status functions from xmenity script.
Introduce errorbox function so we only have to make zenity/xmessage test once.
Move cleanup, onexit to top so we can use for clear_message.
Test for existence of target name using -e rather than -f. (-d would work too,
but in this case we must respond to any name collision, whether file or
directory.)
Instead of terminating on failure, make sudo_mv return rc of the [sudo] mv
command to its caller. If the attempt to move new install to actual viewer
directory fails, restore previous viewer before failing.
When redirecting the script's stderr to updater.log, first save existing
stderr to another file descriptor, and restore it when we launch viewer.
Otherwise updater.log ends up collecting the viewer's duplicate stderr log
output!
The construct 'exec ... | program' doesn't work. In fact it causes any other
redirections on that command to fail too. Remove it -- real fix pending.
Now that the viewer's own background updater logic is responsible for
downloading a new installer, the only functionality we still use in
linux-updater that couldn't be expressed more simply in bash is the UI. But
since most Linux distros capable of running SL at all have zenity, and all
will have xmessage, we can handle even the UI part. Add xmenity wrapper script
so update_install doesn't have to care which is present, and make the bash
script that used to launch linux-updater do the real work.
If you run "/Applications/Second Life Beta Materials.app", and the version
manager directs you to update to a .dmg containing a release candidate whose
app bundle name is "Second Life Viewer.app", update_install.py used to copy
the contents of "Second Life Viewer.app" into "/Applications/Second Life Beta
Materials.app". Changed it so that if the application bundle name differs, we
leave the running app bundle alone, installing instead to the app bundle name
from the .dmg.
Remove mac-updater subtree from viewer source, along with the
update_install bash script that invoked it. Remove all mention of mac-updater
in CMakeLists.txt files and in viewer_manifest.py.
Change Mac update_install bash script references in viewer_manifest.py and in
llupdaterservice.cpp (which invokes it) to new Python update_install.py.
Add update_install.py, messageframe.py (which puts up some Tkinter UI) and
janitor.py (cloned from vita, it's exactly what we need here).
In a number of different places, for different reasons, the viewer wants to
load a UI-related file that might be overridden by a non-default skin; and
within that skin, might further be overridden by a non-default language.
Apparently, for each of those use cases, every individual developer approached
it as an entirely new problem, solving it idiosyncratically for that one case.
Not only is this a maintenance problem, but it rubs one's nose in the fact
that most such solutions consider only a subset of the relevant skin
directories.
Richard and I evolved an API intended to address all such cases: a central
LLDir method returning a list of relevant pathnames, from most general to most
localized, filtered to present only existing files; plus a couple of
convenience methods to specifically obtain the most general and most localized
available file.
There were several load-skinned-file methods (LLFloater::buildFromFile(),
LLPanel::buildFromFile() and LLUICtrlFactory::createFromFile() -- apparently
cloned-and-modified from each other) that contained funky bolted-on logic to
output the loaded data to an optional passed LLXMLNodePtr param. The trouble
is that passing that param forced each of these methods to subvert its normal
search: specifically for that case, it needed to find the baseline XML file
instead of the localized one. Richard agreed that for the intended usage
(reformatting XML files) we should use XML schema instead, and that the hacky
functionality should be removed. Remove it. Also remove
LLUICtrlFactory::getLocalizedXMLNode(), only used for those three special cases.
Some callers explicitly passed the optional LLXMLNodePtr param as NULL. Remove
that.
Remove LLFloaterUIPreview::displayFloater(save) param, which relied on the
optional output LLXMLNodePtr param. Make onClickSaveFloater() and
onClickSaveAll() emit popupAndPrintWarning() about discontinued functionality.
Recast LLFloater::buildFromFile(), LLPanel::buildFromFile(),
LLUICtrlFactory::createFromFile(), LLNotifications::loadTemplates(),
LLUI::locateSkin(), LLFontRegistry::parseFontInfo(),
LLUIColorTable::loadFromSettings(), LLUICtrlFactory::loadWidgetTemplate(),
LLUICtrlFactory::getLayeredXMLNode(), LLUIImageList::initFromFile(),
LLAppViewer::launchUpdater() and LLMediaCtrl::navigateToLocalPage() to use
findSkinnedFilenames(). (Is LLAppViewer::launchUpdater() ever called any more?
