Shachar Shemesh <[EMAIL PROTECTED]> writes:

> Because of you, I actually opened up the wine code that parses this
> thing. If you pull the wine source you will find it in
> server/registry.c, in a function called "load_key". The line you are
> looking for is "if (sscanf( buffer + res, " %d", &modif ) != 1) modif
> = default_modif;". In other words, leaving this number out altogether
> will default to a reasonable number, and this is, in general, the
> modification time of the key (seems to be a simple time_t).

Thanks! This is way more effort than I could have expected!

> 
> Now, the answer you won't like. Messing with this file is risky. You
> have to make sure no wine process (including wineserver) is up while
> you do so. 

In general, I realize that. What seems to be going on on this system
though is that a particular app, on invocation, copies a system.reg
that lives in /opt/<app>/wine-config/etc (*not* in /etc/wine) to the
user's $HOME/.<app>/.wine, and I only operate in
$HOME/.<app>/.wine/system.reg. The app is a singleton, i.e. only one
instance (per user) can run at a time.

Why it is done this way I have no idea. I am only hacking the bloody
thing to make the file associations behave in a sane manner (don't
ask). I definitely don't want to learn winelib or regedit for this
purpose. Hopefully I'll do a working prototype and let others bother
about foolproof implementation.

I wonder though if the fact that the system behaves as described makes
the situation a bit less perilous.

Thanks again for your help,

-- 
Oleg Goldshmidt | [EMAIL PROTECTED]

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to