Merge branch 'development'

.clang-format

@@ -1 +1,13 @@
 BasedOnStyle: LLVM
+SortIncludes: true
+IncludeBlocks: Regroup
+BreakStringLiterals: true
+IncludeCategories:
+  - Regex:           '^"phasar/'
+    Priority:        1
+  - Regex:           '^"(llvm|llvm-c|clang|clang-c)/'
+    Priority:        2
+  - Regex:           '".*"'
+    Priority:        3
+  - Regex:           '<[[:alnum:]_]+>'
+    Priority:        4

.clang-tidy

@@ -55,7 +55,7 @@ CheckOptions:
   - key:             readability-identifier-naming.ParameterIgnoredRegexp
     value:           (d|d1|d2|d3|d4|d5|eP|f|n)
   - key:             readability-identifier-naming.FunctionIgnoredRegexp
-    value:           (try_emplace|from_json|to_json|equal_to)
+    value:           (try_emplace|from_json|to_json|equal_to|to_string)
   - key:             cppcoreguidelines-special-member-functions.AllowSoleDefaultDtor
     value:           1
   - key:             cppcoreguidelines-special-member-functions.AllowMissingMoveFunctions

.github/CODEOWNERS

@@ -7,16 +7,18 @@
 /include/phasar/PhasarLLVM/DataflowSolver/IfdsIde/Solver/IDESolver.h @pdschubert
 
 /include/phasar/PhasarLLVM/DataflowSolver/IfdsIde/FlowEdgeFunctionCache.h @vulder
+/include/phasar/PhasarLLVM/DataflowSolver/IfdsIde/EdgeFunctionSingletonFactory.h @vulder
+/unittests/PhasarLLVM/DataFlowSolver/IfdsIde/EdgeFunctionSingletonFactoryTest.cpp @vulder
 
 /include/phasar/PhasarLLVM/DataflowSolver/Mono @pdschubert
 
 /img @pdschubert
 
-/include/phasar/PhasarLLVM/DataFlowSolver/IfdsIde/Problems/IDEInstInteractionAnalysis.h @pdschubert @vulder
-/include/phasar/PhasarLLVM/DataFlowSolver/IfdsIde/Problems/ExtendedTaintAnalysis/ @fabianbs96
-/include/phasar/PhasarLLVM/DataFlowSolver/IfdsIde/Problems/IDEExtendedTaintAnalysis.h @fabianbs96
-/lib/PhasarLLVM/DataFlowSolver/IfdsIde/Problems/IDEExtendedTaintAnalysis.cpp @fabianbs96
-/lib/PhasarLLVM/DataFlowSolver/IfdsIde/Problems/ExtendedTaintAnalysis/ @fabianbs96
+/include/phasar/PhasarLLVM/DataFlow/IfdsIde/Problems/IDEInstInteractionAnalysis.h @pdschubert @vulder
+/include/phasar/PhasarLLVM/DataFlow/IfdsIde/Problems/ExtendedTaintAnalysis/ @fabianbs96
+/include/phasar/PhasarLLVM/DataFlow/IfdsIde/Problems/IDEExtendedTaintAnalysis.h @fabianbs96
+/lib/PhasarLLVM/DataFlow/IfdsIde/Problems/IDEExtendedTaintAnalysis.cpp @fabianbs96
+/lib/PhasarLLVM/DataFlow/IfdsIde/Problems/ExtendedTaintAnalysis/ @fabianbs96
 
 /include/phasar/PhasarLLVM/AnalysisStrategy/ @pdschubert
 

.github/workflows/ci.yml

@@ -51,8 +51,10 @@ jobs:
               libclang-common-14-dev \
               libclang-14-dev \
               libclang-cpp14-dev \
-              clang-tidy-14
+              clang-tidy-14 \
+              libclang-rt-14-dev
 
+      - uses: swift-actions/setup-swift@v1
       - name: Building Phasar in ${{ matrix.build }} with ${{ matrix.compiler[0] }}
         env:
           BUILD_TYPE: ${{ matrix.build }}
@@ -65,6 +67,7 @@ jobs:
           cmake .. \
             -DCMAKE_BUILD_TYPE=$BUILD_TYPE \
             -DCMAKE_CXX_COMPILER=$CXX \
+            -DBUILD_SWIFT_TESTS=1 \
             -G Ninja
           cmake --build .
 

