Re: Small security patches

2002-12-13 Thread Corinna Vinschen
On Thu, Dec 12, 2002 at 11:34:02AM -0500, Pierre A. Humblet wrote: Christopher Faylor wrote: Actually, if you can get away without using a constructor that would be best. Constructors are a noticeable part of cygwin's startup cost. - Is there a C++ way to initialize a constant class

Re: Small security patches

2002-12-13 Thread Pierre A. Humblet
Corinna Vinschen wrote: On Thu, Dec 12, 2002 at 11:34:02AM -0500, Pierre A. Humblet wrote: Christopher Faylor wrote: Actually, if you can get away without using a constructor that would be best. Constructors are a noticeable part of cygwin's startup cost. What about this idea:

Re: Small security patches

2002-12-13 Thread Pierre A. Humblet
Corinna, Until the initialization issue is settled, here is a patch covering only the internationalization of security.cc It should go in the next cygwin, and I always prefer when there is a sufficiently long bake time. Pierre 2002/12/13 Pierre Humblet [EMAIL PROTECTED] *

Re: Small security patches

2002-12-13 Thread Christopher Faylor
On Fri, Dec 13, 2002 at 09:51:58AM -0500, Pierre A. Humblet wrote: Corinna Vinschen wrote: On Thu, Dec 12, 2002 at 11:34:02AM -0500, Pierre A. Humblet wrote: Christopher Faylor wrote: Actually, if you can get away without using a constructor that would be best. Constructors are a

Re: Small security patches

2002-12-12 Thread Pierre A. Humblet
Christopher Faylor wrote: Actually, if you can get away without using a constructor that would be best. Constructors are a noticeable part of cygwin's startup cost. - Is there a C++ way to initialize a constant class and have it in the .text section, as const int i = 1; would be? - If

Re: Small security patches

2002-12-11 Thread Christopher Faylor
On Wed, Dec 11, 2002 at 11:36:17AM -0500, Pierre A. Humblet wrote: Corinna, here is an internationalization bug fix, and some preliminary definitions for a future well_known_creator approach. Pierre 2002/12/11 Pierre Humblet [EMAIL PROTECTED] * security.h: Declare

Re: Small security patches

2002-12-11 Thread Pierre A. Humblet
Christopher Faylor wrote: Shouldn't the global symbols be marked as NO_COPY? I am not sure why things are as they are. These symbols are initialized in do_global_ctors and never change. Are the constructors running again after a fork? If so, NO_COPY is fine. It would seem more efficient to copy

Re: Small security patches

2002-12-11 Thread Christopher Faylor
On Wed, Dec 11, 2002 at 03:56:17PM -0500, Pierre A. Humblet wrote: Christopher Faylor wrote: Shouldn't the global symbols be marked as NO_COPY? I am not sure why things are as they are. These symbols are initialized in do_global_ctors and never change. Are the constructors running again after a

Re: Small security patches

2002-12-11 Thread Pierre A. Humblet
Christopher Faylor wrote: On Wed, Dec 11, 2002 at 03:56:17PM -0500, Pierre A. Humblet wrote: Christopher Faylor wrote: Shouldn't the global symbols be marked as NO_COPY? I am not sure why things are as they are. These symbols are initialized in do_global_ctors and never change. Are

Re: Security patches

2002-03-11 Thread Christopher Faylor
On Mon, Mar 11, 2002 at 09:38:05PM +0100, Corinna Vinschen wrote: On Sat, Mar 09, 2002 at 07:28:13PM -0500, Pierre A. Humblet wrote: Does RedHat have my copyright assignment after all? We're checking. I come back to you. I'm not clear here. I just checked my records and I've never been

Security patches

2002-03-09 Thread Pierre A. Humblet
Hello Corinna, Attached are 7 diff files, implementing changes discussed last weekend, with two differences: 1) I kept spawn.cc almost intact. I had not considered the possibility of an outside token (from old applications). Also RevertToSelf() is and will remain needed. 2) When a call is made