[9fans] Equis

[9fans] Equis

[Federico G. Benavento]

lotte% contrib/list -v fgb/X11
fgb/X11:
	Description:
		Equis - Xserver for Plan 9 + Xlibs.
	Contents: 115.40Mb in 6023 files
	Modified: Tue May 6 14:42:48 GMT 2008
	Depends:  freetype z
lotte%

if you are curious to see how it's the done, equis' specific code is in
/n/sources/contrib/fgb/root/sys/src/ape/X11/cmd/X/hw/equis

ah, and if you don't want to grab the full source (which is quite big)
you can just fcp the binary /n/sources/contrib/fgb/root/386/bin/equis
you'd also need some fonts in /sys/lib/ape/X11/fonts

enjoy!

Federico G. Benavento

---
/bin/fortune:
Program in place to shed costs

Re: [9fans] Equis

[Pietro Gagliardi]

/n/sources/contrib/pietro/fgbx11.errors

And after that, how do I run? I get some errors and a big X after

	X11/equis

On May 6, 2008, at 10:55 AM, Federico G. Benavento wrote:

> lotte% contrib/list -v fgb/X11
> fgb/X11:
> 	Description:
> 		Equis - Xserver for Plan 9 + Xlibs.
> 	Contents: 115.40Mb in 6023 files
> 	Modified: Tue May 6 14:42:48 GMT 2008
> 	Depends:  freetype z
> lotte%
>
> if you are curious to see how it's the done, equis' specific code is
> in
> /n/sources/contrib/fgb/root/sys/src/ape/X11/cmd/X/hw/equis
>
> ah, and if you don't want to grab the full source (which is quite big)
> you can just fcp the binary /n/sources/contrib/fgb/root/386/bin/equis
> you'd also need some fonts in /sys/lib/ape/X11/fonts
>
> enjoy!
>
> Federico G. Benavento
>
> ---
> /bin/fortune:
> Program in place to shed costs
>
>

Re: [9fans] Equis

[Federico G. Benavento]

hmm... I don't get those errors...

you run it just like any other X server, equis -help is a bit verbose
I run it:
% equis -ac


On Tue, May 6, 2008 at 5:14 PM, Pietro Gagliardi <pietro10@mac.com> wrote:
> /n/sources/contrib/pietro/fgbx11.errors
>
>  And after that, how do I run? I get some errors and a big X after
>
>         X11/equis
>
>
>
>  On May 6, 2008, at 10:55 AM, Federico G. Benavento wrote:
>
>
> > lotte% contrib/list -v fgb/X11
> > fgb/X11:
> >        Description:
> >                Equis - Xserver for Plan 9 + Xlibs.
> >        Contents: 115.40Mb in 6023 files
> >        Modified: Tue May 6 14:42:48 GMT 2008
> >        Depends:  freetype z
> > lotte%
> >
> > if you are curious to see how it's the done, equis' specific code is in
> > /n/sources/contrib/fgb/root/sys/src/ape/X11/cmd/X/hw/equis
> >
> > ah, and if you don't want to grab the full source (which is quite big)
> > you can just fcp the binary /n/sources/contrib/fgb/root/386/bin/equis
> > you'd also need some fonts in /sys/lib/ape/X11/fonts
> >
> > enjoy!
> >
> > Federico G. Benavento
> >
> > ---
> > /bin/fortune:
> > Program in place to shed costs
> >
> >
> >
>
>
>



--
Federico G. Benavento

Re: [9fans] Equis

[john]

> hmm... I don't get those errors...
>
> you run it just like any other X server, equis -help is a bit verbose
> I run it:
> % equis -ac
>
>

I'd just like to comment...  After a loooooong contrib/install, I got
equis running, launched an xterm from my Linux box, and now I'm
running twm and Firefox on the Plan 9 box.  Rio is no longer running
on said box.  X is pretty fast considering that it's going over the
network...  good job Federico!  I get asked all the time "Does it have
X?" and while I will still answer "Plan 9 is based on a different
graphics system", I'll now be able to add, "but we have X11 too".