BreakingChanges.md

@@ -0,0 +1,43 @@
+# Breaking Changes
+
+## v0323
+
+- `EdgeFunctionPtrType` is no longer a `std::shared_ptr`. Instead `EdgeFunction<l_t>` should be used directly. `EdgeFunction` is now a *value-type* that encapsulates its memory management by itself.
+- Concrete `EdgeFunction` types no longer derive from any base-class. Instead they just need to implement the required API functions. `EdgeFunction` implementations should me move-constructible and can be implicitly cast to `EdgeFunction`. To verify that your type implements the edge function interface use the `IsEdgeFunction` type trait. The API functions have been changed as follows:
+    - All API functions of `EdgeFunction` must be `const` qualified.
+    - `EdgeFunctionPtrType composeWith(EdgeFunctionPtrType SecondFunction)` and `EdgeFunctionPtrType joinWith(EdgeFunctionPtrType OtherFunction)` have been changed to `static EdgeFunction<l_t> compose(EdgeFunctionRef<T> This, const EdgeFunction<l_t>& SecondFunction)` and `static EdgeFunction<l_t> join(EdgeFunctionRef<T> This, const EdgeFunction<l_t>& OtherFunction)` respectively. Here, the `This` parameter models the former `shared_from_this()`.
+    - `bool equal_to(EdgeFunctionPtrType Other)const` has been changed to `bool operator==(const T &Other)const noexcept`, where `T` is your concrete edge function type.
+    - `void print(llvm::raw_ostream &OS, bool IsForDebug)` has been changed to `friend llvm::raw_ostream &operator<<(llvm::raw_ostream &OS, const T& EF)`.
+- `EdgeFunction` is tagged with `[[clang::trivial_abi]]`. Hence, you should not rely on any destruction order within a top-level statement that uses temporary `EdgeFunction` objects.
+- `EdgeFunctionSingletonFactory` has been removed. Use `EdgeFunctionSingletonCache` instead.
+- `TaintConfig` has been renamed to `LLVMTaintConfig`. For generic code you may want to use the LLVM-independent `TaintConfigBase` CRTP interface instead.
+- Renamed `phasar/PhasarLLVM/DataFlowSolver/` to either `phasar/DataFlow/` or `phasar/PhasarLLVM/DataFlow/` depending on whether the components need LLVMCore. Analoguous changes in `lib/` and `unittests/`.
+    An incomplete list of moved/renamed files:
+    - `phasar/PhasarLLVM/DataFlowSolver/IfdsIde/Solver/*` => `phasar/DataFlow/IfdsIde/Solver/*`
+    - `phasar/PhasarLLVM/DataFlowSolver/IfdsIde/IDETabulationProblem.h` => `phasar/DataFlow/IfdsIde/IDETabulationProblem.h`
+    - `phasar/DB/LLVMProjectIRDB.h` => `phasar/PhasarLLVM/DB/LLVMProjectIRDB.h`
+    - ...
+- Renamed and split up some libraries:
+    - `phasar_phasarllvm_utils` => `phasar_llvm_utils`
+    - `phasar_typehierarchy` => `phasar_llvm_typehierarchy`
+    - `phasar_ifdside` => `phasar_llvm_ifdside`
+    - `phasar_controlflow` has its LLVM dependent stuff moved to `phasar_llvm_controlflow`
+    - `phasar_db` has its LLVM dependent stuff moved to `phasar_llvm_db`
+    - `phasar_pointer` has its LLVM dependent stuff moved to `phasar_llvm_pointer`
