Browse thread
[Caml-list] Compiler ActiveDVI (CVS)
[
Home
]
[ Index:
by date
|
by threads
]
[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
| Date: | -- (:) |
| From: | Jacques Garrigue <garrigue@k...> |
| Subject: | Re: [Caml-list] Compiler ActiveDVI (CVS) |
Hello Ian, > Jacques> If you start to mess around with the directories, this is not > Jacques> very helpful for anybody. > > Yes, it is: to people with experience on Debian, but none on Ocaml. > Which is exactly the sort of people that would install a binary > package, I think. The point is that the "right" way to compile an extension to ocaml is to use ocamlc, which will automatically add /usr/lib/ocaml to the include path. If the caml subdirectory is not there, this will not be very useful. Using ocamlc will also set properly some compilation flags for thread compliance, and make sure you can build a dll. Since these headers will preferably _not_ be used by gcc directly, they don't need to be in the standard hierarchy. Another problem is that both ocaml's (unfortunately limited) documentation, and ocaml developpers base themselves on a standard installation. If you start messing around, you end-up with the kind of quiproquo we are seeing here. > The way I did it when I built Ocaml from sources was with a symlink, > so that it existed in both places. A symlink avoids all these problems. And if you do it yourself, you know that you're doing something nonstandard, and will no bother the developpers with this. > BTW, /usr/include/ocaml (to match the package name) would be a better > name. That seems a good idea too. By the way, all my comments do not intend to criticize the work of packagers. I know this is a hard task. I just hope we could end up with a better way to combine efforts. Jacques Garrigue ------------------- Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/ To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr