liburlparser 2.1.0__tar.gz → 2.2.0__tar.gz

liburlparser-2.2.0/CHANGELOG.md

@@ -0,0 +1,212 @@
+# Changelog
+
+All notable changes to liburlparser are documented here. The format follows
+[Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and the project follows
+[Semantic Versioning](https://semver.org/). The version of the project is the one in `include/urlparser.h`
+(`URLPARSER_VERSION_MAJOR/MINOR/PATCH`); `pyproject.toml` and CMake read it from there.
+
+The entries up to 2.1.0 were written from the git history and the release tags. Since the commit messages of the 1.x
+releases are short, those entries are brief.
+
+## [2.2.0] - 2026-11-09
+
+### Changed
+- **The CMake package namespace is now `liburlparser::`, like the package.** The public target is
+  `liburlparser::urlparser`, the same name from `add_subdirectory()`, `FetchContent` and
+  `find_package(liburlparser CONFIG)`. A CMake package should export its targets under its own name, and the
+  package is `liburlparser` (`liburlparserConfig.cmake`, `python -m liburlparser --cmake-dir`). **Breaking:** the
+  2.1.0 name `urlparser::urlparser` is gone and there is no alias for it. Change `target_link_libraries(... urlparser::urlparser)` to
+  `liburlparser::urlparser`. The C++ namespace (`urlparser::url`, `urlparser::host`) is unchanged.
+- The `example` program is built only when liburlparser is the top-level project, like the tests and the docs, so it is
+  not compiled for nothing when the library is a dependency.
+
+### Fixed
+- **Using liburlparser from another scikit-build-core project failed.** `SKBUILD` is defined for the whole build, so a
+  project that pulled liburlparser in with `add_subdirectory()` or `FetchContent` made it believe it was itself the wheel
+  being built: it tried to build its own nanobind module (and failed when nanobind was not installed) or stopped
+  with `RETURN()` before the example, tests and docs. The check is now "`SKBUILD` is defined **and** liburlparser is the
+  top-level project". The 2.0.0 release had a guard for this (`c140ea4`); the CMake refactor of 2.1.0 (`ac6ac03`)
+  dropped it.
+
+## [2.1.0] - 2026-10-03
+
+### Added
+- **`find_package(liburlparser CONFIG)`.** The core is a static library that is installed with its headers and a CMake
+  package (`liburlparserConfig.cmake`, targets, version file, and `liburlparserPaths.cmake` with the absolute library
+  and include paths). The Python wheel carries the same tree (`include/`, `lib/`, `cmake/`), and
+  `python -m liburlparser --cmake-dir` / `--include-dir` / `--lib-dir` print the paths, the way `python -m nanobind --cmake_dir` does.
+  `examples/find_liburlparser_via_python` shows it.
+- **IDNA normalization and percent codec APIs.**
+- **scp-like git URLs** (`user@host:path/repo.git`) are parsed.
+- **`url::normalized()`** returns a canonical serialization of a URL.
+- Benchmarks: `search_params` against ada, a reproducible comparison of the PSL structures, benchmark artifacts kept apart
+  from the committed results, and an HTML report target.
+
+### Changed
+- **The CMake targets are unified.** The historical aliases `url::base` and `liburlparser::core` are replaced by the one
+  public alias `urlparser::urlparser`, and `install(EXPORT)` uses the same name. (**Breaking** for anyone who linked the old aliases;
+  renamed again in 2.2.0.)
+- The standalone `idna.cpp` and `percent_codec.cpp` are merged into `urlparser.cpp`, and stale files left from the
+  header/source consolidation were removed.
+- The bundled `public_suffix_list.dat` is the 2026-09-08 version.
+- CI: cibuildwheel is updated.
+
+### Performance
+- AVX2 scanning for the authority of a URL.
+- PSL suffix matching tracks the maximum rule depth, and the truncation scan in `suffix_of` and the TLD suffix are bounded
+  by the maximum length (an O(1) check instead of an O(n) scan).
+- A fast path with interned keys when a `Hostname` is extracted to a Python dict.
+
+## [2.0.0] - 2026-09-11
+
+### Added
+- **Public Suffix List support.** A `psl` class (a singleton that tracks the version of its data) matches suffixes; the
+  list is embedded into the library when CMake configures, and a configure step refreshes the tracked
+  `public_suffix_list.dat` in place when it can download a newer one.
+- **`host`** parsing and classification, with constructor overloads and `string&&` overloads of `host::from_url` and
+  `url::extract_host`. Cross-platform IP address parsing replaces the platform-specific code.
+- **Version macros** `URLPARSER_VERSION_*` and a version class (the prefix was renamed from `LIBURLPARSER_`); CMake reads the version from the
+  header.
+- SIMD scanning in `find_first_of_3` (AVX2, SSE2, NEON).
+- Python: a static `Url.extract_dict`, built from reusable `to_dict` helpers.
+- Tooling: coverage reports for C++ and Python, leak checks with pytest-memray, benchmark and profiling targets, standalone
+  benchmarks (also against ada), Sphinx documentation for the Python API, a CLI reference, and architecture notes.
+
+### Changed
