| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
This time for installation/packaging.
|
|
|
|
Referred from cpuinfo.
|
|
until we are ready to enable media plugins on Windows ARM64.
|
|
Referred from cpuinfo.
|
|
|
|
LLQuad is a typedef of __m128, which is already translated by
sse2neon to float32x4_t (I thought sse2neon wasn't taking effect
and I tried just replacing __m128 with float32x4_t to see that it
didn't make a difference), but then I searched using the keyword
float32x4_t this time and found that others have had a similar
problem:
https://developercommunity.visualstudio.com/t/static-initialization-arm64-neon-datatypes/1238406
https://stackoverflow.com/questions/54016821/error-c2078-when-initializing-uint32x4-t-on-arm
https://github.com/kcat/openal-soft/issues/494
Looking at the type definition, on arm64 it can be initialised
using a designator, the member with the float type and 4 array
elements.
I know it's an MSVC (arm64) problem, but since MSVC is also used
on x64 and only Windows arm64 is suffering from this one in our
case anyway (we only support Windows arm64 building using MSVC so
far), it's just simpler to use the _M_ARM64 preprocessor instead
of _MSC_VER.
|
|
This reverts commit 75012c4f0c0b852c8bfa0b5632bf82c82f240754.
I realised I was in the wrong direction after I avoided the
offending lines by setting GLM_FORCE_XYZW_ONLY, and then by setting
GLM_FORCE_CXX_UNKNOWN, to end up with the same 16 __m128-related
"too many initializers" errors, which I believe were just obscured
by the 160 syntax errors, which are to be fixed in the next commit.
I had already mentioned SSE2NEON being responsible for translating
__m128, but I had forgotten that it was a different package from
GLM, obviously, that __m128 was never GLM_FORCE_NEON's
responsibility.
|
|
Basically just add what the author added to the file.
https://github.com/JochenKalmbach/StackWalker/commit/2a96079ba54a905d29afb3834182cbf06eff9132#diff-0d2beffe54d2ad0b9c31814fc48e5ef5841ba542fe545bba8b8971c767c89911
|
|
Also add cpuinfo to build preparation instruction.
|
|
Somehow when the stream editing to delete anything that begins with 4
spaces is run from CMake (as opposed to directly on shell), it deletes
most of the contributors on Windows. I've tried many methods like using
\w, \s, etc., but what works is just using the first letters of the bug
codenames. There are still empty results (, , ,) among the final parsed
contributors, but we'll ignore them for now.
|
|
275 became 240 when they removed the logos, and then 240 became 220
when MotD became dynamic.
So logos being back but with dynamic MotD should be 255.
|
|
https://megapahit.com/show_bug.cgi?id=163
|
|
|
|
I checked out the related files from 45249de1ca418324c4077312eefc7edc7e22c401,
before the merge with upstream logo removal, but then I had to reapply
c2d491905b668702d5640c7c5472629f7acc27e0, the MotD dynamic resizing, and
remove the unused commented out vivox-related part since we're not
redistributing it.
|
|
Make licenses more readable, neater looking.
|
|
though there are still so many compile errors, but it should be in
the right direction, cause firstly we can't let it use GLM_FORCE_SSE2
otherwise it would include x/emmintrin headers, and secondly, if
none of the GLM_FORCEs is used then the errors are much fewer but
related to __m128 which should be translatable by SSE2NEON (that's
why I believe GLM_FORCE_NEON should be ON for Windows arm64 too).
|
|
This commit is for getting rid of the compile error for now,
we don't know yet if this will work correctly or not.
|
|
Just copying from this:
https://forum.juce.com/t/windows-arm64-support/47572/6
This commit is for getting rid of the compile error for now,
we don't know yet if this will work correctly or not.
|
|
before using it for ColladaDOM library file, cause it seems that
ColladaDOM is the first one that gets built on Windows arm64 before
any other package gets installed and implicitly creating that folder.
|
|
move button_row_ls up a little bit so that the buttons are not cut off by the bottom edge of the floater.
|
|
|
|
It's irrelevant on any distro anyway.
|
|
'std::char_traits<unsigned short>' on FreeBSD.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=282329
|
|
|
|
I felt like they slowed down the flow.
