On Oct 4 20:44, Yaakov S wrote:
> On 29/09/2009 20:11, Yaakov (Cygwin/X) wrote:
>> On 29/09/2009 19:35, Yaakov (Cygwin/X) wrote:
>>> Anyway, to answer the question, AFAICS in glibc, #include
>>> unconditionally[1]. ( is just one line:
>>> #include [2])
>>>
>>> So should I take the first route,
On 05/10/2009 03:37, Corinna Vinschen wrote:
Newlib, methinks.
OK, I'll send a patch there.
I think the newlib kill declaration should be changed to pid_t, since
that's simply correct per POSIX.
I can;t believe the RTEMS people have a problem with that.
Did you forget about the ESTRPIPE ep
On Oct 5 09:18, Yaakov S wrote:
> On 05/10/2009 03:37, Corinna Vinschen wrote:
>> Newlib, methinks.
>
> OK, I'll send a patch there.
>
>> I think the newlib kill declaration should be changed to pid_t, since
>> that's simply correct per POSIX.
>>
>> I can;t believe the RTEMS people have a problem
Normally, posix programs should call abort(), exit(), _exit(), kill() --
or various pthread functions -- to terminate operation (either their
own, or that of some other processes/thread). However, there are two
cases where the win32 ExitProcess and TerminateProcess functions might
justifiably be c
Charles Wilson wrote:
> int main(int argc, char* argv[])
> {
> //cygwin_terminate_process (GetCurrentProcess(),
> STATUS_ILLEGAL_DLL_PSEUDO_RELOCATION);
> cygwin_exit_process (STATUS_ILLEGAL_DLL_PSEUDO_RELOCATION);
> exit (1);
> }
Heh. I see what you did there!
Dave Korn wrote:
> Heh. I see what you did there!
As to the actual patch itself, it looks sane (just reading it by eye, I
haven't tested it), and the design motivation seems reasonable.
> @@ -136,11 +136,19 @@ status_exit (DWORD x)
>
> # define self (*this)
> void
> +pinfo::set_exit_cod
Dave Korn wrote:
> Charles Wilson wrote:
>> //cygwin_terminate_process (GetCurrentProcess(),
>> STATUS_ILLEGAL_DLL_PSEUDO_RELOCATION);
> Heh. I see what you did there!
Oops. Never cut-n-paste, and then edit.
--
Chuck
Dave Korn wrote:
> As to the actual patch itself, it looks sane (just reading it by eye, I
> haven't tested it), and the design motivation seems reasonable.
[snip]
> File-local extern declarations are pure evil, let alone function-local ones.
> Why not fix this badness while you're touching it
Charles Wilson wrote:
> Dave Korn wrote:
>> As to the actual patch itself, it looks sane (just reading it by eye, I
>> haven't tested it), and the design motivation seems reasonable.
> [snip]
>> File-local extern declarations are pure evil, let alone function-local
>> ones.
>> Why not fix thi
Dave Korn wrote:
> Wouldn't see any need to move the declaration any; just think that there
> should be one unique extern declaration and it should be in a common header
> that can be #included both where the var is declared and where it is
> referenced.
Well, the variable IS a global, shared b
Charles Wilson wrote:
> By moving the definition to globals.cc, we get its extern declaration in
> the autogenerated globals.h -- which everybody includes via winsup.h.
Ooh, I forgot we had that. Makes sense then.
cheers,
DaveK
On Oct 5 15:04, Charles Wilson wrote:
> Normally, posix programs should call abort(), exit(), _exit(), kill() --
> or various pthread functions -- to terminate operation (either their
> own, or that of some other processes/thread). However, there are two
> cases where the win32 ExitProcess and Te
Dave Korn wrote:
> Charles Wilson wrote:
>
>> By moving the definition to globals.cc, we get its extern declaration in
>> the autogenerated globals.h -- which everybody includes via winsup.h.
>
> Ooh, I forgot we had that. Makes sense then.
As revised:
2009-10-04 Charles Wilson <...>
On Mon, Oct 05, 2009 at 10:27:22PM +0200, Corinna Vinschen wrote:
>On Oct 5 15:04, Charles Wilson wrote:
>> Normally, posix programs should call abort(), exit(), _exit(), kill() --
>> or various pthread functions -- to terminate operation (either their
>> own, or that of some other processes/threa
Corinna Vinschen wrote:
> I have some doubts that we really need such a functionality externally
> available, outside of the limited scenario of something like
> pseudo-reloc. An API for those knowing what this is about is very
> likely sufficient. What about
>
>cygwin_internal (CW_TERMINATE
On Mon, Oct 05, 2009 at 04:28:34PM -0400, Charles Wilson wrote:
>Dave Korn wrote:
>> Charles Wilson wrote:
>>
>>> By moving the definition to globals.cc, we get its extern declaration in
>>> the autogenerated globals.h -- which everybody includes via winsup.h.
>>
>> Ooh, I forgot we had that.
Christopher Faylor wrote:
> I've checked in the sigExeced part of this patch. Thanks.
The checkin does not yet appear to have taken effect, because
$ cvs update && cvs diff
still shows the sigExeced stuff -- so, it's still in the patch/changelog
below.
> For the other part, if you're willing,
On Mon, Oct 05, 2009 at 04:49:11PM -0400, Charles Wilson wrote:
>Corinna Vinschen wrote:
>> I have some doubts that we really need such a functionality externally
>> available, outside of the limited scenario of something like
>> pseudo-reloc. An API for those knowing what this is about is very
>>
Christopher Faylor wrote:
> On Mon, Oct 05, 2009 at 04:49:11PM -0400, Charles Wilson wrote:
>> hmm...probably
>> cygwin_internal (CW_TERMINATE_PROCESS, HANDLE, UINT)
>> cygwin_internal (CW_EXIT_PROCESS, UINT)
>> right?
>
> Do we really have to provide the ability to kill some other process
19 matches
Mail list logo