+- **liburlparser is a static library**, built with position-independent code, and the default build type is Release.
+- nanobind ABI handling for Python 3.12 and later.
+- The googletest and doxygen-awesome-css submodules are removed, and the `online` extra is gone (the PSL update uses `urllib`).
+- Tests and docs are built only when liburlparser is the top-level project, and a vendored copy does not build its Python
+  bindings.
+- The implementation of the suffix table is hidden behind a forward declaration.
+
+### Fixed
+- Windows compatibility (pull request #4, by CKCat).
+- Regression tests for earlier URL parsing bugs.
+
+### Performance
+- Extra capacity for URL fields, a faster `host` constructor, and the suffix position is computed by arithmetic.
+
+## [1.6.1] - 2025-10-24
+
+### Changed
+- The CLI logic is refactored and covered by tests, and `flatten_dict` moved to the utils module.
+- The release workflow and the wheel builds are improved (`test_cli_subprocess` is excluded in CI).
+
+## [1.6.0] - 2025-05-04
+
+### Added
+- `--parts` option of the command line, and the top-level domain (TLD) of a host.
+- The Public Suffix List file.
+- Doxygen documentation (main page with logo and architecture diagram, doxygen-awesome theme) and the Python documentation,
+  published to GitHub Pages.
+
+## [1.5.0] - 2024-10-18
+
+### Changed
+- The project architecture changed, and the CI runs all tests (C++ and Python).
+- Build requirements are managed with pip-tools; ruff fixes.
+
+## [1.4.6] - 2024-02-11
+
+### Changed
+- macOS is built in CI, the build system changed, and the README got badges.
+
+## [1.4.5] - 2023-11-28
+
+### Changed
+- The Doxygen project information is read from `pyproject.toml`, and the release workflow uploads the release assets.
+
+## [1.4.4] - 2023-11-23
+
+### Changed
+- The C++ tests moved to googletest and were rewritten as parametrized tests, like the Python ones; C++ and Python
+  tests are separate. The wheel files are created.
+
+## [1.4.3] - 2023-11-19
+
+### Fixed
+- A file lock lets the tests run in parallel.
+
+## [1.4.2] - 2023-11-13
+
+### Changed
+- `shared_ptr` replaces `unique_ptr`, since every method is const.
+
+### Fixed
+- The `fulldomain` problem, and the host name was removed again.
+
+## [1.4.1] - 2023-11-11
+
+### Added
+- C++ tests with gtest and gmock.
+- A fast `hostName` that does not parse the host string (with a known problem with `ignore_www`).
+
+### Changed
+- The PSL is held as data instead of a pointer (fixes the pointer-to-data conversion), and `requests` became an optional
+  dependency (the `online` extra).
+
+## [1.4.0] - 2023-11-08
+
+### Changed
+- `urlparser.cpp` is split into a host part and a url part, a url holds its host through a pointer, and the lazy
+  calculation uses a smart pointer.
+
+### Fixed
+- Problems with the operators; the move constructor and assignment were removed.
+
+## [1.3.0] - 2023-11-05
+
+### Fixed
+- `fulldomain` in the substring step; accessibility of the class; test and linter issues. An automatic correction was
+  tried and removed.
+
+## [1.2.1] - 2023-10-10
+
+### Fixed
+- The `fulldomain` issue.
+
+## [1.2.0] - 2023-10-09
+
+### Changed
+- The `fulldomain` attribute was updated.
+
+## [1.1.0] - 2023-07-23
+
+### Added
+- The `ignore_www` option for subdomains.
+
+## [1.0.2] - 2023-07-22
+
+### Added
+- A Ruff lint check.
+
+## [1.0.1] - 2023-07-21
+
+### Fixed
+- `pyproject.toml` and the wheels workflow.
+
+## [1.0.0] - 2023-07-21
+
+First tagged release. The history before it is mostly the setup of the wheel builds with cibuildwheel (PyPy excluded).
+
+[2.2.0]: https://github.com/mohammadraziei/liburlparser/compare/v2.1.0...v2.2.0
+[2.1.0]: https://github.com/mohammadraziei/liburlparser/compare/v2.0.0...v2.1.0
+[2.0.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.6.1...v2.0.0
+[1.6.1]: https://github.com/mohammadraziei/liburlparser/compare/v1.6.0...v1.6.1
+[1.6.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.5.0...v1.6.0
+[1.5.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.6...v1.5.0
+[1.4.6]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.5...v1.4.6
+[1.4.5]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.4...v1.4.5
+[1.4.4]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.3...v1.4.4
+[1.4.3]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.2...v1.4.3
+[1.4.2]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.1...v1.4.2
+[1.4.1]: https://github.com/mohammadraziei/liburlparser/compare/v1.4.0...v1.4.1
+[1.4.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.3.0...v1.4.0
+[1.3.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.2.1...v1.3.0
+[1.2.1]: https://github.com/mohammadraziei/liburlparser/compare/v1.2.0...v1.2.1
+[1.2.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.1.0...v1.2.0
+[1.1.0]: https://github.com/mohammadraziei/liburlparser/compare/v1.0.2...v1.1.0
+[1.0.2]: https://github.com/mohammadraziei/liburlparser/compare/v1.0.1...v1.0.2
+[1.0.1]: https://github.com/mohammadraziei/liburlparser/compare/v1.0.0...v1.0.1
+[1.0.0]: https://github.com/mohammadraziei/liburlparser/releases/tag/v1.0.0

