Detect C++ source and header extensions beyond .cpp

CHANGELOG.md

@@ -8,6 +8,7 @@ The format of this changelog is based on [Keep a Changelog](https://keepachangel
  - `.kts` now reports as Kotlin and `.gradle` as Groovy. Both were missing from the extension table, so Kotlin scripts and every Gradle build script — `build.gradle`, `settings.gradle`, `build.gradle.kts`, `settings.gradle.kts` — reported `Unknown`. Gradle selects the DSL by extension alone, so the mapping is exact. Repositories built with Gradle will now show a small Groovy or Kotlin count from their build scripts.
  - `build.gradle.kts` now carries the same `gradle`, `build` and `dependencies` tags as `build.gradle`. A project using the Gradle Kotlin DSL previously looked like it declared no dependencies.
  - `settings.gradle` and `settings.gradle.kts` are tagged `gradle` and `build`. They declare project structure rather than dependencies, so they do not carry `dependencies`.
+ - `.cc`, `.cxx` and `.c++` now report as `C++`, and `.hpp`, `.hh`, `.hxx` and `.h++` as `C++ Header`, a new language paralleling `C Header`. Only `.cpp` was mapped before, so the rest reported `Unknown`. `.h` is unchanged and still reports `C Header`.
  - `.mm` now reports as `Objective-C++`, a new language. It was missing from the extension table and reported `Unknown`.
  - Swift Package Manager files are recognised. `Package.swift` is tagged `build`, `dependencies`, `Swift` and `SwiftPM`; its lock file `Package.resolved` is tagged `dependencies`, `Swift` and `SwiftPM` and reports as `JSON` instead of `Unknown`. A Swift package previously looked like it declared no dependencies. `Swift` and `SwiftPM` are new tags.
  - CocoaPods files are recognised. `Podfile` and `Podfile.lock` are tagged `dependencies` and `CocoaPods`, and `Podfile` reports as `Ruby` instead of `Unknown`.
@@ -16,7 +17,7 @@ The format of this changelog is based on [Keep a Changelog](https://keepachangel
  All of these are new tags on previously untagged files, so kospex is compatible either way; a re-sync is needed to pick them up.
 
 ### Changed
- - The tag vocabulary grew from 119 to 123 tags, the filetype vocabulary from 86 to 88 and the language vocabulary from 32 to 33.
+ - The tag vocabulary grew from 119 to 123 tags, the filetype vocabulary from 86 to 89 and the language vocabulary from 32 to 34.
  - The Click floor moved from `>=8.3.1` to `>=8.3.3`, so an install can no longer resolve to a vulnerable Click release.
  - The package author email is now `peter@kospex.io`.
 