+- Renamed the phasar tool `phasar-llvm` to `phasar-cli`
+- `LLVMPointsTo[.*]` has been renamed to `LLVMAlias[.*]`
+- The ctor of `LLVMAliasSet` now takes the `LLVMProjectIRDB` as pointer instead of a reference to better document that it may capture the IRDB by reference.
+- The `PointsToInfo` interface has been replaced by the CRTP interface `AliasInfoBase`. Introduced two type-erased implementations of that interface: `AliasInfo` and `AliasInfoRef`. In most cases you should replace `PointsToInfo *` and `LLVMPointsToInfo *` by `AliasInfoRef`, bzw. `LLVMAliasInfoRef`.
+- Introduced a new interface `PointsToInfoBase` and type-erased implementations `PointstoInfo` and `PointsToInfoRef`. Don't confuse them with the old `PointsToInfo`!!! (However, they have different APIs, so you should encounter compilation errors then)
+
+## v1222
+
+- Removed virtual interfaces `CFG<N,F>` and `ICFG<N,F>` and replaced by CRTP interfaces `CFGBase` and `ICFGBase`. Use the concrete types `LLVMBasedICFG` and `LLVMBasedCFG` instead. In template code you can use the type traits `is_crtp_base_of_v` and `is_icfg_v` to check for conformance to the interfaces.
+- The `LLVMBasedICFG` now takes the IRDB as pointer instead of a reference to better document that it may capture the IRDB by reference.
+- Renamed `ProjectIRDB` to `LLVMProjectIRDB` and added a generic CRTP interface `ProjectIRDBBase` that does not depend on LLVM
+- Changed the API of `LLVMProjectIRDB`: The IRDB does no longer link multiple LLVM modules together, i.e. the ctor that reads a module from a file now takes a single filename instead of a vector. If you still want to link multiple LLVM modules together, use LLVM's Linker functionality instead. `ProjecIRDB::getAllModules()` has been removed and `ProjectIRDB::getWPAModule()` has been renamed to `LLVMProjectIRDB::getModule()`.
+- The `LLVMProjectIRDB` ctor that takes a raw-pointer to `llvm::Module` does no longer preprocess the module (i.e. attaching metadata IDs to it). You can still explicitly use the `ValueAnnotationPass` to do so manually.
+- The type `WholeProgramAnalysis` has been removed. Use `AnalysisController` instead.
+- The IFDS and IDE TabulationProblems no longer take all of `LLVMProjectIRDB*`, `LLVMTypeHierarchy*`, `LLVMPointsToInfo*` and `LLVMBasedICFG*` as an argument. Instead, they only get what they need.
+- The `IFDSSolver` and `IDESolver` now take an instance of the `ICFGBase` interface as additional argument to their ctor (because the analysis problems now not necessarily store a reference to it anymore).
+- The `IDETabulationProblem` is now a base class of `IFDSTabulationProblem` (and not vice versa as it was previously). In their ctors they only take the bare minimum of arguments: The IRDB, the entrypoints and optionally the special zero-value. If the zero-value is not passed in the ctor (as it was previously), it has to be set from within the client analysis' ctor. You may use the new function `initializeZeroValue(d_t)` for this.