I would post a screenshot but it looks just like any Linux box running
TWM.


John

Re: [9fans] Equis

[Federico G. Benavento]

oh, nice bug you have there, what's the screen depth when that
happens?

> neat.
>
> when I drawterm from a mac to a plan 9 box it works fine.
> (once I realized that I can only use a DISPLAY of :0
>  if localhost is a known system name,  which it wasn't on my system)
>
> when I run it in plan 9 in parallels on the mac it looks a bit funny,
> but if you learned to read and write slanted letters at school
> everything is ok :-)
>
> http://plan9.cs.utwente.nl/equis-parls.png
>
> I have not investigated yet; do you have an idea?
> anything I could/should try?
>
> Axel.

Re: [9fans] Equis

[Axel Belinfante]

neat.

when I drawterm from a mac to a plan 9 box it works fine.
(once I realized that I can only use a DISPLAY of :0
 if localhost is a known system name,  which it wasn't on my system)

when I run it in plan 9 in parallels on the mac it looks a bit funny,
but if you learned to read and write slanted letters at school
everything is ok :-)

http://plan9.cs.utwente.nl/equis-parls.png

I have not investigated yet; do you have an idea?
anything I could/should try?

Axel.

Re: [9fans] Equis

[Federico G. Benavento]

> when I run it in plan 9 in parallels on the mac it looks a bit funny,
> but if you learned to read and write slanted letters at school
> everything is ok :-)

this is weird, I guess it might be something wrong with the way I use
bytesperline(screen->r, screen->depth), because this only happens
with 16 bpp and with certain windows sizes, that's why I was puzzled,
because it doesn't happen always!

I'll investigate

Re: [9fans] Equis

[Federico G. Benavento]

> I'm having strange problems with wobbling windows and my desktop is
> rotating like a cube. How can that happen?

sorry, but I have no idea what you're talking about, what desktop?
rotating?

> I hope there won't be too much linux apps ported to plan9 in the feature...

I don't get this either.

Re: [9fans] Equis and parallels on mac (long!)

[Axel Belinfante]

> oh, nice bug you have there, what's the screen depth when that
> happens?

I feel stupid - how do I check screen depth?
ehm... I added xdpyinfo output at the end.

anyway, the main issue is not screen depth.

(ok. I see somewhat different colors for the twm bars
between parallels and drawterm, using same twm.
this might be creen depth related)

the main issue could very well be an issue in parallels.

it has something to do, not with when equis is started,
but with when a window is created in rio - whether
parallels is running full screen, or in a big window
(screen-sized) with top-left corner not at top-left of display.

when I run full-screen parallels, create a rio window there,
start equis, all works fine.
http://plan9.cs.utwente.nl/equis-parls2.png

when I have screen-sized parallels window running as window
among other mac windows, with top-left corner somewhere in
the middle of the screen, then create rio window and start
equis I get the slanted stuff.

I tried various times, and the issue is indeed about the
situation at the moment the rio window is created.
that situation seems to be rembered as part of the rio window:
once a rio window is created 'the wrong way', it does not help
to make the parallels window full-screen before starting equis.

it might be possible to see the difference in properties
of the created rio window, if I would known where to look.

when I put the parallels window non-full screen, but with
left border at left border of display, then create a rio window,
then make it full screen again and start equis it worked fine too.

when I move the parallels window to the right, create a rio window,
go full screen and start equis I get the slanted lines.

conclusion:
somehow the horizontal offset between left corner of parallels window
and left corner of display confuses parallels, or plan 9,
or the combination.
also, with the parallels window left-top corner in the middle
of the display, as soon as the fat plan 9 mouse cursor hits the
bottom of the display (or the off-screen bottom of the parallels window?)
a second small mouse cursor (the mac one) appears,
with an offset from the fat plan 9 one. which one is leftmost
depends on the horizontal position of the plan 9 cursor when it
hits the bottom (of display or parallels window, I don't know)
the second cursor disappears again either when it is the first
to go beyond the top or left border of the parallels window,
or when the plan 9 cursor hits such border first, then further
mouse movement pushes the mac mouse cursor to the plan 9 one
till they coincide and the mac mouse cursor disappears.

there is also some interaction with stuff in the dock.
when I see the two mouse cursors and they go low enough
the dock that was hidden at the bottom of the display unhides.
If I only see one mouse cursor that does not happen.

