Add a `CMake_TEST_SERVER_MODE` option that can be set in testing builds
to enable/disable server mode tests explicitly. This will allow testing
in combination with `CMake_TEST_EXTERNAL_CMAKE` or for server mode to be
built on systems that have a python version that cannot run the test.
Provide a way for scripts building CMake to enable server mode
explicitly and assume the risk of a build failure if it is not
supported. This will allow such scripts to ensure that server
mode is available if the build succeeds. It also allows scripts
to explicitly disable server mode even if it would be supported.
71a50587 server-mode: Add project data for unit tests
7b1e60f2 server-mode: Report CMakeCache entries
84553a6e server-mode: Add command to retrieve build system files
ead71873 server-mode: Report information relevant for a codemodel
When hosting CMake's own build with the PGI compiler there are several
warnings that cannot easily be fixed or suppressed. We have enough code
quality tools and warnings from other compilers that it is unlikely a
PGI-only warning is useful.
The `PARSE_ARGV` mode was recently added to help functions properly
parse their arguments even when those arguments may be quoted and
contain literal `;` in their values. Fix the implementation to encode
`;`s in reported multi-value arguments and in `UNPARSED_ARGUMENTS` so
that `;`s in the individual values are preserved in the lists. This
allows clients to access all their argument values correctly.
Do some basic unit tests for "codemodel", "cmakeInputs" and "cache"
commands of the cmake server.
This just calls the commands right now and makes sure the server
thinks it can reply to the request. The data itself is currently not
validated.
On platforms that support shebang lines the `cxxtestgen` script should
be able to execute without an explicit interpreter. Do this if possible
so that the proper interpreter specified by the shebang line is honored.
Ideally we shouldn't even need to find `python` in this case, but that
will be a larger change better made by a dedicated module maintainer.
Closes: #16331
Update the toolset name matching added by commit v3.6.0-rc1~279^2~10
(VS: in Clang/C2 toolset, setup correct compiler settings, 2016-02-18)
to match VS/LLVM toolset names too.
The check added in commit v3.6.0-rc1~293^2 (Diagnose recursive
project/enable_language without crashing, 2016-03-07) broke support for
enabling `RC` explicitly along with other languages like `C`. The
reason is that we enable all listed languages at once so the internal
`enable_language(RC)` that we do while enabling `C` or `CXX` on some
platforms triggers the recursion check if `RC` is explicitly listed.
Ideally we should refactor things to only enable one language at a time,
but for now it is simplest to just exclude `RC` from the explicit list
until other languages are enabled, and then enable it.
Closes: #16330
The `main.cpp` that this test generates needs a newline so that
compilers do not warn about a missing newline. Otherwise the warning
causes RunCMake infrastructure to detect unexpected output.
The 64-bit Windows binary is built in our second batch because it is
built on the same host as the 32-bit Windows binary. When the
documentation generation was first added, it was not included in the
second batch because at the time that batch only included the Cygwin
binary. Therefore the 64-bit Windows binary has been missing the
documentation since it was first created for CMake 3.6.
Since then the Cygwin binary was dropped so the second batch only
contains the 64-bit Windows binary. Add the docs to it.
Closes: #16333
Per-source copyright/license notice headers that spell out copyright holder
names and years are hard to maintain and often out-of-date or plain wrong.
Precise contributor information is already maintained automatically by the
version control tool. Ultimately it is the receiver of a file who is
responsible for determining its licensing status, and per-source notices are
merely a convenience. Therefore it is simpler and more accurate for
each source to have a generic notice of the license name and references to
more detailed information on copyright holders and full license terms.
Our `Copyright.txt` file now contains a list of Contributors whose names
appeared source-level copyright notices. It also references version control
history for more precise information. Therefore we no longer need to spell
out the list of Contributors in each source file notice.
Replace CMake per-source copyright/license notice headers with a short
description of the license and links to `Copyright.txt` and online information
available from "https://cmake.org/licensing". The online URL also handles
cases of modules being copied out of our source into other projects, so we
can drop our notices about replacing links with full license text.
Run the `Utilities/Scripts/filter-notices.bash` script to perform the majority
of the replacements mechanically. Manually fix up shebang lines and trailing
newlines in a few files. Manually update the notices in a few files that the
script does not handle.
We do not require copyright assignment for contributions to CMake.
Update the top-level notice to mention "Contributors" explicitly.
Add a list of the Contributors whose names already appear in
source-level copyright notices. Find the list with the command:
$ git grep -h 'Copyright[^.]' -- *.* \
Auxiliary Modules Packaging Source Templates Tests \
| grep -v Kitware | sed 's/^#\? *//' | sort | uniq
Manually refine the list to de-duplicate entries and filter out
notices from third-party code.
The installer should not put up a message box to confirm uninstall of a
previous version for silent installs. Furthermore, when the uninstaller
is invoked, do not show its confirmation dialog because either (a) the
installer already asked the user or (b) the installer is running in
silent mode.
Closes: #16328
Brad King
Kitware Robot
Matthew Woehlke
Tobias Hunger
Roman Wüger
Dmitry Marakasov
Orion Poplawski
Craig Scott