We are already using `import(..., required: ..)` where `required`
argument was introduced in 0.59.0, so let's update the required version
of meson necessary to build rizin.
* Add an option to provide a custom prefix for portable binaries
* Use bindir's depth to find the portable root path in rz_path_prefix
* Skip the prefix folder check if we are not in bindir
* Move binrz meson directives to binrz/ dir
* Move rizin-shell-parser to a subproject
* Move bochs to a meson subproject
* Move rzqnx to meson subproject
* Move winkd to a meson subproject
* Move rzar to a meson subproject
* Move rzw32dbg_wrap to a meson subproject
* Move ptrace-wrap to a meson subproject
* Move rzgdb to a meson subproject
* Rename w32dbg_wrap to rzw32dbg_wrap
* Update CODEOWNERS
* Remove old references to shlr
* Move shlr/heap stuff in core and remove shlr/arm
* Move spp to a meson subproject
* Remove last references to shlr
* Remove use_webui option
* Move include files directives to librz/include/meson.build
* Move librz meson directives into librz/meson.build
* Handle d/ directories inside the specific module meson.build
* Move plugins listing to specific module meson.build
* Move rzheap to its own subproject
* Fix js linter and licenses
* Use meson.override_dependency for all rz_ modules
* Remove wrong librz.pc.in file
* Add plugindir variable to .pc files to allow easy plugin configuration
* Add CMake config files to find Rizin in CMake projects
* Add CI jobs to check cmake files
* Provide workaround for mesonbuild/meson#9702
* Use .lib suffix and no prefix on MSVC to make CMake files work
* Disable cmake files on static libraries build
* Remove `RIZIN_EXTRAS` because useless
Extras plugin should be treated as any other regular plugin, they should
not have their own directory. I believe we did not have anyway any
plugin using that directory.
* Do not use `<version>` in `/usr/share`
Users can only install one version of Rizin at a time anyway, thus
splitting the `/usr/share/rizin` directories does not really bring
multi-version install possible. At the same time, we stick to what's
common for other tools, which do not use version in `/usr/share`
* Move plugin directory to `/usr/lib64/rizin/plugins`
* Remove `RZ_HOME_BIN` env var because not useful anymore
* Introduce `rz_util/rz_path` to provide rizin paths
* Use `rz_path_system_sdb_types` whenever it makes sense
* Use some others `rz_path_*` functions instead of custom ones
* Use `rz_path` APIs for themes, fortunes and signatures
* Remove `rz_sys_prefix`
* Reduce the number of defines needed to handle home and system paths
* Reduce the number of `rz_path` functions
* Make the lib debug depending on log level and `RZ_LOG_DEBUG`
* Refactor a bit how directory of plugins are loaded
* Keep supporting old plugins directory for at least 0.4.0
The new intermediate language is based on the BAP (Binary Analysis Platform) Core Theory.
It operates on SMT-like bitvectors and their arrays to represent values and memory.
Apart from that, it has also representation for the data and code side effects.
* Rewrite PDB parser to become endianess-independent
* Use type pretty-printing API
* Use `RzCmdStateOutput` instead of `RzOutputMode` to solve invalid JSON issue
* Move PDB-related print functions to `cpdb.c`
* Use flag API instead of calling commands from the code
* Separate printing from parsing and applying the type information
* Delete useless `main.c`
* Added demangling as plugins and libdemangle as external dependency
* Added iDl to return the list of demanglers
* moved demangling tests into rz-libdemangle and keep only bin tests
from the 14.x (next) release, since the snapshot update, interface is available
and allows enough flexibility thus here we re trying to dump data as much
as possible.
* `RZ_LOG_DEBUG` and `RZ_LOG_SILLY` are now only compiled when `RZ_BUILD_DEBUG` is set
* Add `RZ_BUILD_DEBUG` in `rz_userconf.h.in`
* Use rvalue to avoid unused variable error when building in release mode
- Removed oldshell version of `t` (types) commands
- Removed `tk` command because types storage switched to the hashtable
- Added `tfc` command to show the calling convention of the function or set it
- Added `tsl` (RZ_OUTPUT_MODE_LONG) to show structure members' sizes and offsets
- Added `tul` (RZ_OUTPUT_MODE_LONG) to show union members' sizes
- Switched the types storage to the hashtable instead of SDB
- Added load and export from/to SDB for standard type libraries and serialization for Projects
- Changed all `char *type` occurences to the proper `RzType`
- Added `RzCallable` type for the function/class method types
- Changed all function types API to use the `RzCallable` API
- Switched the function types (`RzCallable`) storage to the hashtable instead of SDB
- Added a connection between `RzAnalysisFunction` and `RzCallable` - `rz_analysis_function_derive_type()` API function to derive new `RzCallable` type from the `RzAnalysisFunction` even if it doesn't exist in the RzTypeDB.
