On Mon, 2017-08-14 at 01:41:43 +0900, Vitalie Ciubotaru wrote: > On 08/06/2017 12:28 AM, Guillem Jover wrote: > > > > Well, obviously not entirely as complete but the BSDs (and libbsd > > otherwise) do have something like <sys/bitstring.h>. > > > > The difference between sys/bitstring.h and the package I propose is that > bitstring.h implements a bit array within one unit of data storage (unsigned > char, 1 byte, 8 bits), while my package uses an array of units (array of > unsigned long integers, usually 32 or 64 bits EACH). This means that BSD's bit > array can't be longer that 8 bits, while mine is limited by memory size only.
Nope, <sys/bitstring.h> does support an unlimited amount of bits by using arrays of unsigned chars. You might want to recheck its implementation. :) > > BTW, I noticed when I checked the implementation that you are using > > HAVE_BITFIELD_H as the header macro protector, but that's actually a > > bad idea, as with the typical autoconf based build-system that macro > > will already be defined if the build-system does something like > > AC_CHECK_HEADERS([bitfield.h]), which means the contents will be > > secluded. > > Oh, yes, you are right. I was able to reproduce this problem. The name of the > header guard needs to be changed. Thank you very much for pointing it out. No problem! Regards, Guillem