https://developercommunity.visualstudio.com/t/warnings-in-a-windows-sdk-header-when-compiling-wi/897227
|
|
Parsing failed on macOS' CMake (could be MacPorts, could be the version)
for some reason.
|
|
|
|
|
|
cURL is installed on arm64 cause there doesn't seem to be any easy
way to build OpenSSL 1.1 for Windows arm64 (which LL's cURL fork
depends on).
|
|
Crow Misia's libwebrtc-bin, the one we've been relying on for WebRTC
binaries, doesn't seem to be providing Windows arm64 binaries yet.
|
|
(Eventually - marshalled by [GIRD LOWER])
|
|
Referring to cpack-generators(7) manual, I don't know what to replace
the <compName> in CPACK_NSIS_<compName>_INSTALL_DIRECTORY with, but in
CPACK_NSIS_INSTALL_ROOT explanation, it mentions the full directory
presented to the end user, and from that we can conclude the variable we
need to set for customisation is just CPACK_PACKAGE_INSTALL_DIRECTORY.
|
|
https://stackoverflow.com/questions/5801813/c-usleep-is-obsolete-workarounds-for-windows-mingw
|
|
Uncomment a couple of commented out lines out in
indra/llplugin/slplugin.cpp and adjust to something like:
LLError::setTagLevel("Plugin", LLError::LEVEL_DEBUG);
LLError::logToFile("C:\\Users\\erik\\AppData\\Roaming\\Megapahit\\logs\\slplugin.log");
I got
"The requested service provider could not be loaded or initialized"
This really helped me solve it:
https://travis-ci.community/t/socket-the-requested-service-provider-could-not-be-loaded-or-initialized/1127
|
|
I started with disabling it for having BUILD_SHARED_LIBS on.
Then I enabled it for macOS when I added internal browser support for
macOS.
Then I enabled it for Linux when I added internal browser support for
Linux.
Then I enabled it for FreeBSD when I added internal browser support
attempt for FreeBSD.
I should have left it enabled by default except for platforms that
surely are not supported to have internal browser.
|
|
|
|
|
|
and minimise diff related to LibVLC headers.
|
|
Move categories dropdown button so that it is not cut off by the edge of the floater, also fix "Arts and Culture" & "Parks and Nature" places categories displaying "MissingString"
|
|
in viewer linking stage on Windows.
throw_exception is user defined (it's not picking up its definition from
anywhere on Windows, at least when using vcpkg's Boost).
(Link target) ->
libcollada14dom23-s.lib(daeURI.obj) : error LNK2019: unresolved external symbol
void __cdecl boost::throw_exception(class std::exception const &)
(?throw_exception@boost@@YAXAEBVexception@std@@@Z) referenced in function
void __cdecl boost::re_detail_500::raise_runtime_error<class boost::regex_error>(class boost::regex_error const &)
(??$raise_runtime_error@Vregex_error@boost@@@re_detail_500@boost@@YAXAEBVregex_error@1@@Z)
[C:\Users\erik\Documents\Megapahit\viewer\build-mingw64_nt-10.0-19045-x86_64\newview\megapahit.vcxproj]
|
|
|
|
Less diffs. And only a couple.
|
|
|
|
center floater anymore because it obscures the avatar.
|
|
Put the necessary files into place.
But, none of them is working just yet.
|
|
This reverts commit f0de1898e0b6a3b9f1ee1bcc762ba334b78bbcc6.
|
|
Turns out the CMAKE_BUILD_TYPE setting is necessary, otherwise it would
be built in RelWithDebInfo configuration.
Reorganise the contributors generation and general CPack settings in
indra/newview/CMakeLists.txt.
Running sed when cross-comping for Linux, on FreeBSD, doesn't need to
use Linux's sed, so no need for ${CMAKE_SYSROOT}/usr/bin prefix, this
way that section can be reused for Windows.
I still couldn't get CPack to make NSIS not use the version numbers as
part of the default installation destination.
Using TARGETS for installing llwebrtc would cause the .lib to be
installed too, which isn't necessary, that's why we use PROGRAMS.
contributors.txt still gets generated wrongly.
The executable icon is still SL's test icon.
ColladaDOM's failure to link to Boost throw_exception, from its use of
Boost Regex, is not fixed yet. I got to this stage by temporarily
removing the offending lines in daeURI.cpp (which are the lines where
Boost Regex is used).
|
