Version française
Home     About     Download     Resources     Contact us    

This site is updated infrequently. For up-to-date information, please visit the new OCaml website at

Browse thread
[Caml-list] Ocamlmakefile and Lablgtk
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: -- (:)
From: Jacques Garrigue <garrigue@k...>
Subject: Re: [Caml-list] Ocamlmakefile and Lablgtk
From: Markus Mottl <>
> Florian Hars schrieb am Dienstag, den 13. November 2001:
> > lablgtk comes as a library file, the problem is that every program that
> > uses the library must properly initialize it before it can use it.
> > gtkInit does just this:
> [snip]
> But why not just link the module in the right order with the rest of
> the library? Then things will be initialized as required. Or are there
> circumstances where you don't want initialization?

This is exactly the reason.
Take for instance Unison. It is linked with lablgtk, but does not want
to initialize GTK+ when used in text mode. This is very important,
otherwise you would need an X-server running everytime you start

While Unison was the original reason, I think it is indeed good
practice to initialize by hand, without linking gtkInit.cmo.  Then you
know explicitely what is called and when. This avoids difficulties
with debugging. Note also that in a real application there are things
you want to do immediately after the initialization (or even before),
and this is hard to do if you don't know where it occurs. On the other
hand, if you forget to initialize, you will get a few dozen warnings,
so this should not be too difficult to detect.

So the rule of the thumb is: use gtkInit.cmo if you are developping
with the toplevel and all your program is included in one file, but
switch to explicit initialization as soon as it gets bigger. This way,
OCamlMakefile should not be a problem :-)

Jacques Garrigue
Bug reports:  FAQ:
To unsubscribe, mail  Archives: