Enables the clang-tidy test performance-inefficient-string-concatenation
and replaces all inefficient string concatenations with `cmStrCat`.
Closes: #19555
For each build setting property (such as `COMPILE_DEFINITIONS` or
`INCLUDE_DIRECTORIES`), the value of `$<TARGET_PROPERTY:target,prop>`
includes the values of the corresponding `INTERFACE_*` usage requirement
property from the transitive closure of link libraries of the target.
Previously we computed this by constructing a generator expression
string like `$<TARGET_PROPERTY:lib,INTERFACE_COMPILE_DEFINITIONS>` and
recursively evaluating it with the generator expression engine. Avoid
the string construction and parsing by using the dedicated evaluation
method `cmGeneratorTarget::EvaluateInterfaceProperty`.
Issue: #18964, #18965
In large projects the generation process spends a lot of time evaluating
usage requirements through transitive interface properties on targets.
This can be seen in a contrived example with deep dependencies:
set(prev "")
foreach(i RANGE 1 500)
add_library(a${i} a.c)
target_compile_definitions(a${i} PUBLIC A${i})
target_link_libraries(a${i} PUBLIC ${prev})
set(prev a${i})
endforeach()
For each usage requirement (such as `INTERFACE_COMPILE_DEFINITIONS` or
`INTERFACE_INCLUDE_DIRECTORIES`), the value of the generator expression
`$<TARGET_PROPERTY:target,prop>` includes the values of the same
property from the transitive closure of link libraries of the target.
Previously we computed this by constructing a generator expression
string like `$<TARGET_PROPERTY:lib,INTERFACE_COMPILE_DEFINITIONS>` and
recursively evaluating it with the generator expression engine. Avoid
the string construction and parsing by creating and using a dedicated
evaluation method `cmGeneratorTarget::EvaluateInterfaceProperty` that
looks up the properties directly.
Issue: #18964, #18965
An old workaround for `std::allocator_traits<>::value_type` lints from
IWYU on `std::vector<>` usage breaks IWYU's handling of `<memory>`.
Convert the workaround to use the same approach we already use for a
workaround of `std::__decay_and_strip<>::::__type` lints. Then update
the `<memory>` inclusions to follow the now-correct IWYU lints.
For special properties like `INCLUDE_DIRECTORIES`, the pointer returned
by `cmTarget::GetProperty` is only valid until the next time the same
special property is queried on *any* target. When evaluating a nested
`TARGET_PROPERTY` generator expression we may look up such a property
more than once on different targets. Fix `TargetPropertyNode::Evaluate`
to store the lookup result in locally owned memory earlier.
Fixes: #19286
b53766b205 CUDA: Support compiler id and version generator expressions
b544e34af6 All VersionNode use the same capitalization pattern
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !3085
Extend the genex added by commit ca6ba3fee5 (Genex: Add a SHELL_PATH
expression, 2015-09-24, v3.4.0-rc1~37^2) to accept a `;`-list of paths,
convert them all, and generate a list separated by the native shell
`PATH``` separator.
Suppress some cases in `Source/cmGeneratorExpressionNode.cxx` and
`Source/cmUVHandlePtr.h` where a few older compilers require a
user-defined default constructor (with `{}`).
Adds `Fortran_COMPILER_ID` and `Fortran_COMPILER_VERSION` generator
expression support to match equivalent `C_COMPILER_ID`,
`CXX_COMPILER_ID`, `C_COMPILER_VERSION`, and `CXX_COMPILER_VERSION`
support.
This is very helpful in the case where the C/C++ compiler suite is a
different type of compiler from the platform Fortran compiler and
projects use generator expressions to assign compiler flags and
definitions. (e.g. `GNU` C/C++ and `SunPro` Fortran on Linux)
The old behavior of $<IN_LIST:...> is inconsistent with that of
if(IN_LIST), in that it does not find an empty search item even if
the list contains empty items. This change adds a new policy to
correctly handle empty items and make the behavior more consistent
with if(IN_LIST).
Fixes: #18556
While collecting usage requirements from the `INTERFACE_*` properties of
directly linked targets, we internally generate `TARGET_PROPERTY:` and
`TARGET_OBJECTS:` generator expressions to refer to those properties on
those targets. At the point we generate these expressions we already
have a pointer to an exact `cmGeneratorTarget` instance.
Switch from using the target name in these generator expressions to
using an internal unique name generated for each `cmGeneratorTarget`
instance to be referenced. This avoids depending on the user-facing
target name to find the same target we already have.
If two imported targets in different directories have the same name we
should still be able to propagate transitive usage requirements from
both. Fix the DAG checker to work with target pointers instead of
target names since the pointers will not be duplicated even if the names
are.
Fixes: #18345
Run the `clang-format.bash` script to update all our C and C++ code to a
new style defined by `.clang-format`. Use `clang-format` version 6.0.
* If you reached this commit for a line in `git blame`, re-run the blame
operation starting at the parent of this commit to see older history
for the content.
