|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
Re: ncurses 4.1 security bug
Warner Losh (imp
VILLAGE.ORG)Thu, 9 Jul 1998 14:23:28 -0600
- Messages sorted by: [ date ][ thread ][ subject ][ author ]
- Next message: David Schwartz: "Re: ncurses 4.1 security bug"
- Previous message: Jeffrey Hutzelman: "Re: Linux kernel filesystem oddities"
- In reply to: Alan Cox: "Re: ncurses 4.1 security bug"
- Next in thread: matthew green: "Re: ncurses 4.1 security bug"
In message <m0ytvb6-000aQFCthe-village.bc.nu> Alan Cox writes: : C++ global object constructors are called in pretty much arbitary : order before main() is entererd. That's not entirely correct. C++ global object constructors need to be initialized before they are referenced, even if they are in a dynamically linked in library. This is traditionally done with a call to _main() as the first thing in main(). However, that nit-picking aside, you are correct that you cannot predict when the ctors will be called. : Its an interesting reason not to write setuid apps in C++ 8) Or just don't use global objects that have ctors. It is arguably bad form anyway :-). Well, you could use global objects, so long as they don't need to do privileged things, or carelessly rely on user input.. Warner
- Next message: David Schwartz: "Re: ncurses 4.1 security bug"
- Previous message: Jeffrey Hutzelman: "Re: Linux kernel filesystem oddities"
- In reply to: Alan Cox: "Re: ncurses 4.1 security bug"
- Next in thread: matthew green: "Re: ncurses 4.1 security bug"