Apparently so -- though the linux-updater.bin logic to process the relevant
command-line switch has been disabled. Shrug.) (Is
LLMediaCtrl::navigateToLocalPage() ever used?? If so, why?)
Remove LLUI::setupPaths(), getXUIPaths(), getSkinPath() and
getLocalizedSkinPath(). Remove the skins/paths.xml file read by setupPaths().
The only configuration it contained was the pair of partial paths "xui/en" and
"xui/[LANGUAGE]" -- hardly likely to change. getSkinPath() specifically
returned the first of these, while getLocalizedSkinPath() specifically
returned the second. This knowledge is now embedded in findSkinnedFilenames().
Also remove paths.xml from viewer_manifest.py.
Remove injected xui_paths from LLFontGL::initClass() and
LLFontRegistry::LLFontRegistry(). These are no longer needed since
LLFontRegistry can now directly consult LLDir for its path search. Stop
passing LLUI::getXUIPaths() to LLFontGL::initClass() in LLViewerWindow's
constructor and initFonts() method.
Add LLDir::append() and add() methods for the simple task of combining two
path components separated by getDirDelimiter() -- but only if they're both
non-empty. Amazing how often that logic is replicated. Replace some existing
concatenations with add() or append().
New LLDir::findSkinnedFilenames() method must know current language. Allow
injecting current language by adding an LLDir::setSkinFolder(language) param,
and pass it where LLAppViewer::init() and initConfiguration() currently call
setSkinFolder(). Also add LLDir::getSkinFolder() and getLanguage() methods.
Change LLFLoaterUIPreview's LLLocalizationResetForcer helper to "forcibly
reset language" using LLDir::setSkinFolder() instead of LLUI::setupPaths().
Update LLDir stubs in lldir_stub.cpp and llupdaterservice_test.cpp.
Add LLDir::getUserDefaultSkinDir() to obtain often-overlooked possible skin
directory -- like getUserSkinDir() but with "default" in place of the current
skin name as the last path component. (However, we hope findSkinnedFilenames()
obviates most explicit use of such individual skin directory pathnames.)
Add LLDir unit tests for new findSkinnedFilenames() and add() methods -- the
latter exercises append() as well.
Tweak indra/integration_tests/llui_libtest/llui_libtest.cpp for all the above.
Notably, comment out its export_test_floaters() function, since the essential
LLFloater::buildFromFile(optional LLXMLNodePtr) functionality has been
removed. This may mean that llui_libtest.cpp has little remaining value, not
sure.
Now LLProcess explicitly requests APR to limit the handles passed to any child
process, instead of wantonly passing whatever happens to be lying around the
parent process at the time.
This requires the latest APR build.
Also revert LLUpdateDownloader::Implementation::mDownloadStream to llofstream
(as in rev 1878a57aebd7) instead of apr_file_t*. Using APR for that file was a
Band-Aid -- a single whacked mole -- for the problem more systemically
addressed by apr_procattr_constrain_handle_set().
This handles the case when the target file doesn't exist, just as APR_TRUNCATE
handles the case when it does.
Strengthen error checks concerning downloaded installer file from
ll_apr_warn_status() to ll_apr_assert_status(). Failing to recognize (e.g.)
failure to open that file only leads to mysterious crashes down the road; this
removes the mystery.
On Windows, calling CreateProcess(bInheritHandles=FALSE) is the wrong idea. In
that case, CreateProcess() passes NO handles -- even the files you've
explicitly designated as the child's stdin, stdout, stderr in the STARTUPINFO
struct! Remove LLProcess code to tweak bInheritHandles; we should also remove
the corresponding (useless) APR extension.
Instead, given that the Windows file-locking problem we've observed is
specific to the viewer installer .exe file downloaded by the background
updater logic, use APR file I/O for that specific file. Empirically, both
llofstream and std::ofstream seem to make the open file handle inheritable;
but apr_file_open() documentation says: "By default, the returned file
descriptor will not be inherited by child processes created by
apr_proc_create()." And indeed, it does appear to sidestep the locking problem.
Using a Params block gives compile-time checking against attribute typos. One
might inadvertently set myLLSD["autofill"] = false and only discover it when
things behave strangely at runtime; but trying to set myParams.autofill will
produce a compile error.
However, it's excellent that the same LLProcess::create() method can accept
either LLProcess::Params or a properly-constructed LLSD block.
