On Mon, 10 Jul 2006, Georges Khaznadar wrote: > Aaron M. Ucko a écrit : > > Indeed, though the problem's actually (as I suspected) a 64-bit > > cleanliness issue, which the attached patch addresses; could you > > please apply it when you get a chance? > > Hello Aaron, > > many thanks for this patch: it gives me a valuable hint for other issues > related to 64-bit cleanliness. I would like to ask you a question: > are the int types the same data types both with 32-bit architectures and > 64-bit architectures, when I use gcc version 4.0.3? Or where may I read > informations about this data type?
You can't rely on the size of an int or a long int. If the code doesn't cope with changes of size on those types, then it needs to use special types harcoding the size. In this case "int" have the same size on i386 and amd64 (32bits) but "long int" don't (i386: 32bits, amd64: 64bits). The correct patch is not to replace those "long int" by "int" but by "int32_t" to ensure the size (for this you need to include the header <stdint.h>). > Please can you check if it can be installed correctly with your architecture? > If you have got no time to do so, I'll try to check it in another way, please > just reply me about it, I am already obliged to you. Georges, can you update the patch accordingly and provide a new version of the source package ? Cheers, -- Raphaël Hertzog Premier livre français sur Debian GNU/Linux : http://www.ouaza.com/livre/admin-debian/
