Expose the binutils' machine name (typically used as a prefix on the
tool names) publicly. This is expected to match the `gcc -dumpmachine`
value.
Suggested-by: Ruslan Baratov <ruslan_baratov@yahoo.com>
The NDK provides prebuilt toolchain files in directories named for the
host architecture. The NDK build system calls this `HOST_TAG`.
Expose the value publicly for use by clients that need to pass it
to external tools.
Suggested-by: Ruslan Baratov <ruslan_baratov@yahoo.com>
When this variable is not set by the user or toolchain file, set it to
the default selected. This will be useful for client code that needs to
pass the value to an external tool that needs to find the same toolchain
in the NDK. Leave it empty for a standalone toolchain.
Suggested-by: Ruslan Baratov <ruslan_baratov@yahoo.com>
Since commit v3.7.0-rc1~348^2~3 (FindBISON: Change usage of [VERBOSE
<file>] to [VERBOSE [<file>]], 2016-07-16) we always list the VERBOSE
output file as an output of our custom command even if the option is not
used. This causes the rule to re-run every time (e.g. with Ninja).
Revert one hunk from that change (that looks incorrect) to fix it.
Closes: #16426
Refactoring in commit v3.6.0-rc1~134^2~10 (Autogen: Split out moc file
generation code to dedicated method, 2016-04-18) removed the
unconditional creation of the `<target>_automoc.cpp` file. Now it is
generated only when `AUTOMOC` is enabled. However, if this file is not
created then our internal `GenerateAll` setting is enabled on every
build, causing `AUTORCC` to re-generate its file(s) every time. Fix the
`GenerateAll` setting to be used only for when autogen settings change.
The old logic was left from when we had only automoc.
Closes: #16413
The options to the find_library call to create the imported target
used a literal string "HINTS /path NO_DEFAULT_PATH" instead of a
list of options. This resulted in never finding any library in my
testing.
Refactoring in commit v3.6.0-rc1~72^2 (HDF5: Rework component searching
to correctly find HL for all bindings, 2016-05-12) changed the default
behavior from finding only the C bindings to finding everything for the
enabled languages. Restore the original behavior for compatibility and
because many projects need only the C bindings.
Closes: #16397
Running CMake on it caused the following error:
error: Target "publisher" links to target "Qt5::DBus" but the target
was not found. Perhaps a find_package() call is missing for an
IMPORTED target, or an ALIAS target is missing?
Add the missing DBus component.
`CPACK_RPM_COMPONENT_INSTALL` is the correct variable to set to enable
component packaging. `CPACK_RPM_PACKAGE_COMPONENT` is just set to a
component name when CPack calls corresponding installer.
`CPACK_DEB_COMPONENT_INSTALL` is the correct variable to set to enable
component packaging. `CPACK_DEB_PACKAGE_COMPONENT` is just set to a
component name when CPack calls corresponding installer.
Starting with Xcode 8 the SDK folder also contains an unversioned entry:
MacOSX.sdk
MacOSX10.12.sdk -> MacOSX.sdk
If this unversioned path is used CMake cannot detect the SDK version.
Furthermore, querying the SDK version via
xcodebuild -sdk <sysroot> -version Path
gives bogus results for the Command Line Tools installed into `/`.
The OS X deployment target version and SDK version are not as tied as
they once were, so this check is now more trouble than it is worth.
Simply remove it.
Closes: #16323
Revert commit v3.7.0-rc1~48^2 (Xcode: Convert maybe unversioned OSX
sysroot into versioned SDK path, 2016-09-25). The replacement of
`else()` with `if(CMAKE_OSX_SYSROOT)` defeats the prior handling of
`if("x${CMAKE_OSX_SYSROOT}" MATCHES "/")`. This causes the combination
-DCMAKE_OSX_DEPLOYMENT_TARGET:STRING="" -DCMAKE_OSX_SYSROOT:STRING=/
to not be honored and `-isysroot` to be emitted as a compiler flag
universally. We will need another solution to the problem the
now-reverted commit was meant to address.
Closes: #16394
Fix typo in variable used to communicate between matlab_add_unit_test
and its helper script MatlabTestsRedirect.cmake that was introduced in
commit v3.7.0-rc1~116^2 (FindMatlab: Extend matlab_add_unit_test to run
arbitrary test code, 2016-08-30).
Catch more problematic input during handshake and report failure.
These were caught before when trying to configure, but it is way better
to get these reports early.
Change our message wrapper from
[== CMake Server ==[ ... ]== CMake Server ==]
to
[== "CMake Server" ==[ ... ]== "CMake Server" ==]
to guarantee that no JSON content can ever contain the ending string
(because it would be encoded as `]== \"CMake Server\" ==]`).
The extension flags enabled by commit v3.6.0-rc1~120^2~1 (Features:
Record standard flags for Intel C/C++ on Windows, 2016-04-18) of the
form `-Qstd=gnu++11` are not supported by the Intel C/C++ Compiler for
Windows. Fall back to using the non-extension form of the flags.
Issue: #16384
Add `-fPIE -pie` to the default executable link flags when
`CMAKE_POSITION_INDEPENDENT_CODE` is enabled. This is required by
Android 16 and above for executables to run on the device.
Closes: #16382
If the toolchain file or cache does not set this, enable it
automatically based on the Android API version. Versions 16
and above expect position independent code.
Use the main `CMAKE_POSITION_INDEPENDENT_CODE` setting in favor of
hard-coding `-fpic` or `-fPIC` in the compiler flags for each ABI.
This allows CMake to use `-fpie` or `-fPIE` as needed when sources
are meant for executables, and `-fpic` or `-fPIC` for other sources.
The targets added by commit v3.6.0-rc1~240^2~2 (Ninja: Add `$subdir/all`
targets, 2016-03-11) use as `$subdir` the relative path from the top of
the source tree to the current source directory. This is not correct
when using `add_subdirectory(test test_bin)`. Instead we need to use
the relative path from the top of the binary tree to the current binary
directory as was done for related targets by commit v3.7.0-rc1~268^2
(Ninja: Add `$subdir/{test,install,package}` targets, 2016-08-05).