===== drawterm =========================================
bash-3.2$ name of display:    slurp:0.0
version number:    11.0
vendor string:    The X.Org Foundation
vendor release number:    6600
X.Org version: 0.0.6.600
maximum request size:  262140 bytes
motion buffer size:  256
bitmap unit, bit order, padding:    32, LSBFirst, 32
image byte order:    LSBFirst
number of supported pixmap formats:    5
supported pixmap formats:
    depth 1, bits_per_pixel 1, scanline_pad 32
    depth 8, bits_per_pixel 8, scanline_pad 32
    depth 16, bits_per_pixel 16, scanline_pad 32
    depth 24, bits_per_pixel 24, scanline_pad 32
    depth 32, bits_per_pixel 32, scanline_pad 32
keycode range:    minimum 8, maximum 255
focus:  window 0x60000a, revert to PointerRoot
number of extensions:    4
    RENDER
    SHAPE
    X-Resource
    XInputExtension
default screen number:    0
number of screens:    1

screen #0:
  print screen:    no
  dimensions:    837x575 pixels (212x146 millimeters)
  resolution:    100x100 dots per inch
  depths (6):    1, 8, 16, 24, 32, 32
  root window id:    0x35
  depth of root window:    32 planes
  number of colormaps:    minimum 1, maximum 1
  default colormap:    0x20
  default number of colormap cells:    2048
  preallocated pixels:    black 0, white 16777215
  options:    backing-store NO, save-unders NO
  largest cursor:    837x575
  current input event mask:    0xd0001d
    KeyPressMask             ButtonPressMask          ButtonReleaseMask
    EnterWindowMask          SubstructureRedirectMask PropertyChangeMask
    ColormapChangeMask
  number of visuals:    1
  default visual id:  0x21
  visual:
    visual id:    0x21
    class:    TrueColor
    depth:    32 planes
    available colormap entries:    2048 per subfield
    red, green, blue masks:    0xff0000, 0xff00, 0xff
    significant bits in color specification:    8 bits

==================================================

===== parallels ===============================
name of display:    10.211.55.3:0.0
version number:    11.0
vendor string:    The X.Org Foundation
vendor release number:    6600
X.Org version: 0.0.6.600
maximum request size:  262140 bytes
motion buffer size:  256
bitmap unit, bit order, padding:    32, LSBFirst, 32
image byte order:    LSBFirst
number of supported pixmap formats:    5
supported pixmap formats:
    depth 1, bits_per_pixel 1, scanline_pad 32
    depth 8, bits_per_pixel 8, scanline_pad 32
    depth 16, bits_per_pixel 16, scanline_pad 32
    depth 24, bits_per_pixel 24, scanline_pad 32
    depth 32, bits_per_pixel 32, scanline_pad 32
keycode range:    minimum 8, maximum 255
focus:  PointerRoot
number of extensions:    4
    RENDER
    SHAPE
    X-Resource
    XInputExtension
default screen number:    0
number of screens:    1

screen #0:
  print screen:    no
  dimensions:    604x494 pixels (153x125 millimeters)
  resolution:    100x100 dots per inch
  depths (6):    1, 8, 16, 24, 32, 16
  root window id:    0x35
  depth of root window:    16 planes
  number of colormaps:    minimum 1, maximum 1
  default colormap:    0x20
  default number of colormap cells:    64
  preallocated pixels:    black 0, white 65535
  options:    backing-store NO, save-unders NO
  largest cursor:    604x494
  current input event mask:    0x0
  number of visuals:    1
  default visual id:  0x21
  visual:
    visual id:    0x21
    class:    TrueColor
    depth:    16 planes
    available colormap entries:    64 per subfield
    red, green, blue masks:    0xf800, 0x7e0, 0x1f
    significant bits in color specification:    8 bits

