Re: Re[2]: [Mldonkey-users] pango-20021128b

[Thomas de Grenier de Latour]

On Thu, 28 Nov 2002 17:54:22 +0100 (CET)
address@hidden wrote:

> Here is what I suspect:
>  - if the option doesn't exist, then an md4 is created, but
> the option_hook is not called, so it's okay
>  - if the option exists, option_hooks is called and does a
> "=:=", which calls option_hooks, and it loops again and
> again. So the hook should be a conditional statment that
> only modify the value if it was not correct according to the
> new signature.
> 
> I'm sorry, I can't test anything now because I'm not at
> home. But if I'm right, it will be easy to fix.
> 

I'm back home, and here is the patch (well, not really a patch, but
something to apply by hand in common/commonOptions.ml):

-  option_hook client_md4 (fun _ -> client_md4 =:= mldonkey_md4
!!client_md4)
+  option_hook client_md4 (fun _ -> client_md4!! <- mldonkey_md4
!!client_md4)

Seems to work here... At least, it compiles and launch fine with my old
downloads.ini. My client_md4 has been changed on two bytes. Is that the
expected behavior ?

And sorry for the late submitting, in fact I've already posted a patch
as a reply to myself and not to the list, and then I've again replied to
myself, etc. :)

-- Thomas.