changes/2026-10-cpp-extensions.md

@@ -0,0 +1,75 @@
+# C++ Source and Header Extensions
+
+Closes [#43](https://github.com/kospex/panopticas/issues/43).
+
+## Summary
+
+`.cpp` was the only C++ extension in `EXT_FILETYPES`. Every other common
+spelling reported `Unknown`:
+
+    file.cc    'Unknown'        file.hpp   'Unknown'
+    file.cxx   'Unknown'        file.hh    'Unknown'
+    file.c++   'Unknown'        file.hxx   'Unknown'
+                                file.h++   'Unknown'
+
+A C++ codebase using `.cc` and `.hpp` — a very common pairing — showed no C++
+at all. The gap surfaced during the `.m`/`.h` analysis in #40.
+
+## Changes
+
+Seven entries in `EXT_FILETYPES`, and `C++ Header` added to
+`LANGUAGE_FILETYPES`:
+
+    ".cc":  "C++"            ".hpp": "C++ Header"
+    ".cxx": "C++"            ".hh":  "C++ Header"
+    ".c++": "C++"            ".hxx": "C++ Header"
+                             ".h++": "C++ Header"
+
+## Source
+
+The [GCC documentation](https://gcc.gnu.org/onlinedocs/gcc/Overall-Options.html)
+lists the suffixes it treats as C++:
+
+- **Source:** `.cc`, `.cp`, `.cxx`, `.cpp`, `.CPP`, `.c++`, `.C`
+- **Header:** `.hh`, `.H`, `.hp`, `.hxx`, `.hpp`, `.HPP`, `.h++`, `.tcc`
+
+## Design notes
+
+- **`C++ Header` is a new language name**, paralleling the existing
+  `C Header`. Headers are reported separately from sources because that is
+  what panopticas already does for C. Reporting them as plain `C++` was the
+  alternative; it would have been simpler but inconsistent with `.h`.
+- **`.h` is unchanged.** It is shared by C, C++ and Objective-C — GCC's own
+  entry for it reads "C, C++, Objective-C or Objective-C++ header file" — and
+  cannot be resolved from the path. That is the subject of #40.
+- Additive only: no existing mapping changes, so kospex is unaffected until a
+  re-sync.
+
+## Not mapped
+
+The complete list of C++-related suffixes considered and left out, so none
+needs rediscovering.
+
+| Suffix | What it is | Reports as | Why not mapped |
+|---|---|---|---|
+| `.C` | C++ source (GCC, literal capital) | `C` | panopticas lowercases the extension before lookup, so it is indistinguishable from `.c` |
+| `.H` | C++ header (GCC, literal capital) | `C Header` | Same: indistinguishable from `.h` |
+| `.cp` | C++ source (GCC) | `Unknown` | Rare, and short enough to collide with unrelated uses |
+| `.hp` | C++ header (GCC) | `Unknown` | Rare |
+| `.tcc` | C++ header holding template definitions (GCC) | `Unknown` | Rare |
+| `.ii` | C++ source that should not be preprocessed (GCC) | `Unknown` | Compiler output, not something a repository normally holds |
+| `.ipp`, `.tpp`, `.inl`, `.txx` | Template and inline implementation files | `Unknown` | Project conventions rather than compiler-defined suffixes; not checked against any documentation |
+| `.cppm`, `.ixx`, `.ccm`, `.mpp` | C++20 module interface units | `Unknown` | Vary by compiler; not checked against any compiler's documentation |
+| `.h` | C, C++ or Objective-C header | `C Header` | Shared by three languages and not resolvable from the path — see #40 |
+
+`.CPP` and `.HPP`, which GCC lists separately, need no entry: lowercasing
+resolves them to `.cpp` and `.hpp`, so they report `C++` and `C++ Header`.
+
+The first six rows are from the GCC suffix list above. The template and module
+rows are from general knowledge and would need verifying before any of them
+became a mapping.
+
+## Tests
+
+`tests/test_panopticas.py` covers all seven extensions, asserts `.h` still
+reports `C Header`, and has a fixture for each extension in `src/tests/`.

docs/api.md

@@ -53,7 +53,7 @@ EXT_FILETYPES["dockerfile"]  # "Dockerfile" (special case)
 `EXT_FILETYPES` mixes two kinds of value: things that are programming languages
 (`Python`, `Go`) and things that are file types but not languages (`PNG`,
 `Gitignore`, `Lock`). These two collections classify every value in the table —
-33 languages and 55 non-languages. `get_languages()` and `get_filetypes()` read
+34 languages and 55 non-languages. `get_languages()` and `get_filetypes()` read
 from them.
 
 ```python
@@ -386,7 +386,7 @@ Return every file type panopticas recognises, languages or not.
 from panopticas.core import get_filetypes
 
 get_filetypes()
-# Returns: ['Apache JMeter', 'ASP.NET', ..., 'ZIP']  (88 file types)
+# Returns: ['Apache JMeter', 'ASP.NET', ..., 'ZIP']  (89 file types)
 ```
 
 **Returns:** Sorted list of file type strings
@@ -399,7 +399,7 @@ Return every file type that is a programming language.
 from panopticas.core import get_languages
 
 get_languages()
-# Returns: ['C', 'C Header', 'C#', 'C++', 'CSS', ..., 'Vue']  (33 languages)
+# Returns: ['C', 'C Header', 'C#', 'C++', 'C++ Header', ..., 'Vue']  (34 languages)
 ```
 
 **Returns:** Sorted list of language names — the subset of `get_filetypes()`

docs/cli.md

@@ -396,8 +396,8 @@ count.
 | Command | Shape |
 |---|---|
 | `tags` | `{"tags": [...], "count": 123}` |
-| `languages` | `{"languages": [...], "count": 33}` |
-| `filetypes` | `{"filetypes": [...], "count": 88}` |
+| `languages` | `{"languages": [...], "count": 34}` |
+| `filetypes` | `{"filetypes": [...], "count": 89}` |
 
