This is the mail archive of the
cygwin-developers
mailing list for the Cygwin project.
Re: Output of "uname -s" and "uname -o"
On Mon, Jun 09, 2008 at 04:21:37PM +0200, Corinna Vinschen wrote:
>On Jun 9 09:54, Igor Peshansky wrote:
>> On Mon, 9 Jun 2008, Corinna Vinschen wrote:
>> > I don't think we can change this. The sysname field is the only one
>> > which you can use to identify the system. There's no other room for
>> > this.
>>
>> You're right, both release and version fields are pretty much filled to
>> capacity. I was somehow under the impression that the OS name was a field
>> returned by Cygwin -- I now see from the sources that it's a coreutils
>> invention.
>>
>> Technically, the version field would be the right place to store the
>> information about the underlying Windows version.
>>
>> I just looked at the Linux utsname.h, and it defines the length of all the
>> fields to be 65 characters. Is there a compelling reason for Cygwin to be
>> limited to 20?
>
>Old header files plus backward compatibility. The structure is exposed
>to applications. Every change here requires to check for the Cygwin
>version under which the application has been built. I cursed the short
>field length myself at onepoint, but the added complexity looks a bit
>questionable.
Ditto. I cursed the short field names when I first added the _{95,NT}
field to uname. It's possible to work around with #define magic like
we do in other places but, as Corinna says, it doesn't seem worth the
effort, especially given that it is now potentially entrenched in many
configure programs out there.
cygwin-developers FAQ anyone?
cgf