Raul Miller <[EMAIL PROTECTED]> wrote: > So? When the program is running, both the GPLed code and the Qt code > exist together in the same virtual memory image. > > Ultimately, there's no difference between run-time linking and > compile-time linking except that run-time linking happens at run time > while compile-time linking happens at compile time.
And run-time is post-distribution while compile-time is pre-distribution, in the case of distribution of dynamically linked binaries. And no copyright-based licence has anything to say about what end users can do with the code thay legally obtained (except redistribution). Despite what some licences pretend (e.g., QPL section 5). Nobody is going to distribute a post-dynamic-linkage virtual memory image, so its composition is rather irrelevant. If a dynamically linked object file contains no actual code from the library (which I am assuming; I'm no expert at this) then the copyright of the library does not directly affect the object file's distribution any more than the copyright of a web page I am linking to affects the (web-)distribution of my web pages. The copyright of the library (sources) may still come into scope indirectly via section 2 of the GPL (or some similar clause of a different licence, if that applies), but that is a slightly different issue that I don't intend to discuss here. The point is, there is a good basis for making the distinction between static and dynamic linkage, in some cases. By the way, I assume that Microsoft does not forbid distribution of binaries for programs that run under MS Windows (that would certainly decrease the popularity of their platform). Is this because they explicitly gave permission, or simply because their permission is not required? I honestly don't know, but I would bet it is the second possibilty. Does anybody have more definite information on such issues? Marc van Leeuwen Universite de Poitiers http://wwwmathlabo.univ-poitiers.fr/~maavl/
