Sven Luther <[EMAIL PROTECTED]> writes: >> So you shall not give this pointer which does not reflect what's >> going to be. > > Oh come one, i gave it to Stefano, who was not following all our > previous discution, and told him that this where the being worked on > package, and that there where not (yet) ready.
:-) >> The big problem will probably be ocamlrun and the hability >> to have #!/usr/bin/ocaml-<version>. Is it currently possible? > > No problem, you just modify the script that gets copied into each > bytecode program. I can do that if needed in the debian/rules, but am > looking for a more clean solution, with a configure time switch, which i > can even send upstream if needed. It would be better indeed. >> Same thing will ocamlmklib and such ocaml programs that call >> other ocaml programs. > > Yes, these are the problematic ones. Let me time to look at it and > provide a propper patch. > > Other problems are : > > ocamldoc.sty (a simple alternative should do it). alternative would mean prioritise ocaml versions. > emacs mode (should install in a subdirectory and provide alternatives ?) Hmm, we can put it in a separate package ocaml-mode and only one ocaml-mode would be allowed to be installed at a time. > In the meantime is it more important to fix the remaining bugs and other > problems that stop ocaml from entering testing. Damned testing. >> So I fear that we won't be able to reach our goal. > > Huh, do you believe i will not be able to achieve a nice working > ocaml-3.06 package, which will be parallel installable ? I think it is > possible, and as said, hope even to provide a patch upstream. I will do > it in a clean way though. > > If this fails, no problem i can simply provide a ocaml 3.06-16 package > with the latest fixes that went into ocaml-3.06 3.06-16. In fact i have > such a thing almost ready. I think it would be wise to get our changes blessed by the OCaml team. Cheers, -- Jérôme Marant http://marant.org
