The test regularly fails updating the `vc*.pdb` compiler-generated
PDB file.  Add the `/Z7` flag as the compiler suggests for this.

Since commit ddcd1469 (MSYS: Add support for running under MSYS
runtime environment, 2021-04-01) the test uses `std::size_t` without
including `<cstddef>`.  Also, the type does not exist under `std::`
on VS 6.  Use just `size_t` instead.

Sergey authored 3 years ago and Brad King committed 3 years ago

New option `-A` passed to test binary allows to run all testcases at
once and prints a report in a standard format - TestAnythingProtocol v.13 [1].
Execution of test whose names will be passed after an option will be skipped.

Sample of output:

    TAP version 13
    1..6
    ok 1 TestCryptoHash # 0.030000
    ok 2 TestCryptoRand # 0.008000
    not ok 3 TestCryptoCipher # 0.005000
    ok 4 TestCryptoProtectData # 0.000000
    cbPlainText: 21 cbCipherText: 32
    PlainText: MySecretPassword123! (cbPlainText = 21, cbCipherText = 32)
    Decrypted CipherText: MySecretPassword123!
    ok 5 TestCryptoProtectMemory # 0.014000
    ok 6 TestCryptoCertEnumCertificatesInStore # 0.000000

1. https://testanything.org/

Fixes: #19367

Orgad Shaneh authored 3 years ago and Brad King committed 3 years ago

Detect MSYS as CYGWIN, with the required adaptations.

Addresses #18654

Consider the case where the find_package call for QtCore is wrapped in a
function call. Then AutoGen cannot determine the Qt version, because it
only looks at variables and directory properties. The former don't leave
the function scope and the latter are not set by default.

As a fallback, locate the moc executable via its target and call it with
the --version argument to determine the Qt version.

Issue: #22028

Fixes: #22096

Nils Gladitz authored 3 years ago and Brad King committed 3 years ago

Merge use of SetFilterOption() into more abstract thread count
in cmArchiveWrite constructor.

libarchive defaulting of threads for threads == 0 seems to be
configuration dependent. Preemptively default thread count via
std::thread::hardware_concurrency().

Also allow negative values for the thread count in which case
the detected hardware concurrency is also used but the given
absolute thread count is used as an upper limit.

Format message content for `message(FATAL_ERROR)`.

Alexander Akhundzhanov authored 3 years ago and Craig Scott committed 3 years ago

Co-Authored-By: Craig Scott <craig.scott@crascit.com>

Marc Chevrier authored 3 years ago and Craig Scott committed 3 years ago

This option enables the replacement of any leading tilde with the path
to the user's home directory.

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

This MR extend the support of 'DEPFILE' to buildsystem version 1.

Issue: #20286

nic-kaczinsky authored 3 years ago and Brad King committed 3 years ago

Fix the warning: `potential leak of memory pointed to by "vp"`.

Marc Chevrier authored 3 years ago and Brad King committed 3 years ago

Issue: #20286

This facility is very useful for 'Ninja Multi-Config' and required
as well for future support of DEPFILE in 'Xcode' and 'Visual Studio'
generators (#20286).

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

In v3 of the presets, generator and buildDir can be omitted to fall
back to regular cmake behavior when these values are not explicitly
provided by the user.

Fixes: #21987

These messages are incidental scheduling output.  Avoid cluttering
the normal verbose output showing tests starting and finishing.

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

Rolf Eike Beer authored 3 years ago and Brad King committed 3 years ago

This needlessly produces warnings during the test runs that no-one
sees but that are distracting when actually inspecting the logs.

- None is a valid CMAKE_BUILD_TYPE
- Most of distros uses None as CMAKE_BUILD_TYPE
- When CMAKE_BUILD_TYPE=None, main_other.cpp will be compiled and linked
  into final executable, this program requires some symbols only exist
  when CUSTOM_CFG_OTHER is defined.
- And CMake also allows other CMAKE_BUILD_TYPE, too, CMake documentation
  specificaly mentions funny CMAKE_BUILD_TYPE like ReLeAsE [1]

Let's define them when non main stream like None is specified as CMAKE_BUILD_TYPE.

[1]: https://cmake.org/cmake/help/v3.20/variable/CMAKE_BUILD_TYPE.html

Scriplet tags should not be added to generated
spec files if scripts weren't provided as those
tags are otherwise present in generated rpm file
even if the script wasn't provided thereby
generating unneeded dependency on shell.

Fixes: #21345

Fixes: #22007

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