Complete.Org: Mailing Lists: Archives: freeciv-dev: October 2001:
[Freeciv-Dev] Re: [Patch] Extended connect dialog (PR#977)
Home

[Freeciv-Dev] Re: [Patch] Extended connect dialog (PR#977)

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: Andreas Kemnade <akemnade@xxxxxxxxxxx>
Cc: freeciv-dev@xxxxxxxxxxx, bugs@xxxxxxxxxxxxxxxxxxx
Subject: [Freeciv-Dev] Re: [Patch] Extended connect dialog (PR#977)
From: Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>
Date: Fri, 5 Oct 2001 17:02:49 +0200
Reply-to: rf13@xxxxxxxxxxxxxxxxxxxxxx

On Fri, Oct 05, 2001 at 04:33:40PM +0200, Andreas Kemnade wrote:
> Raimar Falke writes:
>  > On Mon, Oct 01, 2001 at 03:21:00PM +0200, Andreas Kemnade wrote:
>  > Content-Description: message body text
>  > > Here is a new version of the patch. 
>  > > I hope all the issues mentioned before are fixed.
>  > > connectdlg_g.h is extended and contains some new functions.
>  > > 
>  > > This patch is only for the gtk client. I will do the stuff for the
>  > > win32 client. 
>  > > I haven't done xaw and mui development for a long time, so I can only
>  > > provide stubs (and call gui_server_connect() in popup_main_menu).
>  > 
>  >  - WNOHANG wasn't defined in client/gui-gtk/connectdlg.c
> Doesn't running autoheader and autoconf solve the problem?

AFAI remember I had.

>  >  - in the connect dialog should be a "Back" button
> Instead of quit or do you want an extra button?

I withdraw my comment. Either we do it (having next/prev buttons) in
all dialogs or we don't do it.

>  >  - I think that the control logic should be in a client/*.c file with
>  >  specific gui depended callbacks. Or put the fork/wait/pipe code into 
> common.
> 
> The problem is the symbol SOCKET_ZERO_ISNT_STDIN. If we put the stuff into
> common, we had to put a lot of #ifdefs into it, too. SIGCHLD has also
> to be tested, which leads to additional #ifdefs. That code would
> not look very nice.

This is IMHO better than copy it all over the place and slightly
change it after the copying.

> start_server_load_game(char *), start_server_for_new_game():
> I haven't put them into client/*.c because gui-win32 and gui-mui
> clients will certainly have a start_server(char *commandline) function
> because the underlying API does not support char **argv.

So is there is a reason to not make start_server(char *commandline)
the default in client/*.c?

> I don't think that simple callbacks should have stuff in client/*.c.
> example:  new_game_callback
>   in connectdlg.c: 
>      static void new_game_callback(GtkWidget *w, gpointer data)
>      {
>        start_new_game();
>      }
>      void start_new_game_gui(void)
>      {
>        input_dialog_create(...);
>      }
>   in client/*.c
>      void start_new_game(void)
>      {
>        start_new_game_gui();
>      }
> 
> I think this is really not sensible. 
> 
> Please explain to me which control logic you would like to have in
> client/*.c.

The real amount which can be extracted can only examined after a
second client is converted. Still missing are command line options
like "-f" and a "--skip-main-menu". The code for these would probably
in client/*.c.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 1 + 1 = 3, for large values of 1


[Prev in Thread] Current Thread [Next in Thread]