Albin Ahlbäck

@fredrik-johansson, what do you think about a creating generic header for Arb, like FLINT has `flint.h`? I know `arb.h` is already taken, but I have two reasons why I think...

GCC complains at some points that `arf_[src]ptr` is used in the header and that `arf_t` is used in the function definition (or if it was the other way around). This...

By now, I seriously doubt that `XXX_t` and `XXX_ptr` is equivalent in function definitions. I am brewing on a PR for Flint to suppress warnings, and it seems like the...

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102810 If I understand this correctly: ```c void fun(int a[]); // Nemas problemas void fun(int * a); // Nemas problemas void fun(int a[3]); // ? haven't encountered a problem with...

> This is still mysterious to me. arf_t (for example) is defined as an array of arf_struct of length 1, so it should always be safe to read exactly sizeof(arf_struct)...

I'm not sure, but can it be because of this? ```c #include int main() { long int a[3] = {0, 1, 2}; printf("&a = %p\n", &a); printf("a + 0 =...

Any thoughts into this? As mentioned, I can prepare a PR which shouldn't take too long with some `sed`.

Yeah, my example is incorrect. I read somewhere that something similar is fixed in the current state of GCC, so let's wait until GCC 12 is released until we draw...

> GCC complains at some points that `arf_[src]ptr` is used in the header and that `arf_t` is used in the function definition (or if it was the other way around)....

Neither `arb_get_str` nor `arb_set_str` says in their documentation that they preserve the radius. It only states that the old interval is contained in the new interval.