{liburlparser-2.1.0 → liburlparser-2.2.0}/CMakeLists.txt

@@ -38,9 +38,20 @@ if (CYGWIN OR UNIX)
 endif()
 
 
+# URLPARSER_SKBUILD is true only when liburlparser itself is the scikit-build-core
+# project being built (the wheel build). SKBUILD alone is not enough: it is also
+# defined when another scikit-build-core project (e.g. httpp) pulls this one in with
+# add_subdirectory() or FetchContent, and then it must behave as a plain C++ library:
+# no nanobind module, no early RETURN() before the liburlparser::urlparser target and the tests.
+if(DEFINED SKBUILD AND CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
+    set(URLPARSER_SKBUILD ON)
+else()
+    set(URLPARSER_SKBUILD OFF)
+endif()
+
 find_package(Python3 REQUIRED COMPONENTS Interpreter)
 
-if(DEFINED SKBUILD)
+if(URLPARSER_SKBUILD)
     set(PYTHON_PROJECT_NAME "${SKBUILD_PROJECT_NAME}")
 elseif(BUILD_PYTHON)
     set(PYTHON_PROJECT_NAME "${CMAKE_BINARY_DIR}")
@@ -102,7 +113,7 @@ embed_file_as_string(
 # Install layout: the wheel carries include/, lib/ and cmake/ inside the
 # `liburlparser` package (nanobind/pybind11 pattern, see get_cmake_dir() in
 # the Python package); a plain `cmake --install` uses the usual prefix layout.
-if(DEFINED SKBUILD)
+if(URLPARSER_SKBUILD)
     set(_urlparser_lib_dest   liburlparser/lib)
     set(_urlparser_inc_dest   liburlparser/include)
     set(_urlparser_cmake_dest liburlparser/cmake)
@@ -113,16 +124,16 @@ else()
 endif()
 
 add_library(urlparser STATIC ${SOURCES})
-# The public target is `urlparser::urlparser` - the same name whether the
+# The public target is `liburlparser::urlparser` - the same name whether the
 # library comes from FetchContent/add_subdirectory (this alias) or from an
 # installed package (find_package(liburlparser), exported below).
-add_library(urlparser::urlparser ALIAS urlparser)
+add_library(liburlparser::urlparser ALIAS urlparser)
 
 set_target_properties(urlparser PROPERTIES
     POSITION_INDEPENDENT_CODE ON        # linked into the Python extension
     CXX_VISIBILITY_PRESET hidden
     VISIBILITY_INLINES_HIDDEN ON
-    EXPORT_NAME urlparser               # installed as urlparser::urlparser (see install(EXPORT))
+    EXPORT_NAME urlparser               # installed as liburlparser::urlparser (see install(EXPORT))
     OUTPUT_NAME urlparser               # liburlparser.a / urlparser.lib
 )
 target_include_directories(urlparser
@@ -175,7 +186,7 @@ if(DEFINED PYTHON_PROJECT_NAME)
 
     target_link_libraries(${NB_MODULE} PRIVATE urlparser)
 
