Probably incorrect formatting in timespec.3type

Probably incorrect formatting in timespec.3type

[Alejandro Colomar]

[-- Attachment #1.1: Type: text/plain, Size: 418 bytes --]

Hi наб!

I was moving timespec to a separate type, when I noticed a formatting 
(.I) that I think is misplaced.  I didn't fix it, because I wanted to 
confirm with you what was the intention.

Check timespec(3type) and search for 'syscall', followed by 'long', and 
confirm if the formatting is what you wanted to put or not.

Cheers,

Alex

-- 
Alejandro Colomar
<http://www.alejandro-colomar.es/>

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Probably incorrect formatting in timespec.3type

[наб]

[-- Attachment #1: Type: text/plain, Size: 696 bytes --]

Hi!

On Thu, Jul 21, 2022 at 03:38:46PM +0200, Alejandro Colomar wrote:
> Check timespec(3type) and search for 'syscall', followed by 'long', and
> confirm if the formatting is what you wanted to put or not.

It is:
  syscall is emphasised, since that's both the fundamental point
  and a subtle difference as compared to the very strong binding of
  the "long" which also designates a C type, and
  long is Some Object, not a type itself.

You may prefer sth like I[syscall] B[long]
(which I'd say is slightly worse because the whole point is that the
 syscall long is not long on X32, so the long is not literal),
but having them both be I would be misleading.

Best,
наб

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Probably incorrect formatting in timespec.3type

[Alejandro Colomar]

[-- Attachment #1.1: Type: text/plain, Size: 1099 bytes --]

Hi!

On 7/21/22 16:39, наб wrote:
> Hi!
> 
> On Thu, Jul 21, 2022 at 03:38:46PM +0200, Alejandro Colomar wrote:
>> Check timespec(3type) and search for 'syscall', followed by 'long', and
>> confirm if the formatting is what you wanted to put or not.
> 
> It is:
>    syscall is emphasised, since that's both the fundamental point
>    and a subtle difference as compared to the very strong binding of
>    the "long" which also designates a C type, and
>    long is Some Object, not a type itself.

What is exactly "syscall long".  Is it what the kernel understands 
internally as a long (which I guess doesn't necessarily match with what 
userspace understands as long)?  I'm not so used to kernel details...

> 
> You may prefer sth like I[syscall] B[long]
> (which I'd say is slightly worse because the whole point is that the
>   syscall long is not long on X32, so the long is not literal),
> but having them both be I would be misleading.

That makes sense.

Thanks,

Alex

> 
> Best,
> наб

-- 
Alejandro Colomar
<http://www.alejandro-colomar.es/>

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Probably incorrect formatting in timespec.3type

[наб]

[-- Attachment #1: Type: text/plain, Size: 531 bytes --]

On Fri, Jul 22, 2022 at 12:45:37AM +0200, Alejandro Colomar wrote:
> What is exactly "syscall long".  Is it what the kernel understands
> internally as a long (which I guess doesn't necessarily match with what
> userspace understands as long)?  I'm not so used to kernel details...

Yeah, if you're looking at it that way then you could call it
"the kABI long", as you say ‒ on X32 the kABI is amd64=LP64,
but the userspace is ILP32: the kABI long is a u64,
and the smallest (only) userspace u64 is long long.

наб

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]