Fix remaining occurrences of the issue addressed in commit 5962db4389
(Use C++11 nullptr, 2017-08-22) that are only showing up on macOS.
Signed-off-by: Matthias Maennich <matthias@maennich.net>
The opened XML elements were not closed, so an assert was triggered when the
file was finally closed. If CMake is built with assertions disabled then an
invalid XML file will be produced.
Calls to this method may dominate generation time in some cases.
Measurements for configuring cmake itself show a cache hit rate of ~57%
(7753 total calls, 4453 cache hits). For a larger project (that also
makes use of custom targets as prerequisite for all compile targets),
the measured cache hit ratio is ~96% (2530827 total calls, 2433124 cache
hits). For this project the observable cmake runtime could be reduced
from 40s to 30s.
Signed-off-by: Matthias Maennich <matthias@maennich.net>
The commit v3.7.0-rc1~339^2 (Ninja: Fix inter-target order-only
dependencies of custom command, 2016-07-20) might cause performance
degradations for larger projects. Especially when using custom
commands as an input for each compilation rule (e.g. generated headers).
For reference in the following I am referring to
Source/cmGlobalNinjaGenerator.cxx:
-> cmGlobalNinjaGenerator::AppendTargetDependsClosure
-> cmGlobalNinjaGenerator::ComputeTargetDependsClosure
It turned out that the mentioned commit is doing (indirectly) some
redundant work that might impact performance when generating large
projects.
Imagine the dependency tree of custom targets:
A
\
C - D - E
/
B
For each target the transitive closure is calculated recursively, but as
the TargetDependsClosures are only cached on the top most level, everything
downstream has to be recalculated. I.e.
A->C->D->E
B->C->D->E
This ultimately leads to a lot of redundant calls to AppendTargetOutputs.
The recursive nature of the algorithm itself is not significant to the
problem, but reducing the work to actually to be done work, eliminates the
performance problem.
This patch changes the way, intermediate results are cached. Rather than
caching the closure of targets, we cache the closure of outputs. Such that
in the example above at B->C the cache already would kick in.
Caching the outputs has one disadvantage that the patch takes care of.
In case of such a structure
A E
\ / \
C - D G
/ \ /
B F
the calling order for A would be
A->C->D->E->G (at which time G is seen to the recursion)
then the recursion returns until it reaches
A->C->D->F (at which the seen G would prevent to recurse down to G)
But this would poison the cache for F with a wrong value (without G).
Hence we use a local result set to ensure the cache is still consistently
populated.
For a large C++ project with around 25k targets this reduced the CMake
configure / generate time from ~40s to ~29s.
Signed-off-by: Matthias Maennich <matthias@maennich.net>
Filter out `-Winvalid-command-line-argument` warnings from Clang (that
can be caused by user-specified flags) so that they do not break our
checks for C++ feature availability.
Add special parsing of the flags given in `/MANIFESTUAC:"..."` in order
to map them correctly to `.vcxproj` elements.
Keep the old incorrect flag table entries for `uiAccess` and `level`
flags for compatibility even though they do not really exist.
Fixes: #16563
67529aab Doc: document that CoverageExtraFlags will come first
69fac3c3 pass arguments as vector to cmCTest::RunCommand()
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !1261
714ce728 bootstrap: Make libuv available during bootstrap
6a2d967d bootstrap: Require compiler mode aware of C99 on Solaris
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !1131
On UNIX, build only the parts of libuv we need for the filesystem,
process, and poll abstractions using the POSIX poll() backend. This
avoids many platform-specific conditions. On Windows, build all of
libuv; there are no conditional alternatives anyway.
The only 2 callers took care to construct a properly escaped string, but not
using the documented way, and that string was passed only to be immediately
split into tokens again. Start with a vector and join it only for logging,
avoiding needless quotes during that.
This Qt flag enables UI scaling based on scale factor provided by the
operating system. Only supported on versions of Qt >= 5.6.
The flag has no effect on macOS, where high DPI support is already
enabled as a setting in the bundle's info.plist.
Fixes: #17026
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'
Add variable `CMAKE_CODEBLOCKS_EXCLUDE_EXTERNAL_FILES` to optionally
exclude files from outside the project root from the project file
written by the CodeBlocks extra generator. This optionally restores
logic that had been removed by commit v2.8.3~40^2 (CodeBlocks Generator:
Do not omit files in the project file listing, 2010-10-05) in response
to QTCREATORBUG-2250.
Issue: #12110Fixes: #17188
This commit changes the internal -E__run_iwyu to be -E__run_co_compile. This
is used for co-compile commands. These are tools that want to mirror the
compiler. For each compiler invocation the tool will be invoked first. This
started as a way to implement include what you use (iwyu), but has expanded
to include cpplint, cppcheck and others. Likely there will be more in the
future as well. This commit implements each one in its own function and
provides a way to add additional ones in the future with less work.
Target dependencies of the origin target were mistakenly forwarded to
the _autogen target as *file* dependencies. This patch introduces
proper distinction between *target* and *file* dependencies of the
_autogen target.
This patch also changes when PRE_BUILD is used for AUTOGEN in the Visual
Studio generator. Formerly PRE_BUILD was disabled when the origin target
depended on *any* other target. Now PRE_BUILD is only disabled if a
dependency of the _autogen target to an additional *file* is detected.
Fixes: #17278, #17205
The KWSys `SystemTools::SplitProgramFromArgs` implementation goes into
an infinite loop when the value is just " " (a space). Since the
"program path with unquoted spaces plus command-line arguments"
operation it is trying to provide is poorly defined (string parsing
should not depend on filesystem content), just stop using it.
Instead consider the main two use cases the old approach tried to handle:
* The value is the name or absolute path of a program with no quoting
or escaping, but also no command-line arguments. In this case we
can use the value as given with no parsing, and assume no arguments.
* The value is a command-line string containing the program name/path
plus arguments. In this case we now assume that the command line
is properly quoted or escaped.
Fixes: #17262
Since commit v3.9.0-rc1~309^2 (include_external_msproject: Honor
MAP_IMPORTED_CONFIG_<CONFIG>, 2017-04-04) we accidentally honor
`MAP_IMPORTED_CONFIG_<CONFIG>` while generating the `.sln` file entries
for normal targets. This causes `devenv.com`-driven builds to use the
mapping incorrectly for normal targets. Check that a target really
comes from `include_external_msproject` before considering the map.
Furthermore, when we do use the map, we should only take the first entry
if more than one configuration is specified. Otherwise we end up giving
VS a configuration name with a `;` in it.
Fixes: #17276
Kitware Robot
Matthias Maennich
Rolf Eike Beer
Brad King
Pavel Solodovnikov
comicfans
Luis Caro
Alexandr (Sagrer) Gridnev
Bill Hoffman
Sebastian Holtermann
Minmin Gong