Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/tierra-colada/cppguts
Tool aimed at C/C++ source code correction that allows to automatically find and copy/paste new function definition
https://github.com/tierra-colada/cppguts
c cpp edit-c edit-cpp parse parse-c parse-cpp parser-c parser-cpp python
Last synced: about 2 months ago
JSON representation
Tool aimed at C/C++ source code correction that allows to automatically find and copy/paste new function definition
- Host: GitHub
- URL: https://github.com/tierra-colada/cppguts
- Owner: tierra-colada
- License: mit
- Created: 2021-09-18T16:56:46.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-05-26T18:57:45.000Z (over 2 years ago)
- Last Synced: 2024-11-18T17:59:58.792Z (2 months ago)
- Topics: c, cpp, edit-c, edit-cpp, parse, parse-c, parse-cpp, parser-c, parser-cpp, python
- Language: Python
- Homepage:
- Size: 67.4 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# cppguts
If your project depends on some external C/C++ projects and
you want to make some changes in external functions/methods
and then copy/paste these changes automatically - this package may help you.
There are two tools:
1) `editcpp` used to edit source code;
2) `dumpcpp` used to print some information about C/C++ translation units
We will discuss `editcpp` as it is the objective tool.
**`editcpp` doesn't work with templates.**
Same tool aimed at editing `python` files is also available as [pythonguts](https://github.com/tierra-colada/pythonguts)
## Installation
`cppguts` is available at [PyPI](https://pypi.org/project/cppguts/):
```bash
pip install cppguts
```
## The idea behind `editcpp` tool
`editcpp` uses `libclang` to find function/method definition start/end lines in text file (`.c`, `.h`, `.hpp`, `.cpp` or whatever C/C++ extension). `libclang` parses each `dest.cpp` and `src.cpp` and everything that is
included by `#include` preprocessor directives. Then `editcpp` tool
selects all functions and methods defined in `dest.cpp` and `src.cpp`
(it assumes that you know where old and new function/method definition resides) and tries to find matching functions/methods. After that `editcpp` copies function/method definition from `src.cpp` and pastes it to the `dest.cpp` while deleting old funcion/method definition.
To find common function `editcpp` checks:
* are they both _definitions?_
* are they both _const?_
* are they both _static?_
* are they both _vitual?_
* are they both _functions?_
* are they both _methods?_
* do they both have the same name?
* do they both have the same return type and arg types?
* do they both have the same semantic parent (classname)? (for methods only)
## Notes
* you are free to pass any clang argument after `editcpp` otions but don't pass file without flag to clang (as we usually do when using clang)
* if you parse files that includes non standard headers then you probably need to pass `-I` flag to clang to include directories;
* if you parse `C++` files then it may be necessary to pass `-std=c++03` (or higher) flag to inform clang that the file is `C++`;
* if your new function/method definition uses external types then
these types must be preliminary declared (not necessary to define them);
Remember that after `editcpp` finds common functions/methods
it will simply copy selected text lines from one file to another
## Example
original function/method definition file **dest.h**:
```cpp
#include
// helper class that is used by the target class `Src`
class SrcPrivate {
public:
SrcPrivate(){};
void add(int v){
val += v;
}
void substract(int v){
val -= v;
}
private:
int val;
}
// target class
class Src {
// method defined inside the class
void add(SrcPrivate p, int v){
p.add(v);
}
// method defined outside the class
void substract(SrcPrivate p, int v);
// we won't tuch this method
void untouched_print(int v){
std::cout << "The value is:\t" << v << std::endl;
}
}
void Src::substract(SrcPrivate p, int v){
p.substract(v);
}
// simple function
void foo(int &v){
v--;
}
namespace ns {
// function in namespace
void bar(int &v){
v++;
}
}
```
new function/method definition file **src.h**:
```cpp
class SrcPrivate; // declare helper type
// target class
class Src {
// NEW method definition inside class
void add(SrcPrivate p, int v){
p.add(v) + 10;
}
// NEW method definition ouside class
void substract(SrcPrivate p, int v);
}
void Src::substract(SrcPrivate p, int v){
p.substract(v) - 10;
}
// NEW simple function definition
void foo(int &v){
v -= 10;
}
// NEW function definition in namespace
namespace ns {
void bar(int &v){
v += 10;
}
}
```
Run:
`editcpp --src-file=src.h --dest-file=dest.h --oldfile-keep -std=c++03`
The `-std=c++03` tells the clang to parse the files as C++. Also you may need to use any other clang flags like `-I` to include directories that are required by the files.
`--oldfile-keep` (default) is used to keep the original file (it will be renamed by adding `_OLD_N` suffix). Otherwise use `--oldfile-delete` to delete the original file.
Another option is to run the test (though the test deletes all the generated files so you better take a look in `/tests` dir):
`python -m unittest cppguts.tests.test_cppguts`