=====================================================

Re: [9fans] Equis

[Federico G. Benavento]

> I have run into the same thing running Plan 9 natively on 16-bit color
> mode.  It seems to be partially a function of the window size/shape
> and partially a result of some perverse gremlin somewhere in my
> machine; if I keep killing and restarting equis while twiddling window
> size, I can eventually get a working display.

it has to do with the reminder of some division, I'm fixing it, let me test it
and I'll push it.

thanks

Re: [9fans] Equis

[hiro]

I'm having strange problems with wobbling windows and my desktop is
rotating like a cube. How can that happen?
I hope there won't be too much linux apps ported to plan9 in the feature...

Re: [9fans] Equis and parallels on mac (long!)

[Federico G. Benavento]

> depth: 16 planes

thanks, I'm at it.

--
Federico G. Benavento

Re: [9fans] Equis

[andrey mirtchovski]

>  sorry, but I have no idea what you're talking about, what desktop?
>  rotating?

those are "metacity"'s window effects. gnome's window manager window
manager. don't ask how i know :)

Re: [9fans] Equis

[john]

>> when I run it in plan 9 in parallels on the mac it looks a bit funny,
>> but if you learned to read and write slanted letters at school
>> everything is ok :-)
>
> this is weird, I guess it might be something wrong with the way I use
> bytesperline(screen->r, screen->depth), because this only happens
> with 16 bpp and with certain windows sizes, that's why I was puzzled,
> because it doesn't happen always!
>
> I'll investigate

I have run into the same thing running Plan 9 natively on 16-bit color
mode.  It seems to be partially a function of the window size/shape
and partially a result of some perverse gremlin somewhere in my
machine; if I keep killing and restarting equis while twiddling window
size, I can eventually get a working display.


John

Re: [9fans] Equis

[Federico G. Benavento]

ok, there is a new binary on sources which hopefully will fix this issue,
it also wont try to restart when rio sends equis a hangup

thanks for the bug reports :)

On Wed, May 7, 2008 at 12:15 PM, Federico G. Benavento
<benavento@gmail.com> wrote:
> > I have run into the same thing running Plan 9 natively on 16-bit color
>  > mode.  It seems to be partially a function of the window size/shape
>  > and partially a result of some perverse gremlin somewhere in my
>  > machine; if I keep killing and restarting equis while twiddling window
>  > size, I can eventually get a working display.
>
>  it has to do with the reminder of some division, I'm fixing it, let me test it
>  and I'll push it.
>
>  thanks
>
>



--
Federico G. Benavento

Re: [9fans] Equis

[hiro]

>  thanks for the bug reports :)

You're welcome.

>  > I hope there won't be too much linux apps ported to plan9 in the feature...
>
>
> I don't get this either.

I intended to write future instead of feature, although it could be
feature future.

You're a good man:)
Perhaps my fear is not appropriate yet. Perhaps it evolves the right way.

Re: [9fans] Equis

[Federico G. Benavento]

I guess, you don't get it, I'm the one that tries to live in a Plan 9
exclusive environment, I only have one machine, which means
I have to reboot everytime I have to use a real web browser, before
abaco my life was completely different, so I either run Plan 9 in
vmware or try to make it play the better possible with the rest of
the world.

hah, I don't even have a lunix here, I installed linux in 2002, but that
was long ago.


On Wed, May 7, 2008 at 9:21 PM, hiro <23hiro@googlemail.com> wrote:
> >  thanks for the bug reports :)
>
>  You're welcome.
>
>
>  >  > I hope there won't be too much linux apps ported to plan9 in the feature...
>  >
>  >
>  > I don't get this either.
>
>  I intended to write future instead of feature, although it could be
>  feature future.
>
>  You're a good man:)
>  Perhaps my fear is not appropriate yet. Perhaps it evolves the right way.
>
>



--
Federico G. Benavento

Re: [9fans] Equis

[hiro]

Oh, so you're not only a good man, but also very classy. Can't say
that of myself.
I don't want to malign your great work, but I rather find this
interesting for myself,  since I could swap my old plan9 machine with
my main machine now:)