This is the mail archive of the cygwin-developers@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

RE: SHOWSTOPPER: [was Re: vfscanf in newlib]


> -----Original Message-----
> From: Christopher Faylor [mailto:cgf@redhat.com]
> Sent: Monday, April 23, 2001 9:16 AM
> To: cygwin-developers@cygwin.com
> Subject: Re: SHOWSTOPPER: [was Re: vfscanf in newlib]
> 
> 
> On Mon, Apr 23, 2001 at 08:59:47AM +1000, Robert Collins wrote:
> >I might be a cancellation point being checked before MTInterface is
> >initialised. If thats the case, I suggest that MTInterface is
> >initialised earlier.
> 
> It should be initialized pretty early.  I don't see any way 
> for /etc/passwd
> to be read prior to this initialization.
> 
> This is the initialization in question, right:
> 
>   user_data->threadinterface->Init (user_data->forkee);
> 

Yes.

Actually can you tell me - the if (forked) below only runs in spawned
cygwin process's ? If so I've got a few bugs to fix.. 

Rob


void
MTinterface::Init (int forked)
...
  if (forked)
    return;

  /* shm areas a inherited when forking */
  shm_head = NULL;
  /* possible the atfork lists should be inited here as well */

  mainthread.win32_obj_id = myself->hProcess;
  mainthread.setThreadIdtoCurrent ();
  /* store the main thread's self pointer */
  TlsSetValue (thread_self_dwTlsIndex, &mainthread);


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]