- Moved type links to the RzAnalysis where they belong
- Removed TCC and MPC-based parsers
- Added Tree-Sitter-based parser for C grammar (see `librz/type/parser/*`)
- Reworked type databases in `librz/analysis/d/` to be consistent with C standard, libc, platform-specific definitions
- Switched the type links from `char *` to `RzType`
- Switched the type links from SDB to the hashtable
- Added concept of the "type paths", e.g. `a.b[20].c` where `b` is a member of `a` (`a` is either structure or a union) and `c` is a member of `b`, and `b` is an array where we take 20-th element.
- Added concept of the typeclasses (a bit similar to Haskell) for numerical types: `None` (the most generic one), `Num` (includes `Integral` and `Floating`), `Integral` (includes `Signed Integral` and `Unsigned Integral`), `Floating`.
- Added concept of type equality (`rz_types_equal(a, b)` API function)
- Added Doxygen documentation to every API function in the `librz/type/*`.
- Updated the project format from **V2** to **V3**, added migration.
* Remove the unnecessary `sdb` pointer of `RzArchTarget`
The SDB pointer in `RzArchTarget` is not needed as `sdb_load_arch_profile_by_path`
creates a new SDB pointer and takes care of our requirements.
* Introduce `RzArchPlatformTarget` and `RzArchPlatformItem` and the new APIs
Also define a new flagspace: `RZ_FLAGS_FS_PLATFORM_PORTS`
* Introduce a new configuration variable `asm.platform` to choose the
platform profile
* Define and call the APIs for platform profiles
* Add the SDB file for BCM 2835 and a unit test
* Check for NULL at various places and run `clang-format`
* Use `RZ_NONULL` for the parameters of `rz_arch_platform_init`
* Add `RZ_NONNULL` to all of the arguments which gets `rz_return[_val]_if()`-ed
* Let `platform` be NULLABLE and fix the ASAN tests
* Introduce `RzArchProfile` and add the CPU profiles
* Create `rz_arch.h` and `arch_profile.c`
* Add the CPU Profile's SDB files
* Add the `meson.build` for compiling SDB files
* Add the hashtables and the flagspaces
* Add a test and compare the `type` in `rz_sysreg_get`
* Update the list of AVR CPUs in `asm_avr`
* Run clang-format and include rz_arch.h in rz_analysis.h
* Fix the install directory of the compiled SDB files
* Introduce `RzArchTarget` which houses `RzArchProfile`
* Support for loading a new profile when `asm.cpu` is changed
* Change the variables names to lowercase and add tests
* Add a test for loading a new CPU profile when CPU is changed
* Add Doxygen documentation, run `clang-format` and correct the requested changes
* Remove unnecessary zero-initializations and other improvements
* Reversed a condition while creating the path in `rz_arch_profiles_init`
* Fixed four tests: three regarding printing the flagspaces and one from printing flags after analysis in AVR
* Free the required pointers and check for `NULL`
* Free the pointers that was allocated previously, before returning `NULL`
* Free the pointer in `rz_sysregs_db_new`
* Support for adding a fallback profile for unsupported CPUs
* RzType introduction
* Move format processing to RzTypes from RzUtil
* Rename RzType to RzTypeDB
* Join RzType and RzAnalysis noreturn functions
* Add noreturn column in `aflt` output
* Remove D (disassembly) pf tests
* Fix `tn` test
* Add function noreturn serialization
* Modify meaningless type test
* Fix wrongly written `Cf` test
* Fix `za t` (zignature type) command
* Fix signature type test
* Set default meson buildtype to debugoptimized
* Add separate Codecov build and remove linux-meson-gcc-build build
* Set test buildtypes to highest optimization level possible
* Run Codecov build only on dev branch
`portable` makes rizin use paths relative to the executable binary path,
so that all resource files (e.g. SDB files) can be retrieved without
having an hard-coded prefix in the binary. By doing so, rizin
installation directory can be moved from location to location and still
work. This is useful when having a static rizin that you want to execute
on another system.
It makes sure that all PATHs defined in rz_userconf.h are just relative
to the prefix and that, in the code, the prefix is prepended to all
these paths.