-    if(DEFINED SKBUILD)
+    if(URLPARSER_SKBUILD)
         install(TARGETS ${NB_MODULE}
                 CONFIGURATIONS Release
                 LIBRARY DESTINATION
@@ -187,7 +198,7 @@ if(DEFINED PYTHON_PROJECT_NAME)
     endif()
 endif()
 
-# --- install: find_package(liburlparser CONFIG) -> urlparser::urlparser ---
+# --- install: find_package(liburlparser CONFIG) -> liburlparser::urlparser ---
 include(CMakePackageConfigHelpers)
 install(TARGETS urlparser EXPORT liburlparserTargets
     ARCHIVE DESTINATION ${_urlparser_lib_dest}
@@ -195,7 +206,7 @@ install(TARGETS urlparser EXPORT liburlparserTargets
 install(DIRECTORY include/ DESTINATION ${_urlparser_inc_dest})
 install(EXPORT liburlparserTargets
     FILE liburlparserTargets.cmake
-    NAMESPACE urlparser::
+    NAMESPACE liburlparser::
     DESTINATION ${_urlparser_cmake_dest})
 
 # urlparser_LIB_PATH / urlparser_INCLUDE_PATH for consumers that want plain
@@ -228,12 +239,15 @@ install(FILES
     ${PROJECT_BINARY_DIR}/liburlparserConfigVersion.cmake
     DESTINATION ${_urlparser_cmake_dest})
 
-if(DEFINED SKBUILD)
+if(URLPARSER_SKBUILD)
     RETURN()
 endif()
 
-add_executable(example examples/main.cpp)
-target_link_libraries(example PRIVATE urlparser::urlparser)
+# The example is only for the top-level build; as a dependency it would just be compiled for nothing.
+if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
+    add_executable(example examples/main.cpp)
+    target_link_libraries(example PRIVATE liburlparser::urlparser)
+endif()
 
 enable_testing()
 find_package(Threads REQUIRED)

{liburlparser-2.1.0 → liburlparser-2.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: liburlparser
-Version: 2.1.0
+Version: 2.2.0
 Summary: Fastest Url parser in the world
 Author-Email: Mohammad Raziei <mohammadraziei1375@gmail.com>
 Classifier: Development Status :: 4 - Beta
@@ -243,7 +243,7 @@ you can see all methods in python we can use in c++ very easily
 ### C++
 
 Requires a C++17 compiler and CMake >= 3.19. The library is a static library plus one header (`urlparser.h`),
-and its CMake target is **`urlparser::urlparser`** - the same name no matter how you get it:
+and its CMake target is **`liburlparser::urlparser`** - the same name no matter how you get it:
 
 #### 1. FetchContent
 
@@ -256,7 +256,7 @@ FetchContent_Declare(
 )
 FetchContent_MakeAvailable(liburlparser)
 
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 #### 2. From the pip package
@@ -274,7 +274,7 @@ execute_process(
     COMMAND_ERROR_IS_FATAL ANY)
 
 find_package(liburlparser CONFIG REQUIRED)
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 For build systems that prefer plain paths over the target, `find_package` also sets:
@@ -321,7 +321,7 @@ Then, in your project (add `-DCMAKE_PREFIX_PATH=<prefix>` if you used a custom p
 
 ```cmake
 find_package(liburlparser CONFIG REQUIRED)
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 ### Python and Command Line:

{liburlparser-2.1.0 → liburlparser-2.2.0}/README.md

@@ -219,7 +219,7 @@ you can see all methods in python we can use in c++ very easily
 ### C++
 
 Requires a C++17 compiler and CMake >= 3.19. The library is a static library plus one header (`urlparser.h`),
-and its CMake target is **`urlparser::urlparser`** - the same name no matter how you get it:
+and its CMake target is **`liburlparser::urlparser`** - the same name no matter how you get it:
 
 #### 1. FetchContent
 
@@ -232,7 +232,7 @@ FetchContent_Declare(
 )
 FetchContent_MakeAvailable(liburlparser)
 
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 #### 2. From the pip package
@@ -250,7 +250,7 @@ execute_process(
     COMMAND_ERROR_IS_FATAL ANY)
 
 find_package(liburlparser CONFIG REQUIRED)
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 For build systems that prefer plain paths over the target, `find_package` also sets:
@@ -297,7 +297,7 @@ Then, in your project (add `-DCMAKE_PREFIX_PATH=<prefix>` if you used a custom p
 
