qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org, shaharh@gmail.com, kwolf@redhat.com
Subject: [Qemu-devel] Re: Qemu does not pass pressed caps lock to client
Date: Fri, 12 Feb 2010 13:39:32 +0100	[thread overview]
Message-ID: <4B754C04.4080907@redhat.com> (raw)
In-Reply-To: <b0ae6fcc1002120309ya8e9160p414e94d9875c7956@mail.gmail.com>

On 02/12/2010 12:09 PM, Shahar Havivi wrote:
> It's not true that SDL is not sending up event like the comment say,
>
> On Fedora 12 it behave like a toggle button, first press/release will send
> caps-down event second press/release send caps-up event
 >
> On Ubuntu 9.10 it work like any other key, i.e. pressing caps will generate two
> events down and up.

True.  To see why, start at 
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=317010 and 
http://archive.ubuntu.com/ubuntu/pool/main/libs/libsdl1.2/libsdl1.2_1.2.13-4ubuntu4.diff.gz 
-- it's a nice code reading exercise, so I'll suggest a possible 
solution before pointing out the reason for this behavior.  The solution 
would be to put hack after hack, i.e. something like this (untested):

     case 0x3a: /* caps lock */
         /* SDL will usually send only 2 events instead of 4, so we
            generate the missing ones.  However, on Debian/Ubuntu
            systems it may generate 4; in this case we have to discard
            the extra events.  On Debian/Ubuntu ev->key.keysym.mod
            will always be zero, but for other systems we need the
            complicated condition below.  */
	if ((ev->key.keysym.mod & KMOD_CAPS) ==
             (ev->type == SDL_KEYDOWN ? KMOD_CAPS : 0)) {
             kbd_put_keycode(keycode);
             kbd_put_keycode(keycode | 0x80);
         }
         return;
     case 0x45: /* num lock */
         /* Same as above.  */
	if ((ev->key.keysym.mod & KMOD_NUM) ==
             (ev->type == SDL_KEYDOWN ? KMOD_NUM : 0)) {
             kbd_put_keycode(keycode);
             kbd_put_keycode(keycode | 0x80);
         }
         return;

Now, the solution of the riddle.  The patch was correctly submitted as

+	SDL_UseLockKeys = getenv ("SDL_DISABLE_LOCK_KEYS") == NULL;

...

+	use_lock_keys = SDL_UseLockKeys;

...

+				if (!use_lock_keys)
+					break;

(i.e. by default do not change anything) but the maintainer apparently 
morphed it into

+	SDL_UseLockKeys = getenv("SDL_DISABLE_LOCK_KEYS");

...

+ 	use_lock_keys = ( SDL_UseLockKeys && *SDL_UseLockKeys );

...

+ 				if ( ! use_lock_keys )
+ 					break;

which changed the default and the meaning of SDL_DISABLE_LOCK_KEYS.

I initially thought about removing the caps lock/num lock hack 
altogether and add the following, however it would need SDL 1.2.14 
because SDL_NO_LOCK_KEYS support was added exactly two months after 
1.2.13 was released. :-( :-(

     /* There are two versions around of a Debian patch that changes the
        way Caps Lock and Num Lock are handled.  The first version
        by default sends only one of the KeyDown/KeyUp events, unless
        SDL_DISABLE_LOCK_KEYS is present in the environment.  The second
        version instead by default sends both events, unless
        SDL_DISABLE_LOCK_KEYS is present and not empty.  This version
        is the most commonly found (and a totally braindead idea).

        Upstream instead supports SDL_NO_LOCK_KEYS which, if set to 1,
        will generate all four events---which is what we want.  Luckily,
        there is a combination of environment variable that will satisfy
        all variant.  */

     putenv ("SDL_DISABLE_LOCK_KEYS", "");
     putenv ("SDL_NO_LOCK_KEYS", "1");

Yes, I love Debian.

Paolo

  parent reply	other threads:[~2010-02-12 12:39 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-02-11 21:13 [Qemu-devel] Qemu does not pass pressed caps lock to client Shahar Havivi
2010-02-12  9:54 ` Kevin Wolf
2010-02-12 11:09   ` Shahar Havivi
2010-02-12 11:43     ` Kevin Wolf
2010-02-12 12:39     ` Paolo Bonzini [this message]
2010-02-12 15:15       ` [Qemu-devel] " Anthony Liguori
2010-02-12 18:17         ` Paolo Bonzini
2010-02-12 18:44           ` Anthony Liguori
2010-02-12 20:49             ` Dustin Kirkland
2010-02-13 11:21               ` Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4B754C04.4080907@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=shaharh@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).