Update sigaction(2) to reflect the implemention in src/sys/sys/signal.h

[sevan]

Summary

Bug 211365
Konstantin Stroykovskiy <[e-mail]>

[bcr]

Gee, You really make me wish I had a src bit to approve this. :-\
I can't help here, sorry.

[jilles]

Note the paragraph below starting with "The above declaration of struct sigaction is not literal.". The man page is inconsistent now.

Also, I don't think applications are supposed to use the __sigaction_u, __sa_handler and __sa_sigaction identifiers.

The definition in the man page could be clarified using a C11 anonymous union, but this cannot be used in the implementation, since we do not want to require using recent compilers and standards to compile application code.

[bjk]

In D12006#248864, @bcr wrote:

Gee, You really make me wish I had a src bit to approve this. :-\
I can't help here, sorry.

(unmentored) doc committers have privileges to commit man page changes to src without additional src approval. Why would one think that the situation would be different when approving mentored doc committers for similar purposes?

[bcr]

Right, I sometimes hesitate because I can't grasp the extend of the change and whether what the man page change does reflects the src tree and what is done there.

[sevan]

@jilles best to just revert the commit & leave it as-is?

[jilles]

In D12006#249341, @sevan wrote:

@jilles best to just revert the commit & leave it as-is?

As far as I can see, nothing was committed so there is nothing to revert.

Perhaps something could be changed to clarify somehow inside the synopsis that sa_handler and sa_sigaction may overlap.

[sevan]

Ah , I got my reviews mixed up. Abandoning.
Thanks for the input :)