 ```cmake
 find_package(liburlparser CONFIG REQUIRED)
-target_link_libraries(my_app PRIVATE urlparser::urlparser)
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)
 ```
 
 ### Python and Command Line:

{liburlparser-2.1.0 → liburlparser-2.2.0}/benchmarks/cpp/CMakeLists.txt

@@ -25,7 +25,7 @@ FetchContent_Declare(
 FetchContent_MakeAvailable(liburlparser ada)
 
 add_executable(bench_cpp bench.cpp)
-target_link_libraries(bench_cpp PRIVATE urlparser::urlparser ada::ada)
+target_link_libraries(bench_cpp PRIVATE liburlparser::urlparser ada::ada)
 target_compile_definitions(bench_cpp PRIVATE
     URLPARSER_BENCH_DOMAINS_FILE="${URLPARSER_BENCH_DOMAINS_FILE}"
     URLPARSER_BENCH_URLS_FILE="${URLPARSER_BENCH_URLS_FILE}"

{liburlparser-2.1.0 → liburlparser-2.2.0}/cmake/liburlparserConfig.cmake

@@ -3,9 +3,9 @@
 # Usage (CMake >= 3.15):
 #
 #   find_package(liburlparser CONFIG REQUIRED)
-#   target_link_libraries(app PRIVATE urlparser::urlparser)
+#   target_link_libraries(app PRIVATE liburlparser::urlparser)
 #
-# `urlparser::urlparser` is the same target you get from
+# `liburlparser::urlparser` is the same target you get from
 # FetchContent/add_subdirectory. It carries everything a consumer needs: the
 # compiled static library and the include directory for <urlparser.h>. All
 # paths are resolved relative to this file, so the package works wherever it

{liburlparser-2.1.0 → liburlparser-2.2.0}/docs/index.html.in

@@ -188,10 +188,10 @@ FetchContent_Declare(
 )
 FetchContent_MakeAvailable(liburlparser)
 
-target_link_libraries(my_app PRIVATE urlparser::urlparser)</code></pre>
+target_link_libraries(my_app PRIVATE liburlparser::urlparser)</code></pre>
             </div>
             <div class="install-note">
-                Requires CMake 3.19+ and a C++17-compatible compiler. Link against <code>urlparser::urlparser</code> and include <code>urlparser.h</code>.
+                Requires CMake 3.19+ and a C++17-compatible compiler. Link against <code>liburlparser::urlparser</code> and include <code>urlparser.h</code>.
             </div>
         </div>
 

{liburlparser-2.1.0 → liburlparser-2.2.0}/examples/find_liburlparser_via_python/CMakeLists.txt

@@ -19,7 +19,7 @@ execute_process(
 find_package(liburlparser CONFIG REQUIRED)
 
 add_executable(app main.cpp)
-target_link_libraries(app PRIVATE urlparser::urlparser)
+target_link_libraries(app PRIVATE liburlparser::urlparser)
 
 # Plain paths are available too, for build systems that don't use the target:
 message(STATUS "urlparser_LIB_PATH     = ${urlparser_LIB_PATH}")

{liburlparser-2.1.0 → liburlparser-2.2.0}/include/urlparser.h

@@ -48,7 +48,7 @@
 //   - pyproject.toml's [tool.scikit-build.metadata.version] regex provider
 //   - version.py (bump/tag helper)
 #define URLPARSER_VERSION_MAJOR 2
-#define URLPARSER_VERSION_MINOR 1
+#define URLPARSER_VERSION_MINOR 2
 #define URLPARSER_VERSION_PATCH 0
 
 #define URLPARSER_VERSION_ENCODE(maj,min,pat) (((maj)*10000)+((min)*100)+(pat))