Fixes: #22096

Since commit f65f2093 (Autogen: Avoid processing CSharp targets,
2020-11-12, v3.20.0-rc1~301^2) we collect all sources for a target
earlier than previously.  Clear the sources cache so that it will be
re-computed later after AUTOGEN processing.

Fixes: #22085

Refactoring in commit bdc40742 (CMakeDetermineCompilerId: Test without
COMPILER_ID_FLAGS if REQUIRE_SUCCESS, 2021-02-27, v3.20.0-rc3~6^2) added
an extra macro layer through which flag strings are passed.  That caused
an extra level of argument re-parsing, and broke flags with backslashes.
Pass flags to the helper macro through variable names instead.

Fixes: #22041

The NVHPC compiler does not have any implicit link directories that can
be used to detect `CMAKE_LIBRARY_ARCHITECTURE`, but it does have
implicit object files.  Extract implicit object file paths from link
lines and check them for the `CMAKE_LIBRARY_ARCHITECTURE` pattern.

Issue: #22024

Robert Maynard authored 3 years ago and Brad King committed 3 years ago

The change in commit 657fc3a9 (CMakeDetermineCompilerABI: Parse
library arch from versioned paths, 2021-02-03, v3.20.0-rc1~40^2) caused
`CMAKE_LIBRARY_ARCHITECTURE` to be populated on non-multiarch platforms
if their compilers happen to use `$arch/$version` library directories.
Revert the use of versioned library paths.

Fixes: #22024

Fixes: #22002

Fixes: #22007

Many tests covering CMake scripting commands can work without
initializing a full generator.

Fixes: #21994

With generator expressions in a custom command's `OUTPUT` and
`BYPRODUCTS`, it is possible to have no outputs at all for a particular
configuration.  Generate no rule in this case.

Fixes: #21989

Ninja versions prior to 1.6.0 occasionally print this message.

Fixes #21966

Correctly compute the next parse index for `ZeroOrOne` when
it is zero, and for OneOrMore.

Fixes: #21973

Under the CMP0118 NEW behavior, sources generated in one directory
should be visible when added to targets in other directories.  This was
accidentally left out of commit 6624b65b (GENERATED prop: Add
implementation for policy CMP0118 being set to NEW, 2020-11-09,
v3.20.0-rc1~393^2~1).

Fixes: #18399

In commit ce2dee9e (Xcode: Don't add framework as -framework argument
in linker info list, 2020-09-28, v3.19.0-rc1~47^2) we split up the path
to a framework into the directory and framework name parts, but only
retained the quoting on the directory part.  Restore quoting of the
framework name.

Fixes: #21910

Refactoring of the ExternalProject and FetchContent modules moved
the commands into CMake scripts. This broke custom commands that
used shell redirection or special build tool variables of the form
$(MakeVar). Undo the sequence of commits that performed this
refactoring and follow-up fixes associated with it.

The following commits are reverted by this change:

4f3d1abb (ExternalProject: Refactor pre-configure steps to support
no-target uses, 2021-02-05)

17e5516e (FetchContent: Invoke steps directly and avoid a separate
sub-build, 2021-01-29)

bd876f38 (FetchContent: Restore patch command support,
2021-02-18)

404cddb7 (ExternalProject: Fix misuse of IS_NEWER_THAN in
timestamp checks, 2021-02-21)

b0da6712 (FetchContent: Don't update timestamps if files don't
change, 2021-02-18)

Fixes: #21892

Use `CMAKE_ANDROID_NDK_VERSION` to check for files expected based
on the version of the NDK.

Kyle Edwards authored 4 years ago and Brad King committed 4 years ago

Fixes: #21858

Upstream ninja commit `ad3d29fb53` (Put builder output through status
interface) from ninja PR 1899 changed the status output from stdout to
stderr.  In particular, `ninja: no work to do` is now printed on stderr.
Update our RunCMake tests to accept this difference.

A few RunCMake test cases check for `ninja: no work to do`.  For those,
move the message to stdout using `RunCMake_TEST_OUTPUT_MERGE`.  The rest
of the test cases do not care about the message, so remove it from the
actual stderr content before comparing against that expected.

The main `cmake --preset` argument for configure presets supports both
forms, so support it for `ctest --preset` too.

Fixes: #21855

The main `cmake --preset` argument for configure presets supports both
forms, so support it for `cmake --build --preset` too.

Issue: #21855

This form already works.  Document it and add tests.

Issue: #21855

Marc Chevrier authored 4 years ago and Brad King committed 4 years ago

If target property LINKER_LANGUAGE is set, LINK_LANGUAGE generator
expression evaluation must be always successful.

This fix can be helpful to elaborate a solution for issue #21818.

Some tests fail because Xcode runs `POST_BUILD` commands before signing
the binaries they run.  Tell the linker to perform ad-hoc codesign even
though Xcode normally tells it not to.

Other tests fail because `install_name_tool` does not revise ad-hoc
signatures without the codesign `linker-signed` flag.  Add that flag
ourselves where needed by our tests.

For now these changes help our test suite pass so we can use it to cover
everything else.  Both of these cases may need further investigation to
update CMake to help projects in general.

Issue: #21845, #21854

These were added by commit e29a92f5 (Xcode: Make BundleTests
compatible with Xcode 11, 2019-09-01, v3.16.0-rc1~158^2).  The
real problem is that the `.app` is constructed with text files
under the `.app/Contents/MacOS/` folder.  Move those, and drop
the special no-signing configuration.

Daan De Meyer authored 4 years ago and Brad King committed 4 years ago

Following commit 7155e358 (ExternalProject: Add CONFIGURE_HANDLED_BY_BUILD
option, 2020-12-16, v3.20.0-rc1~168^2), modify the CONFIGURE_HANDLED_BY_BUILD
test to sleep 1.125 seconds to make sure the file timestamp is always
updated regardless of the resolution of the underlying filesystem.

Fixes: #21830

Craig Scott authored 4 years ago and Brad King committed 4 years ago

The refactoring in 17e5516e (FetchContent: Invoke steps directly and
avoid a separate sub-build, 2021-01-29) uses a different way of writing
out the step scripts and updating time stamps when steps are executed.
That inadvertently always wrote out the scripts for custom commands,
even when the contents didn't change. This caused their timestamp to
always be updated, resulting in those steps always being seen as
out-of-date and needing to be re-executed.

The way timestamps were checked to determine whether to re-execute
a step also did not adequately account for file systems which only have
second-resolution timestamps. The IS_NEWER_THAN if condition also
returns true when timestamps are the same, so one needs to use the
negative form to get a true "is newer than" test.

ExternalProject is not susceptible to this problem because it uses
file(GENERATE) to write out the script files and that only updates the file's
timestamp if the contents change. It also mostly leaves timestamp
checking to the build tool.