 ```console
 $ panopticas languages --json | jq -r '.count'

src/panopticas/constants.py

@@ -20,8 +20,11 @@
 
 EXT_FILETYPES = {
     ".c": "C",
+    ".c++": "C++",
+    ".cc": "C++",
     ".class": "Java Class",
     ".cpp": "C++",
+    ".cxx": "C++",
     ".cs": "C#",
     ".cjs": "JavaScript",   # CommonJS module, explicit
     ".cts": "TypeScript",   # TypeScript emitting CommonJS
@@ -42,6 +45,10 @@
     ".groovy": "Groovy",
     ".gsp": "Groovy Server Pages",
     ".h": "C Header",
+    ".h++": "C++ Header",
+    ".hh": "C++ Header",
+    ".hpp": "C++ Header",
+    ".hxx": "C++ Header",
     ".aspx": "ASP.NET",
     ".ascx": "ASP.NET User Control",
     ".htm": "HTML",
@@ -179,6 +186,7 @@
     "C Header",
     "C#",
     "C++",
+    "C++ Header",
     "CSS",
     "Dockerfile",
     "Go",

src/tests/file.c++

@@ -0,0 +1,6 @@
+#include <iostream>
+
+int main() {
+    std::cout << "hello" << std::endl;
+    return 0;
+}

src/tests/file.cc

@@ -0,0 +1,6 @@
+#include <iostream>
+
+int main() {
+    std::cout << "hello" << std::endl;
+    return 0;
+}

src/tests/file.cxx

@@ -0,0 +1,6 @@
+#include <iostream>
+
+int main() {
+    std::cout << "hello" << std::endl;
+    return 0;
+}

src/tests/file.h++

@@ -0,0 +1,10 @@
+#pragma once
+
+namespace example {
+
+class Greeter {
+public:
+    void greet() const;
+};
+
+}  // namespace example

src/tests/file.hh

@@ -0,0 +1,10 @@
+#pragma once
+
+namespace example {
+
+class Greeter {
+public:
+    void greet() const;
+};
+
+}  // namespace example

src/tests/file.hpp

@@ -0,0 +1,10 @@
+#pragma once
+
+namespace example {
+
+class Greeter {
+public:
+    void greet() const;
+};
+
+}  // namespace example

src/tests/file.hxx

@@ -0,0 +1,10 @@
+#pragma once
+
+namespace example {
+
+class Greeter {
+public:
+    void greet() const;
+};
+
+}  // namespace example

tests/test_panopticas.py

@@ -83,6 +83,18 @@ def test_kotlin(self):
     def test_kotlin_script(self):
         assert get_extension_filetype(".kts") == "Kotlin"
 
+    @pytest.mark.parametrize("ext", [".cc", ".cxx", ".c++"])
+    def test_cpp_source_extensions(self, ext):
+        assert get_extension_filetype(ext) == "C++"
+
+    @pytest.mark.parametrize("ext", [".hpp", ".hh", ".hxx", ".h++"])
+    def test_cpp_header_extensions(self, ext):
+        assert get_extension_filetype(ext) == "C++ Header"
+
+    def test_plain_h_is_still_a_c_header(self):
+        # .h is shared by C, C++ and Objective-C and is not resolved here.
+        assert get_extension_filetype(".h") == "C Header"
+
     def test_podfile_is_ruby(self):
         # CocoaPods loads an extensionless Podfile as Ruby.
         assert get_extension_filetype("Podfile") == "Ruby"
@@ -776,6 +788,18 @@ def test_gradle_kotlin_settings_detected(self, fixtures_exist):
     def test_objective_cpp_detected(self, fixtures_exist):
         assert get_language("file.mm") == "Objective-C++"
 
+    @pytest.mark.parametrize("filename", [
+        "file.cc", "file.cxx", "file.c++",
+    ])
+    def test_cpp_sources_detected(self, fixtures_exist, filename):
+        assert get_language(filename) == "C++"
+
+    @pytest.mark.parametrize("filename", [
+        "file.hpp", "file.hh", "file.hxx", "file.h++",
+    ])
+    def test_cpp_headers_detected(self, fixtures_exist, filename):
+        assert get_language(filename) == "C++ Header"
+
     def test_podfile_detected(self, fixtures_exist):
         assert get_language("Podfile") == "Ruby"