Subject: some notes about XFree86-4.5.0 on NetBSD/sparc
To: NetBSD/sparc Discussion List <port-sparc@NetBSD.org>
From: Greg A. Woods <woods@planix.com>
List: port-sparc
Date: 11/14/2006 15:08:57
--pgp-sign-Multipart_Tue_Nov_14_15:08:51_2006-1
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: quoted-printable

For the record you'll find the detailed xdpyinfo output below from my
first use of XFree86-4.5.0 on NetBSD/sparc (1.6.2_STABLE).

I was pleasantly surprised when it at least ran and displayed a login
window without me touching my /etc/X11/xdm/* files.  However things
rapidly went downhill from there.

I must say I feel like I've stepped back in time, WAY back!  Some of
these are the kinds of problems I remember with X11R3 or earlier.

I can no longer change the key repeat rate:

	$ xset r rate 200 10
	Xlib:  extension "XFree86-Misc" missing on display "very.weird.com:0.0".

The default rate is far too slow for my liking.  Why isn't the
XFree86-Misc extension included in a build of an XFree86 Xserver?
Perhaps the "-ar2" option on the Xserver command line would have the
desired effect, but I won't know until the next time I can logout and
try it.

The screen blanker blanks my screen even while I'm typing or moving the
mouse (though of course it instantly lights up again if I keep typing or
moving the mouse).  That's VERY VERY VERY disconcerting!

Similarly there's some much more prominent and disconcerting window
flashing (windows appear to redraw twice) when I switch between
workspaces when using ctwm.  It's disconcerting because it looks exactly
like an overlaying window has closed upon the switch and an underlying
window is redrawing itself.  This was much smoother and less jarring
before, on the same hardware.

The caps-lock key for a short while turned on the num-lock light, but
now it doesn't toggle any lights after I tried fiddling with the new
(but undocumented) xled toy.  (Note that actually the "old" xsrc/xc
server had the lights all screwed up too, but I don't remember ever
disabling them accidentally.)  (See xdm log entries below too)

The scroll-lock key doesn't do anything.  XKeyCaps suggests it should
produce a Scroll_Lock keysym when pressed, but Emacs see it only as
<S-key-20> and I'm not sure who's wrong, and I don't remember what
happened with the older Xserver.

I can't change the pitch or duration of the display bell, neither with
xbell nor with the traditional "xset b":

	$ xbell 100 100 300 900
	Setting:
	   bell_percent       =3D 100
	   bell_pitch         =3D 300
	   bell_duration      =3D 900
	Have:
	   key_click_percent  =3D 0
	   bell_percent       =3D 100
	   bell_pitch         =3D 300
	   bell_duration      =3D 900
	   led_mask           =3D 0x0
	   global_auto_repeat =3D 1

It always just makes this really wimpy little squeak, and I'm reasonably
certain it was a lot more robust sounding with xsrc/xc, and that it
could be adjusted properly.

"xclock -digital" now dies unless you give it the "-norender" option
(though it sometimes did that before, but without any decent error
message, and that old problem may have been related).

	$ xclock -digital
	Fontconfig error: Cannot load default config file
	Xlib:  extension "RENDER" missing on display "very.weird.com:0.0".
	Warning: Cannot convert string "" to type XftFont
	Memory access violation (core dumped)=20

A freshly compiled Swisswatch doesn't display any face markings or
hands, just a white circle (or black, if I change the colours).  The old
binary running on another machine works fine though.  The new compile
also doesn't start with a default geometry -- one must be supplied
explicitly:

	$ swisswatch
	Error: Shell widget swisswatch has zero width and/or height

They're both built from pkgsrc, swisswatch-0.06nb1, the only difference
being what libraries they're linked against.


The default login manager looks very pretty and all with the nice new
NetBSD logo, even on my monochrome bwtwo hi-res screen, but it doesn't
seem to honour the same Xresources, which I still have set as follows:

xlogin*login.translations: # specific overrides of defaults \
        Ctrl<Key>L: restart-session()\n\
        Ctrl<Key>R: abort-display()\n\
        <Key>F1: set-session-argument(failsafe) finish-field()\n\
        <Key>Left: move-backward-character()\n\
        <Key>Right: move-forward-character()\n\
        <Key>Home: move-to-begining()\n\
        <Key>End: move-to-end()\n\
        Ctrl<Key>KP_Enter: set-session-argument(failsafe) finish-field()\n\
        <Key>KP_Enter: set-session-argument() finish-field()\n\
        Ctrl<Key>Return: set-session-argument(failsafe) finish-field()\n\
        <Key>Return: set-session-argument() finish-field()

As far as I could tell those settings used to work just fine (or was I
dreaming?), but now nothing I pressed gave me a failsafe login, and
nothing killed the session or aborted the Xserver, and for example "F1"
just beeps.  The "Tab" key works to go to the next field, as do the
cursor and home/end keys move back and forth, but that's it.  See below
though....

(and how do I over-ride  CTRL-ALT-BACKSPACE to make that traditional key
sequence work???  I can't find an example of expressing two modifiers.)