CMakeLists.txt

@@ -53,6 +53,13 @@ include("phasar_macros")
 
 option(PHASAR_BUILD_UNITTESTS "Build all tests (default is ON)" ON)
 
+option(BUILD_SWIFT_TESTS "Builds the Swift tests (Swift compiler has to be installed manually beforehand!)" OFF)
+if (BUILD_SWIFT_TESTS)
+  set(CMAKE_Swift_FLAGS_RELEASE "-g")
+  set(CMAKE_Swift_FLAGS_RELWITHDEBINFO "-g")
+  enable_language(Swift)
+endif(BUILD_SWIFT_TESTS)
+
 option(PHASAR_BUILD_OPENSSL_TS_UNITTESTS "Build OPENSSL typestate tests (require OpenSSL, default is OFF)" OFF)
 
 option(PHASAR_BUILD_IR "Build IR test code (default is ON)" ON)
@@ -107,10 +114,14 @@ include_directories(
   ${PHASAR_SRC_DIR}/include
 )
 
-set(PHASAR_INSTALL_LIBDIR ${CMAKE_INSTALL_LIBDIR}/phasar)
-
 if (NOT PHASAR_IN_TREE)
-  set(CMAKE_INSTALL_RPATH ${CMAKE_INSTALL_PREFIX}/${PHASAR_INSTALL_LIBDIR})
+  set(CMAKE_INSTALL_RPATH ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR})
+
+  if (NOT "${CMAKE_INSTALL_LIBDIR}" STREQUAL "lib")
+    message(STATUS "Detected CMAKE_INSTALL_LIBDIR that deviates from 'lib': ${CMAKE_INSTALL_LIBDIR}. Add ${CMAKE_INSTALL_PREFIX}/lib to the RPATH as json-schema-validator needs it")
+    list(APPEND CMAKE_INSTALL_RPATH ${CMAKE_INSTALL_PREFIX}/lib)
+  endif()
+
   set(CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
 endif()
 
@@ -120,6 +131,17 @@ else()
   set(PHASAR_STD_FILESYSTEM stdc++fs)
 endif()
 
+set(PHASAR_CUSTOM_CONFIG_INSTALL_DIR "" CACHE STRING "If set, customizes the directory, where configuration files for PhASAR are installed (default is /usr/local/.phasar-config)")
+if ("${PHASAR_CUSTOM_CONFIG_INSTALL_DIR}" STREQUAL "")
+  set(PHASAR_CONFIG_INSTALL_DIR "${CMAKE_INSTALL_PREFIX}/.phasar-config/")
+else()
+  set(PHASAR_CONFIG_INSTALL_DIR "${PHASAR_CUSTOM_CONFIG_INSTALL_DIR}")
+endif()
+
+add_compile_definitions(PHASAR_CONFIG_DIR="${PHASAR_CONFIG_INSTALL_DIR}")
+add_compile_definitions(PHASAR_DIR="${PHASAR_SRC_DIR}")
+
+
 ### Adding external libraries
 # Threads
 find_package(Threads)
@@ -144,7 +166,7 @@ set(CMAKE_CXX_CLANG_TIDY "")
 set(BUILD_TESTS OFF CACHE BOOL "Build json-schema-validator-tests")
 
 # Make nlohmann_json_schema_validator happy by telling it how to find the single include of nlohmann_json
-include_directories(external/json/single_include/)
+include_directories(SYSTEM external/json/single_include/)
 
 if (PHASAR_IN_TREE)
   set_property(GLOBAL APPEND PROPERTY LLVM_EXPORTS nlohmann_json_schema_validator)
@@ -301,7 +323,7 @@ endif()
 
 set(INCLUDE_INSTALL_DIR include/ CACHE PATH "Install dir of headers")
 
-# Install targets of phasar-llvm, other executables, and libraries are to be
+# Install targets of phasar-cli, other executables, and libraries are to be
 # found in the individual subdirectories of tools/
 
 # Install Phasar include directory
@@ -326,15 +348,17 @@ install(DIRECTORY external/googletest/googletest/include/gtest/
 # Install Phasar utils helper scripts
 install(DIRECTORY utils/
   DESTINATION bin
-  FILES_MATCHING PATTERN "phasar-*"
+  FILES_MATCHING
+  PATTERN "CodeGen" EXCLUDE # CodeGen does not contain files to install
+  PATTERN "phasar-*"
   PERMISSIONS OWNER_EXECUTE OWNER_WRITE OWNER_READ
               GROUP_EXECUTE GROUP_READ
               WORLD_EXECUTE WORLD_READ
 )
 
-# Install the Phasar config files into ~/.config/phasar/
+# Install the Phasar config files into CMAKE_INSTALL_PREFIX/.phasar-config/
 install(DIRECTORY config/
-  DESTINATION $ENV{HOME}/.config/phasar
+  DESTINATION ${PHASAR_CONFIG_INSTALL_DIR}
   PATTERN "config/*"
   PERMISSIONS OWNER_WRITE OWNER_READ
               GROUP_WRITE GROUP_READ

Config.cmake.in

@@ -12,15 +12,24 @@ find_package(Boost 1.65.1 COMPONENTS program_options graph REQUIRED)
 
 set(PHASAR_COMPONENTS
   utils
-  config
-  phasarllvm_utils
   passes
+  config
   db
   pointer
-  typehierarchy
   controlflow
+
+  llvm_utils
+  llvm_db
+  llvm_pointer
+  llvm_typehierarchy
+  llvm_controlflow
+
   taintconfig
-  ifdside
+  mono
+  llvm
+  llvm_ifdside
+  analysis_strategy
+  controller
 )
 
 foreach(component ${PHASAR_COMPONENTS})

