This is the mail archive of the
cygwin-developers@cygwin.com
mailing list for the Cygwin project.
Re: 1.5.8 release announcement
- From: Christopher Faylor <cgf at alum dot bu dot edu>
- To: cygwin-developers at cygwin dot com
- Date: Mon, 15 Mar 2004 16:22:17 -0500
- Subject: Re: 1.5.8 release announcement
- References: <20040315192728.GA25871@redhat.com> <40561593.F4C3AC2@ieee.org>
- Reply-to: cygwin-developers at cygwin dot com
On Mon, Mar 15, 2004 at 03:44:03PM -0500, Pierre A. Humblet wrote:
>
>
>Christopher Faylor wrote:
>>
>> If all of my tests continue to work, I plan on releasing this tomorrow.
>>
>> Any additions or corrections to the below text are welcome.
>
>Here is another one:
>- Allow Administrators to also "kill" all top level Cygwin processes. (Pierre Humblet)
>
>>- Implement sighold. (Christopher Faylor)
>
>I notice that sighold's argument sig is an integer, while that of
>sigpause is a mask (in Cygwin & BSD), which makes sigpause identical to
>sigsuspend.
>
>sighold and sigpause are supposed to work together, see "Application
>Usage" in http://www.opengroup.org/onlinepubs/007908799/xsh/signal.html
>So now that sighold is present it would make sense to adopt the Posix
>definition for sigpause (there is no sighold in BSD).
sigpause was exported in its current form in 1998. We can't just change
its behavior now? sigpause mimics the linux man page description for
this function although the header file allows different usages depending
on various defines.
There is no linux man page that I can find for sighold but cygwin
seems to match the definition in the linux header file.
I suppose that we could attempt to mimic the way linux does it using a
NEW_FUNCTIONS OBSOLETE_FUNCTIONS mechanism but since I don't see any
complaints about "sigpause" in my search of the archives, I don't think
it's a big problem.
cgf