LLProcessLauncher had the somewhat fuzzy mandate of (1) accumulating
parameters with which to launch a child process and (2) sometimes tracking the
lifespan of the ensuing child process. But a valid LLProcessLauncher object
might or might not have ever been associated with an actual child process.
LLProcess specifically tracks a child process. In effect, it's a fairly thin
wrapper around a process HANDLE (on Windows) or pid_t (elsewhere), with
lifespan management thrown in. A static LLProcess::create() method launches a
new child; create() accepts an LLSD bundle with child parameters. So building
up a parameter bundle is deferred to LLSD rather than conflated with the
process management object.
Reconcile all known LLProcessLauncher consumers in the viewer code base,
notably the class unit tests.
On a Posix platform (vfork()/execv() implementation), if for any reason the
execv() failed (e.g. executable not on PATH), the viewer would never know, nor
the user: the vfork() child produced no output, and terminated with rc 0! Add
logging, make child terminate with nonzero rc.
Remove pointless addArgument(const char*) overload: this does nothing for you
that the compiler won't do implicitly.
In llupdateinstaller.cpp, remove pointless c_str() call in addArgument() arg:
we were starting with a std::string, then extracting its c_str(), only to
construct a whole new std::string from it!
This library was only needed on the Mac, and only with the OS X 10.4 SDK. As
of October 2010, we no longer build the viewer with that SDK. The 10.5 SDK
we're currently using directly supports the functionality for which we
originally brought in Pth.
accessed through the static LLThread::tldata().
Currently this object contains two (public) thread-local
objects: a LLAPRRootPool and a LLVolatileAPRPool.
The first is the general memory pool used by this thread
(and this thread alone), while the second is intended
for short lived memory allocations (needed for APR).
The advantages of not mixing those two is that the latter
is used most frequently, and as a result of it's nature
can be destroyed and reconstructed on a "regular" basis.
This patch adds LLAPRPool (completely replacing the old one),
which is a wrapper around apr_pool_t* and has complete
thread-safity checking.
Whenever an apr call requires memory for some resource,
a memory pool in the form of an LLAPRPool object can
be created with the same life-time as this resource;
assuring clean up of the memory no sooner, but also
not much later than the life-time of the resource
that needs the memory.
Many, many function calls and constructors had the
pool parameter simply removed (it is no longer the
concern of the developer, if you don't write code
that actually does an libapr call then you are no
longer bothered with memory pools at all).
However, I kept the notion of short-lived and
long-lived allocations alive (see my remark in
the jira here: https://jira.secondlife.com/browse/STORM-864?focusedCommentId=235356&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-235356
which requires that the LLAPRFile API needs
to allow the user to specify how long they
think a file will stay open. By choosing
'short_lived' as default for the constructor
that immediately opens a file, the number of
instances where this needs to be specified is
drastically reduced however (obviously, any
automatic LLAPRFile is short lived).
***
Addressed Boroondas remarks in https://codereview.secondlife.com/r/99/
regarding (doxygen) comments. This patch effectively only changes comments.
Includes some 'merge' stuff that ended up in llvocache.cpp
(while starting as a bug fix, now only resulting in a cleanup).
***
Added comment 'The use of apr_pool_t is OK here'.
Added this comment on every line where apr_pool_t
is correctly being used.
This should make it easier to spot (future) errors
where someone started to use apr_pool_t; you can
just grep all sources for 'apr_pool_t' and immediately
see where it's being used while LLAPRPool should
have been used.
Note that merging this patch is very easy:
If there are no other uses of apr_pool_t in the code
(one grep) and it compiles, then it will work.
***
Second Merge (needed to remove 'delete mCreationMutex'
from LLImageDecodeThread::~LLImageDecodeThread).
***
Added back #include <apr_pools.h>.
Apparently that is needed on libapr version 1.2.8.,
the version used by Linden Lab, for calls to
apr_queue_*. This is a bug in libapr (we also
include <apr_queue.h>, that is fixed in (at least) 1.3.7.
Note that 1.2.8 is VERY old. Even 1.3.x is old.
***
License fixes (GPL -> LGPL). And typo in comments.
Addresses merov's comments on the review board.
***
Added Merov's compile fixes for windows.
- Replaced all existing usages of LLDir::getNextFileInDir() with the new directory iterator object.
- Removed platform specific LLDir::getNextFileInDir() implementation.
Rider Linden