| Age | Commit message (Collapse) | Author |
|---|
|
LLVM became more complex, use LLVM's CMake functionality directly instead
of replicating this behaviour in KLEE's build system.
Use the correct build flags provided by LLVM itself.
This is influenced by the way LLVM is built in the first place.
Remove older CMake support (< 3.0).
|
|
|
|
|
|
This is a patch that I made few months ago as Fedora forbids bundling and
using pieces of software provided by other packages in its repositories but
forgot to upstream it at that time. [1]
It has been rebased and improved so that it also reflects changes made
in #1458.
This should also make the compilation of unittests easier for our users
as they don't need to clone googletest from GitHub anymore and just use
package manager in the distro of their choice, provided that the gtest
package includes a corresponding CMake module.
[1]: https://src.fedoraproject.org/rpms/klee/blob/4c81b78/f/use-system-gtest.patch
|
|
|
|
|
|
|
|
|
|
|
|
Before, we would use gtest_main as provided by either LLVM or Google
Test itself; the TestMain.cpp was ignored (contrary to its source code
comment). In newer versions of Google Test (1.8.1+), gtest_main uses
`__FILE__` for its "Running main() from" line, but llvm-lit (which we
use to invoke unit tests) currently matches exactly "Running main()
from gtest_main.cc" for determining whether to skip this line. This
results in spurious "tests" that will be shown as unresolved.
|
|
If the used CMake version is recent enough to support the necessary
policy CMP0077, set INSTALL_GTEST using normal variables instead of
cache variables. We never want to allow Google Test to add to the
install target, so we should not allow a user to interfere with this.
In addition, we set BUILD_GMOCK=OFF (and the necessary BUILD_GTEST=ON
for Google Test 1.8.0), as our current tests do not require Google Mock.
|
|
As far as I can tell, the variable intended to be set here was never
called GTEST_INSTALL.
|
|
Later versions of googletest also ship googlemock alongside googletest.
Thus, the include directory we are looking for is located in a
subdirectory. The source directory, however, does not change as
googletest/CMakeLists.txt references variables set in CMakeLists.txt of
the root directory and is not intended to be included directly.
|
|
|
|
The previous version left unnecessary intermediate nodes behind, sometimes
leading to very long paths in the tree.
|
|
|
|
* move global theRNG into Executor
* pass theRNG via ctor to searchers
* remove some type warnings from RNG.cpp
Fixes #1023.
|
|
|
|
|
|
|
|
|
|
using "../"
|
|
|
|
|
|
|
|
Using KLEE's `ref<>` shared ptr requires the referenced object
to contain a reference counter to be added and initialised to 0
as part of the constructor.
To support better reuse of the `ref<>` ptr add a `ReferenceCounter`
struct.
Just adding this struct to a new class/struct as member enables
reference counting with `ref<>` - no additional counter management
needed.
|
|
reference assigned.
|
|
|
|
|
|
|
|
directory. This improves the organization of the code, and also makes it easier to reuse Expr outside KLEE.
|
|
Fixes an issue that occurs with USE_CMAKE_FIND_PACKAGE_LLVM=ON
and LLVM compiled from sources, which then exports gtest and
gtest_main targets.
In case gtest and gtest_main targets are not imported from LLVM
and GTEST_SRC_DIR is not set, CMake can now reuse the googletest
sources from LLVM_BUILD_MAIN_SRC_DIR (if available) with
USE_CMAKE_FIND_PACKAGE_LLVM=ON. This last limitation is due to
LLVM making modifications to the CMakeLists.txt of googletest
that requires add_llvm_library() from AddLLVM.cmake.
|
|
|
|
This should not change the behaviour of KLEE and mimics the old API.
- functions moved from util into time namespace
- uses time points and time spans instead of double
- CLI arguments now have the form "3h5min8us"
Changed command line parameters:
- batch-time (double to string)
- istats-write-interval (double to string)
- max-instruction-time (double to string)
- max-solver-time (double to string)
- max-time (double to string)
- min-query-time-to-log (double to string)
- seed-time (double to string)
- stats-write-interval (double to string)
- uncovered-update-interval (double to string)
- added: log-timed-out-queries (replaces negative max-solver-time)
|
|
to have only solver options.
|
|
|
|
|
|
|
|
|
|
configuration, TravisCI scripts and Dockerfile build appropriately.
There are a bunch of clean ups this enables but this commit doesn't
attempt them. We can do that in future commits.
|
|
|
|
other a regression test for #562
|
|
'const unsigned long long')
|
|
correctly applied
|
|
These were changes that I forgot to make in
dda296e09ee53ed85ccf1c3f08e7e809adce612e .
|
|
This is based off intial work by @jirislaby in #481. However it
has been substantially modified.
Notably it includes a separate build sytem to build the runtimes which
is inspired by the old build system. The reason for doing this is
because CMake is not well suited for building the runtime:
* CMake is configured to use the host compiler, not the bitcode
compiler. These are not the same thing.
* Building the runtime using `add_custom_command()` is flawed
because we can't automatically get transitive depencies (i.e.
header file dependencies) unless the CMake generator is makefiles.
(See `IMPLICIT_DEPENDS` of `add_custom_command()` in CMake).
So for now we have a very simple build system for building the runtimes.
In the future we can replace this with something more sophisticated if
we need it.
Support for all features of the old build system are implemented apart
from recording the git revision and showing it in the output of
`klee --help`.
Another notable change is the CMake build system works much better with
LLVM installs which don't ship with testing tools. The build system
will download the sources for `FileCheck` and `not` tools if the
corresponding binaries aren't available and will build them. However
`lit` (availabe via `pip install lit`) and GTest must already be
installed.
Apart from better support for testing a significant advantage of the
new CMake build system compared to the existing "Autoconf/Makefile"
build system is that it is **not** coupled to LLVM's build system
(unlike the existing build system). This means that LLVM's
autoconf/Makefiles don't need to be installed somewhere on the system.
Currently all tests pass.
Support has been implemented in TravisCI and the Dockerfile for
building with CMake.
The existing "Autoconf/Makefile" build system has been left intact
and so both build systems can coexist for a short while. We should
remove the old build system as soon as possible though because it
creates an unnecessary maintance burden.
|
|
* Add unittest to check that the `Assignment` class can evaluate
expressions containing a `NotOptimizedExpr`.
* Fix the `AssignmentTest.FoldNotOptimized` unit test by
teaching the `ExprEvaluator` to fold `NotOptimizedExpr` nodes.
|
|
|
|
|
|
be included by tools that needs to link against MetaSMT. Apart from
making the Makefile code cleaner this allowed the Solver unit test
linking to succeed when not building with STP support.
Unfortunately when using MetaSMT as the core solver the Solver unit
tests do not pass. Clearly no one tried this before... :(
|
