> > What about implementation in random.cxx and creating random.hxx so all > > the random stuff is in one place? > > Why do you want this abstraction? Following KISS I would just start > implementing the functions in ScInterpreter and extract commonly used > parts into an own method. If at some point other parts of Libreoffice > also need random functions with known statistical properties we can > still extract them.
I'm happy to propose a patch this way. Just need to be aware that RANDBETWEEN() which is part of scaddins would also need a replacement of rand(). Also, another consideration is the way to implement different non-uniform distribution, just wondering if anyone has a strong view on this: (1) implementation similar to rand() and srand(), where the state of the generator is saved in a global variable (which is not exposed via a header file and can only be changed by srand), and all other non-uniform generators draw from the same rand(). this requires an srand() call ideally at the start of libreoffice or calc, currently done in ./sc/source/core/data/global.cxx (2) each rand_nonuniform() function draws from it's own private rand() generator with its own state, which is defined as a static variable and initialised at the first call. For a general purpose library one would probably do (1) (or (1) and encapsulated in a class) but for calc both versions are fine in my opinion. Memory requirements for boost::mt19937 is 625*sizeof(uint32_t) so I guess even if we had 10 copies (one for each non-uniform rand) of it that wouldn't be a problem? _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
