3 ms·

(title omits: in Idris) But it's really not fair to C to exclude all the special compiler support we do have with -Werror=format=2. Given the existence of the

by o11c 2y ago

(title omits: in Idris) But it's really not fair to C to exclude all the special compiler support we do have with -Werror=format=2. Given the existence of the format/format_arg attributes it's really quite powerful; the only thing that's missing is support for custom formats, which requires a compiler plugin. In C++ you can even make a wrapper macro that safely converts arbitrary C++ classes into suitable types for C printf. If you know all your types ahead of time you can do similar with C11 _Generic I think.

mananaysiempre 2y ago
> in Idris For a lower-tech approach, you can even do one in Standard ML[1,2], more or less in plain Hindley–Milner (no substantial use of modules). The error messages are admittedly miserable. Perhaps it would be more correct to call that one a type-safe middle ground between printf and std::ostream::operaror<<, though, and I have to admit that loses what for me is one of the principal advantages of (POSIX) printf: the format string is a string, so localization—which inevitably involves rearranging placeholders—is much easier. It might be that the correct solution involves a string with placeholders, but with formatting instructions—and all dependencies on argument types—pushed outside (to the argument list or similar). I’ve toyed with something like that in C; unfortunately, it’s not as succint as normal printf unless I want to tie it to dynamic memory allocation. For C, there’s also the advantage that you can avoid pulling in floating-point formatting code if you don’t need it, without resorting to GCC-specific tricks like Cosmopolitan does[3], and that code is the biggest (code and data) space hog that keeps printf out of embedded and other kinds of lean programs. [1] http://mlton.org/Printf http://mlton.org/Printf [2] http://mlton.org/Fold http://mlton.org/Fold [3] https://justine.lol/sizetricks/ https://justine.lol/sizetricks/
mimd 2y ago
Not to fully encapture C's variadic printf, but on the macro front, I've been having fun creating type checking for macros by a constexpr allocation from a _generic switch type check of the args and then checking the result with static_asserts. I think there was a hack before as you mentioned but would have to dig through my tests to find it, but the C23 version is quite clean. Might want see if I can get it to be clean for a variadic printf version.

by o11c 2y ago

mananaysiempre 2y ago

> in Idris For a lower-tech approach, you can even do one in Standard ML[1,2], more or less in plain Hindley–Milner (no substantial use of modules). The error messages are admittedly miserable. Perhaps it would be more correct to call that one a type-safe middle ground between printf and std::ostream::operaror<<, though, and I have to admit that loses what for me is one of the principal advantages of (POSIX) printf: the format string is a string, so localization—which inevitably involves rearranging placeholders—is much easier. It might be that the correct solution involves a string with placeholders, but with formatting instructions—and all dependencies on argument types—pushed outside (to the argument list or similar). I’ve toyed with something like that in C; unfortunately, it’s not as succint as normal printf unless I want to tie it to dynamic memory allocation. For C, there’s also the advantage that you can avoid pulling in floating-point formatting code if you don’t need it, without resorting to GCC-specific tricks like Cosmopolitan does[3], and that code is the biggest (code and data) space hog that keeps printf out of embedded and other kinds of lean programs. [1] http://mlton.org/Printf http://mlton.org/Printf [2] http://mlton.org/Fold http://mlton.org/Fold [3] https://justine.lol/sizetricks/ https://justine.lol/sizetricks/

mimd 2y ago

Not to fully encapture C's variadic printf, but on the macro front, I've been having fun creating type checking for macros by a constexpr allocation from a _generic switch type check of the args and then checking the result with static_asserts. I think there was a hack before as you mentioned but would have to dig through my tests to find it, but the C23 version is quite clean. Might want see if I can get it to be clean for a variadic printf version.