Dockerfile

@@ -33,12 +33,11 @@ RUN apt-key adv --fetch-keys https://apt.llvm.org/llvm-snapshot.gpg.key && \
               libclang-common-14-dev \
               libclang-14-dev \
               libclang-cpp14-dev \
-              clang-tidy-14
+              clang-tidy-14 \
+              libclang-rt-14-dev
 
 RUN pip3 install Pygments pyyaml
 
-# installing boost
-RUN apt install libboost-all-dev -y
 
 
 # installing wllvm
@@ -58,4 +57,4 @@ RUN mkdir -p build && cd build && \
             -G Ninja && \
           cmake --build .
 
-ENTRYPOINT [ "./build/tools/phasar-llvm/phasar-llvm" ]
+ENTRYPOINT [ "./build/tools/phasar-cli/phasar-cli" ]

README.md

@@ -6,12 +6,12 @@ PhASAR a LLVM-based Static Analysis Framework
 [![C++ Standard](https://img.shields.io/badge/C++_Standard-C%2B%2B17-blue.svg?style=flat&logo=c%2B%2B)](https://isocpp.org/)
 [![GitHub license](https://img.shields.io/badge/license-MIT-blueviolet.svg)](https://raw.githubusercontent.com/secure-software-engineering/phasar/master/LICENSE.txt)
 
-Version 1222
+Version 0323
 
 Secure Software Engineering Group
 ---------------------------------
 
-+ Philipp Dominik Schubert (philipp.schubert@upb.de) and others
++ Fabian Schiebel (fabian.schiebel@iem.fraunhofer.de), Martin Mory (martin.mory@upb.de), Philipp Dominik Schubert (philipp.schubert@upb.de) and others
 + Please also refer to https://phasar.org/
 
 Required version of the C++ standard
@@ -30,9 +30,13 @@ fully-automated manner on the specified LLVM IR target code. Computing points-to
 information, call-graph(s), etc. is done by the framework, thus you can focus on
 what matters.
 
+Breaking Changes
+----------------
+To keep PhASAR in a state that it is well suited for state-of-the-art research in static analysis, as well as for productive use, we have to make breaking changes. Please refer to [Breaking Changes](./BreakingChanges.md) for detailed information on what was broken recently and how to migrate.
+
 How do I get started with PhASAR?
 ---------------------------------
-We have some documentation on PhASAR in our [_**wiki**_](https://github.com/secure-software-engineering/phasar/wiki). You probably would like to read
+We have some documentation on PhASAR in our [_**Wiki**_](https://github.com/secure-software-engineering/phasar/wiki). You probably would like to read
 this README first and then have a look on the material provided on https://phasar.org/
 as well. Please also have a look on PhASAR's project directory and notice the project directory
 examples/ as well as the custom tool tools/myphasartool.cpp.
@@ -45,6 +49,21 @@ prerequisites and compilation. It is recommended to compile PhASAR yourself in
 order to get the full C++ experience and to have full control over the build
 mode.
 
+As a shortcut for the very first PhASAR build on your system, you can use our [bootstrap](./bootstrap.sh) script:
+
+```bash
+$ ./bootstrap.sh
+```
+
+Note: If you want to do changes within PhASAR, it is recommended to build it in Debug mode:
+```bash
+$ ./bootstrap.sh -DCMAKE_BUILD_TYPE=Debug
+```
+
+The bootstrap script may ask for superuser permissions (to install the dependencies); however it is not recommended to start the whole script with `sudo`.
+
+For subsequent builds, see [Compiling PhASAR](#compiling-phasar-if-not-already-done-using-the-installation-scripts).
+
 Please help us to improve PhASAR
 --------------------------------
 You are using PhASAR and would like to help us in the future? Then please
@@ -65,10 +84,12 @@ PhASAR using an Ubuntu or Unix-like system.
 
