On Tue, Oct 13, 2009 at 08:52:41PM +0000, Sylvain Le Gall wrote: > > I think you're cheating with this example, because a change in the OCaml > > compiler can pretty much change everything, and that's exactly why (also > > *before* dh_ocaml) we were keeping versioned dependencies on the ABI of > > OCaml itself. > > I am not cheating ;-) > > The decision we took of versioned dependencies is a "safe guard". Some > people can argue that we are too much safe. In certain case, it should > be possible to run a bytecode exec with newer OCaml version. This, AFAICT, has nothing to do with the context you provided your example in. I was arguing that dynamic loading of .so should have the same checks (at the OCaml loader level) that are performed by the OCaml linker at (static) link time. If there were a way to extract the information on the basis of which such (currently totally hypothetical) test, than we could use them to extract proper ABI checksums. Your example simply explains why we have a dependency between bytecode executables to specific OCaml runtime versions which (again, AFAICT), has nothing to do with the above reasoning. Even if there were a way to extract the checksum, for sure we would not remove the specific dependency on a given OCaml version. Cheers. -- Stefano Zacchiroli -o- PhD in Computer Science \ PostDoc @ Univ. Paris 7 zack@{upsilon.cc,pps.jussieu.fr,debian.org} -<>- http://upsilon.cc/zack/ Dietro un grande uomo c'è ..| . |. Et ne m'en veux pas si je te tutoie sempre uno zaino ...........| ..: |.... Je dis tu à tous ceux que j'aime
Attachment:
signature.asc
Description: Digital signature