The xdm log has weird and possibly related stuff in it that I've not
identified the source of yet either:

--------------------
The XKEYBOARD keymap compiler (xkbcomp) reports:
> Error:            Cannot open "compiled/server-0.xkm" to write keyboard d=
escription
>                   Exiting
(EE) Couldn't load XKB keymap, falling back to pre-XKB keymap
--------------------

and:

--------------------
Warning: translation table syntax error: Unknown modifier name:   specific
Warning: ... found while parsing 'specific overrides of defaults        Ctr=
l<Key>L: restart-session()'
Warning: String to TranslationTable conversion encountered errors
--------------------

So much for compatability and even documentation:

            xlogin*login.translations: #override\
                 Ctrl<Key>R: abort-display()\n\/&
                 <Key>F1: set-session-argument(failsafe) finish-field()\n\

XFree86                   Version 4.5.0                        16

XDM(1)                                                     XDM(1)

                 <Key>Return: set-session-argument() finish-field()
            xlogin*borderWidth: 3
            xlogin*greeting: CLIENTHOST
            #ifdef COLOR
            xlogin*greetColor: CadetBlue
            xlogin*failColor: red
            #endif

I'll try and remove what I thought was a comment, or rather change it
back to exactly what's shown in the documentation, but I'm having
trouble finding the code that's supposed to parse that table.


and also:

--------------------
Failed to set keyboard lights: Interrupted system callFailed to set keyboar=
d lights: Interrupted system callFailed to set keyboard lights: Interrupted=
 system callFailed to set keyboard lights: Interrupted system call
--------------------
Hmmm....  I wonder....


$ xdpyinfo -ext all
name of display:    very.weird.com:0.0
version number:    11.0
vendor string:    The XFree86 Project, Inc
vendor release number:    40500000
XFree86 version: 4.5.0
maximum request size:  16777212 bytes
motion buffer size:  256
bitmap unit, bit order, padding:    32, MSBFirst, 32
image byte order:    MSBFirst
number of supported pixmap formats:    1
supported pixmap formats:
    depth 1, bits_per_pixel 1, scanline_pad 32
keycode range:    minimum 8, maximum 132
focus:  window 0x40000e, revert to PointerRoot
number of extensions:    20
    BIG-REQUESTS
    DEC-XTRAP
    DOUBLE-BUFFER
    DPMS
    Extended-Visual-Information
    FontCache
    LBX
    MIT-SCREEN-SAVER
    MIT-SHM
    MIT-SUNDRY-NONSTANDARD
    RECORD
    SECURITY
    SHAPE
    SYNC
    TOG-CUP
    X-Resource
    XC-APPGROUP
    XC-MISC
    XKEYBOARD
    XTEST
default screen number:    0
number of screens:    1

screen #0:
  dimensions:    1600x1280 pixels (452x361 millimeters)
  resolution:    90x90 dots per inch
  depths (1):    1
  root window id:    0x24
  depth of root window:    1 plane
  number of colormaps:    minimum 1, maximum 1
  default colormap:    0x21
  default number of colormap cells:    2
  preallocated pixels:    black 1, white 0
  options:    backing-store NO, save-unders NO
  largest cursor:    1600x1280
  current input event mask:    0xd0001d
    KeyPressMask             ButtonPressMask          ButtonReleaseMask    =
   =20
    EnterWindowMask          SubstructureRedirectMask PropertyChangeMask   =
   =20
    ColormapChangeMask      =20
  number of visuals:    1
  default visual id:  0x20
  visual:
    visual id:    0x20
    class:    StaticGray
    depth:    1 plane
    available colormap entries:    2
    red, green, blue masks:    0x0, 0x0, 0x0
    significant bits in color specification:    1 bits

MIT-SHM version 1.1 opcode: 129, base event: 65, base error: 128
  shared pixmaps: yes, format: 2

XKEYBOARD version 1.0 opcode: 137, base event: 70, base error: 140

SHAPE version 1.0 opcode: 128, base event: 64

SYNC version 3.0 opcode: 136, base event: 68, base error: 138
  system counters: 1
    SERVERTIME  id: 0x00000022  resolution_lo: 4  resolution_hi: 0

Xlib:  extension "XFree86-VidModeExtension" missing on display "very.weird.=
com:0.0".
XFree86-VidModeExtension extension not supported by server

Xlib:  extension "XFree86-Misc" missing on display "very.weird.com:0.0".
XFree86-Misc extension not supported by server

XTEST version 2.2 opcode: 131

DOUBLE-BUFFER version 1.0 opcode: 141, base error: 143
  Double-buffered visuals on screen 0
    visual id 0x20  depth 1  perflevel 0

RECORD version 1.13 opcode: 139, base error: 141

Xlib:  extension "XInputExtension" missing on display "very.weird.com:0.0".
XInputExtension extension not supported by server

RENDER extension not supported by server

Xlib:  extension "XINERAMA" missing on display "very.weird.com:0.0".
XINERAMA extension not supported by server

DMX extension not supported by server


I hope, but I don't expect, xorg will be back to the present at least.  :-)

--=20
						Greg A. Woods

H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack <woods@robohack.ca>
Planix, Inc. <woods@planix.com>       Secrets of the Weird <woods@weird.com>

--pgp-sign-Multipart_Tue_Nov_14_15:08:51_2006-1
Content-Type: application/pgp-signature
Content-Transfer-Encoding: 7bit

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
MessageID: apeGOEwqawU+X7A8JnzJbjr8l3p6AJzi

iQA/AwUBRVoiWGZ9cbd4v/R/EQKz2wCeP1yICgE1GgZRAP9vhKZFQp5RrJgAoLHE
lERjwh776T//fPPjW1zUCZP8
=4+hO
-----END PGP SIGNATURE-----

--pgp-sign-Multipart_Tue_Nov_14_15:08:51_2006-1--