 Therefore, we provide an installation script. To install PhASAR, just navigate to the top-level
 directory of PhASAR and use the following command:
-```
-$ sudo ./bootstrap.sh
+```bash
+$ ./bootstrap.sh --install
 ```
 
+The bootstrap script may ask for superuser permissions.
+
 Done!
 
 
@@ -86,8 +107,7 @@ $ export CXX=/usr/local/bin/clang++
 You may need to adjust the paths according to your system. When you cloned PhASAR from Github you need to initialize PhASAR's submodules before building it:
 
 ```
-$ git submodule init
-$ git submodule update
+$ git submodule update --init
 ```
 
 If you downloaded PhASAR as a compressed release (e.g. .zip or .tar.gz) you can use the `init-submodules-release.sh` script that manually clones the required submodules:
@@ -101,27 +121,25 @@ Navigate into the PhASAR directory. The following commands will do the job and c
 ```
 $ mkdir build
 $ cd build/
-$ cmake -DCMAKE_BUILD_TYPE=Release ..
-$ make -j $(nproc) # or use a different number of cores to compile it
-$ sudo make install # if you wish to install PhASAR system wide
+$ cmake -G Ninja -DCMAKE_BUILD_TYPE=Release ..
+$ ninja -j $(nproc) # or use a different number of cores to compile it
+$ sudo ninja install # only if you wish to install PhASAR system wide
 ```
 
 When you have used the `bootstrap.sh` script to install PhASAR, the above steps are already done.
 Use them as a reference if you wish to modify PhASAR and recompile it.
 
 After compilation using cmake the following two binaries can be found in the build/ directory:
 
-+ phasar-llvm - the actual PhASAR command-line tool
++ phasar-cli - the actual PhASAR command-line tool (previously called `phasar-llvm`)
 + myphasartool - an example tool that shows how tools can be build on top of PhASAR
 
 Use the command:
 
-`$ ./phasar-llvm --help`
+`$ ./phasar-cli --help`
 
 in order to display the manual and help message.
 
-`$ sudo make install`
-
 Please be careful and check if errors occur during the compilation.
 
 When using CMake to compile PhASAR the following optional parameters can be used:
@@ -130,7 +148,8 @@ When using CMake to compile PhASAR the following optional parameters can be used
 |-----------|--------|
 | <b>BUILD_SHARED_LIBS</b> : BOOL | Build shared libraries (default is ON) |
 | <b>CMAKE_BUILD_TYPE</b> : STRING | Build PhASAR in 'Debug' or 'Release' mode <br> (default is 'Debug') |
-| <b>CMAKE_INSTALL_PREFIX</b> : PATH | Path where PhASAR will be installed if <br> “make install” is invoked or the “install” <br> target is built (default is /usr/local/phasar) |
+| <b>CMAKE_INSTALL_PREFIX</b> : PATH | Path where PhASAR will be installed if <br> "ninja install” is invoked or the “install” <br> target is built (default is /usr/local/phasar) |
+| <b>PHASAR_CONFIG_INSTALL_PREFIX</b> : PATH | Path where PhASAR's configuration files will be installed if <br> ninja install” is invoked or the “install” <br> target is built. Expression will be evaluated within CMAKE_INSTALL_PREFIX, so prefer absolute paths (default is $(HOME)/.config/) |
 | <b>PHASAR_BUILD_DOC</b> : BOOL | Build PhASAR documentation (default is OFF) |
 | <b>PHASAR_BUILD_UNITTESTS</b> : BOOL | Build PhASAR unit tests (default is ON) |
 | <b>PHASAR_BUILD_IR</b> : BOOL | Build PhASAR IR (required for running the unit tests) (default is ON) |
@@ -148,7 +167,7 @@ C++'s long compile times are always a pain. As shown in the above, when using cm
 ### Running a test solver
 To test if everything works as expected please run the following command:
 
-`$ phasar-llvm --module test/build_systems_tests/installation_tests/module.ll -D ifds-solvertest`
+`$ phasar-cli -m test/build_systems_tests/installation_tests/module.ll -D ifds-solvertest`
 
 If you obtain output other than a segmentation fault or an exception terminating the program abnormally everything works as expected.