* optionally depend on lzma, bzip2 and openssl for additional zip support
* add use_sys_libzip_openssl meson option to choose if openssl should be
used in libzip or not
* avoid optional dependencies on windows to avoid troubles
* Remove tree-sitter/sdb git submodules and move them to meson subprojects/wraps
* Use meson subprojects/wraps to get capstone
* Remove capstone patches as they are already applied to rizinorg/capstone repository
* Use meson from master branch to create the release tarball, as all released versions so far have a bug that prevent wrap-git to be patched correctly when creating the tarball with meson dist
* Add check_meson_subproject script to do a quick check if the currently downloaded subprojects are updated
* Add subprojects_check meson option to skip the check_meson_subproject script
* Update documentation to remove references to git submodules.
* Ad-hoc sign rizin binary with debugger entitlements on MacOS
* Use re.rizin.rizin and move OSX pkg files in dist/osx
* Use dist/osx/build_osx_package.sh to create OSX pkg with meson
* Mention how to create OSX pkg in doc/PACKAGERS.md
* Remove doc/macos.md
* Disable colors on Windows because terminal looks bad otherwise
* Use environment on Windows to set PATH and make sure unit tests use
the compiled libraries and not the installed ones
* Define ssize_t on Windows as well by using SSIZE_T
Update test_analysis_block.c
Update test_intervaltree.c
Update test_unum.c
* Do not use setenv after fork, it's not async-signal-safe
* Do not use mutexes after fork
If pipe2 is not available on a system, we try to simulate the atomic
setting of O_CLOEXEC in different ways (depending on whether O_CLOEXEC
is at least available or not). However this involves using mutexes. When
forking, only the calling thread is forked, so it is important to lock
the mutex before the fork. However, in the child between the fork and
the exec* it is recommended to use only async-signal-safe functions. mutex
operations are not async-signal-safe, so we just keep track of whether we are
in the child or not and avoid using the mutex as necessary.
* Replace pthread_atfork with rz_sys_fork
* Simplify a bit meson.build
* Use meson summary and move fortunes install dir
* Use 0.55.0 meson version
* Use /CURRENTUSER to select local installation
By default, pipes file descriptors are not closed on exec. This means
they are inherited by the child process. However, this may be a problem
in particular in multi-threading programs (e.g. rz-test), because EOF is
not sent to a pipe until the last pipe file descriptor is closed.
This commit makes sure that by default pipes are created with O_CLOEXEC,
so they are closed whenever an exec syscall is executed. Usually pipe
file descriptors are duplicated to be stdin/stdout/stderr of the
children process and those dupped file descriptors won't be closed on
exec.
In case pipe2 is not available (e.g. MacOS), we do a pipe + fcntl, which
is in theory racy (see man open(2)). For this reason pipe, exec, system,
etc. are wrapped in rz_sys and they use a global mutex to make sure a
pipe is either created and set as O_CLOEXEC or not created at all before
an exec operation is done.
On very old systems, O_CLOEXEC may not be available at all. In those
cases, which should be rare, we just keep a list of file descriptors
created with rz_sys_pipe that we need to close before executing another
process.
Interaction with external libs may be tricky. We expect external libs to
create files/pipes that need to be closed on exec with the right flags
and we expect rizin code to manually close pipes file descriptors as
appropriate if the exec syscall is executed by an external library
(non-rizin code).
Make sure execl is found by acr/make
This is to make it easier for people to install rizin without dealing
with LD_LIBRARY_PATH, while at the same time avoiding RPATH when things
are installed in the system-wide path (for most systems, at least). In
any case, people can disable this behaviour by using `-Dlocal=disabled`.
On some platforms (e.g. FreeBSD) by default -Wl,--no-undefined is passed
to the compiler, thus when generating the util library there is an
error, because `environ` is defined only in an executable. Let's
automatically add a linker option to ignore those errors on FreeBSD.
Multiline scripts had issues on my windows system while building
from `cmd`. I removed some "pre_sdb" scripts used to read symlinks
because there are only two and it's probably easier to just avoid
symlinks for better portability.
Also moved a bunch of inlined script to separate files in sys/.
* Create RzCmdDesc generator from YAML file
The commands tree can be entirely described in a YAML file, that is
parsed at build-time and generates cmd_descs.c/cmd_descs.h files. The
developer then just needs to provide the handler with the proper name to
make sure things work.
* Add PyYAML to CI jobs
