Re: [Bug-gnulib] Re: gnulib module for uintmax_t? (and best practice questions)

[Bruno Haible]

Simon Josefsson wrote:
> Having to replace all occurrences of uint32_t with @address@hidden
> in every package not only produce hard to read source code

Not much worse than 'gint' in GNOME...

> interesting type casting problems: is it OK to cast idn_uint32_t to
> nls_uint32_t?

As integer types, certainly, because even if one of them is 'unsigned int'
and the other is 'unsigned long', both cover the range of 32-bit
integers, and the values are assumed to be 32-bit integers.

As pointer types, no: 'idn_uint32_t *' and 'nls_uint32_t *' are different.

> > I would propose to let the autoconf macro determine the full pathname of
> > the system's <stdint.h> (through "$CC -E" and grepping for #line
> > statements), and do a #include of this full pathname at the beginning of
> > our stdint.h.
>
> What if the system has several compilers, each with its own stdint.h?

I make the assumption that question 1) has been solved by _not_ installing
the generated/parametrized stdint.h.

Bruno