* See the parent commit for instructions to rebase a change across this
style transition commit.
Evaluate to false on `$<COMPILE_LANGUAGE:Lang>` if language `Lang`
is not loaded. This is helpful in exported targets consumed in other
projects that may not enable all the same languages.
Fixes: #17952
* Change some functions to take `std::string` instead of
`const char*` in the following classes: `cmMakeFile`, `cmake`,
`cmCoreTryCompile`, `cmSystemTools`, `cmState`, `cmLocalGenerator`
and a few others.
* Greatly reduce using of `const char*` overloads for
`cmSystemTools::MakeDirectory` and `cmSystemTools::RelativePath`.
* Remove many redundant `c_str()` conversions throughout the code.
The set of compile flags used for a target's C and C++ sources is based
on the linker language. By default this is always the C++ flags if any
C++ sources appear in the target, and otherwise the C flags. Therefore
we can define the `COMPILE_LANGUAGE` generator expression in
`INCLUDE_DIRECTORIES` to match the selected language.
This is not exactly the same as for other generators, but is the best VS
and Xcode can do. It is also sufficient for many use cases since the
set of include directories for C and C++ is frequently similar but may
be distinct from those for other languages like CUDA.
Fixes: #17435
The set of compile flags used for a target's C and C++ sources is based
on the linker language. By default this is always the C++ flags if any
C++ sources appear in the target, and otherwise the C flags. Therefore
we can define the `COMPILE_LANGUAGE` generator expression in
`COMPILE_DEFINITIONS` to match the selected language.
This is not exactly the same as for other generators, but is the best VS
and Xcode can do. It is also sufficient for many use cases since the
set of definitions for C and C++ is frequently similar but may be
distinct from those for other languages like CUDA.
Issue: #17435
Since commit v3.9.0-rc4~3^2~1 (VS: Fix target_compile_options for CUDA,
2017-06-21), the evaluation of `COMPILE_LANGUAGE` receives the proper
language. The set of compile flags used for a target's C and C++
sources is based on the linker language. By default this is always the
C++ flags if any C++ sources appear in the target, and otherwise the C
flags. Therefore we can define the `COMPILE_LANGUAGE` generator
expression in `COMPILE_OPTIONS` to match the selected language.
This is not exactly the same as for other generators, but is the best VS
can do. It is also sufficient for many use cases since the set of
allowed flags for C and C++ is almost the same in Visual Studio.
Furthermore, since the VS generator moves many of the flags to
declarative `.vcxproj` elements, it will automatically avoid passing
C++ flags for C sources.
Issue: #17435
When rejecting the `COMPILE_LANGUAGE` generator expression on include
directories and compile definitions with Xcode, add `file(GENERATE)` to
the allowed set in the message. It is allowed and already covered
by the `RunCMake.File_Generate` test `COMPILE_LANGUAGE-genex` case.
std::{begin,end} are part of C++11, std::{cbegin,cend} are part of C++14
and an standard compliant implementation has been introduced within the
'cm' namespace: cm::{cbegin,cend}.
std::size is only part of C++17, hence exposing a compliant implementation
within namespace cm (cm::size).
where possible, the standard implementations are reused.
We now require C++11 support including `override`. Drop use of
the old compatibility macro. Convert references as follows:
git grep -l CM_OVERRIDE -- '*.h' '*.hxx' '*.cxx' |
xargs sed -i 's/CM_OVERRIDE/override/g'
Teach install() and export() to handle the actual object files.
Disallow this on Xcode with multiple architectures because it
still cannot be cleanly supported there.
Co-Author: Brad King <brad.king@kitware.com>
Previously the `TARGET_OBJECTS` generator expression was limited
only to use in a buildsystem context so that Xcode's placeholders
in object file paths can be evaluated. Lift this restriction so
that the expression can at least be used in most settings.
Co-Author: Brad King <brad.king@kitware.com>
Move the diagnostic that rejects the TARGET_OBJECTS generator expression
in non-buildsystem context until after the check for whether the named
target is an object library. This order will makes more sense than the
previous order once TARGET_OBJECTS is allowed in non-buildsystem
context.
This allows a single condition to be used to choose between two
alternatives. Without this the condition must be duplicated with
one surrounded by `NOT`.
Closes: #15585
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.
When `$<LINK_ONLY:...>` is used outside of linking we may evaluate it
without a `dagChecker`. Do not dereference the NULL pointer and issue a
diagnostic instead.
Closes: #16287
Replace use of cmsys::auto_ptr with a CM_AUTO_PTR macro that maps to
our own implementation adopted from the KWSys auto_ptr implementation.
Later we may be able to map CM_AUTO_PTR to std::auto_ptr on compilers
that do not warn about it.
Automate the client site conversions:
git grep -l auto_ptr -- Source/ | grep -v Source/kwsys/ | xargs sed -i \
's|cmsys::auto_ptr|CM_AUTO_PTR|;s|cmsys/auto_ptr.hxx|cm_auto_ptr.hxx|'
