From: zyx z...@gmx.us<mailto:z...@gmx.us>
> I guess you know threads are here much much much longer than C++11 and > even PoDoFo itself has compile time options to enable thread safety > using other means than C++, do you? I know it does things, and I assume these are going to work in practice, with the compilers tested. But I also know that there is not, and has never been, any guarantee beyond what individual compilers may promise. > I agree > that C++ "wrappers" can make life easier, I only do not agree that it's > a requirement to make things work. Compilers are free to make optimization under the “as-if” rule. Under C++03, that never included “watch out for other threads accessing the same data,” and compilers were allowed to make optimizations (or other changed to the code) that are not valid in a multithreaded situation. Again, I am aware that multiple compiler vendors made promises, not always the same ones, about not breaking multithreaded code in certain situations, but the introduction of standard thread support in C++11 went far beyond “wrappers.” The memory model started making portable thread-safe code possible. Does the current threading support in PoDoFo contain thread-local storage? I only found mutexes, and as I argued before, those are not the right tool for the recursion guard. Cheers, Christopher The MathWorks GmbH | Friedlandstr.18 | 52064 Aachen | District Court Aachen | HRB 8082 | Managing Directors: Bertrand Dissler, Steven D. Barbo, Jeanne O’Keefe
_______________________________________________ Podofo-users mailing list Podofo-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/podofo-users
