- Aug 04, 2009
-
-
Nat Goodspeed authored
Resources subdir without requiring user to explicitly set DYLD_LIBRARY_PATH. Engage the SHARED_LIB_STAGING_DIR mechanism for Mac as well as Windows and Linux so unit-test executables, too, can find libllcommon.dylib. Perform the POST_BUILD copy of libllcommon.dylib to SHARED_LIB_STAGING_DIR everywhere but Windows -- but make it conditional on SHARED_LIB_STAGING_DIR.
-
- Jul 31, 2009
-
-
Nat Goodspeed authored
-
Nat Goodspeed authored
-
Nat Goodspeed authored
-
Palmer Truelson authored
To be reviewed by Mani
-
Nat Goodspeed authored
self-identifies as @executable_path/../Resources/filename. This allows the SL executable to find it at runtime in its official place in the app bundle. Change viewer_manifest.py to copy libllcommon.dylib, libapr*.dylib to app bundle's Resources subdir.
-
- Jul 30, 2009
-
-
brad kittenbrink authored
-
brad kittenbrink authored
DEV-36893 make missing eventhost module a non-fatal error. We now print a warning and fail gracefully.
-
palmer@eniac54.lindenlab.com authored
Moves libllcommon.so to a staging dir for unit tests to work and gets rid of LL_COMMON_API in forward declarations
-
brad kittenbrink authored
-
brad kittenbrink authored
this doesn't actually fix anything yet though.
-
Nat Goodspeed authored
Update Mac apr_suite package, introducing libapr-1.0.3.7.dylib and libaprutil-1.0.3.8.dylib. We want to use dynamic APR libraries, but libllcommon.dylib isn't allowed to depend on .so libraries: must be .dylibs themselves. Tweak APR.cmake to reference the aforementioned .dylibs when on Mac with LLCOMMON_LINK_SHARED set. Eliminate some redundancy in Windows section. For .dylibs, RunBuildTest.cmake must set DYLD_LIBRARY_PATH: LD_LIBRARY_PATH only affects search for .so files. PENDING: CMake dependencies broken? With LLCOMMON_LINK_SHARED on, it seems we need to build libllcommon.dylib explicitly before full build can succeed. This snapshot does NOT yet copy the llcommon, apr or aprutil dylibs into the app bundle.
-
- Jul 29, 2009
-
-
Mark Palange (Mani) authored
DEV-35405 Added creation of <exe_dir>/components directory in debug builds to quell the savage llmozlib.
-
Mark Palange (Mani) authored
DEV-35405 Viewer manifest needed some love for supporting the debug build, but its still broke on llmozlib
-
- Jul 28, 2009
-
-
Mark Palange (Mani) authored
-
Mark Palange (Mani) authored
Fixing up string escaping on calls to RunBuildTest.cmake Also added ilasm.bat to the FindMono.cmake search
-
Bryan O'Sullivan authored
-
Nat Goodspeed authored
-
Mark Palange (Mani) authored
Now using RunBuildTest.cmake to run tut and lscript_lsl tests, inorder to set path to llcommon.dll Exported a few llcommon apis needed by server components/tests.
-
- Jul 27, 2009
-
-
brad kittenbrink authored
New package of xmlrpc-epi that actually links to the expat libs we ship. I have no idea how this has worked for so long.
-
- Jul 24, 2009
-
-
Mark Palange (Mani) authored
-
Mark Palange (Mani) authored
- Added LLNotificationsInterface class. - Removed LLLoginInstance use of LLNotifications EventAPI
-
- Jul 23, 2009
-
-
brad kittenbrink authored
-
- Jul 22, 2009
-
-
Mark Palange (Mani) authored
-
- Jul 20, 2009
-
-
Nat Goodspeed authored
-
Nat Goodspeed authored
-
Nat Goodspeed authored
-
Palmer authored
-
- Jul 16, 2009
-
-
brad kittenbrink authored
-
brad kittenbrink authored
-
- Jul 15, 2009
-
-
brad kittenbrink authored
-
Palmer authored
Cmake changes and fixes to python script to get it to work properly Mani reviewed most of this
-
brad kittenbrink authored
-
- Jul 13, 2009
-
-
brad kittenbrink authored
-
- Jul 10, 2009
-
-
brad kittenbrink authored
-
brad kittenbrink authored
First draft of cleaning up eventhost module on viewer shutdown. Still need to make the whole system more generalized and data driven.
-
brad kittenbrink authored
Fix for crash when running windows debug build. std::string::replace cannot take NULL for it's argument.
-
Palmer authored
Switching eventlet to use subrepo. By Brad, reviewed by Palmer
-
brad kittenbrink authored
Merged in my latest changes including switching LLAPRSockStream to be fully non-blocking (and thus not horribly inefficient) and a bunch of new event listener implementations.
-
brad kittenbrink authored
-