[9fans] Page ported to plan9port

[9fans] Page ported to plan9port

[Kris Maglione]

Hi,

I've ported page(1) to plan9port. Since the default ghostscript on most 
systems doesn't support plan9 bitmaps, I have it output windows bitmaps 
and pipe them through bmp(1) (messing with ghostscript isn't worth the 
trouble, unless Russ wants to pester them to put it into the default 
distro).

You can get it here: http://suckless.org/~jg/page.tgz

There are a few things that aren't as well as I'd like:
1. I switched to libthread, but I still have to fork/exec for gs, 
because it needs 5 file descriptors, and threadspawnl allows only 3. I 
looked at the threadspawnl code, and it forks/execs itself, and I 
haven't had any issues.

2. To get the screen size, I used an ugly hack of spawning "xdpyinfo | 
grep dimensions:" and parsing it. I'd like to know if libdraw provides 
another way to get these.

3. There are some conversion commands which specify programs in fb/, 
which exists neither on plan 9, nor in p9p. Does anyone know what/where 
they were?

Re: [9fans] Page ported to plan9port

[erik quanstrom]

those are relics of 8½.

- erik

On Thu Feb  8 18:11:26 EST 2007, bsdaemon@comcast.net wrote:
> 3. There are some conversion commands which specify programs in fb/, 
> which exists neither on plan 9, nor in p9p. Does anyone know what/where 
> they were?

Re: [9fans] Page ported to plan9port

[Russ Cox]

Thanks.

> There are a few things that aren't as well as I'd like:
> 1. I switched to libthread, but I still have to fork/exec for gs,
> because it needs 5 file descriptors, and threadspawnl allows only 3. I
> looked at the threadspawnl code, and it forks/execs itself, and I
> haven't had any issues.

It would be nice to make do with three, since at some
point we'll want to run on other systems, but this is fine for now.

> 2. To get the screen size, I used an ugly hack of spawning "xdpyinfo |
> grep dimensions:" and parsing it. I'd like to know if libdraw provides
> another way to get these.

I should probably add that to the library.  xdpyinfo is fine for now.

> 3. There are some conversion commands which specify programs in fb/,
> which exists neither on plan 9, nor in p9p. Does anyone know what/where
> they were?

These are relics, but not from 8½ -- they're conversion programs
from the second edition.  Don't worry about them.

Russ

Re: [9fans] Page ported to plan9port

[Charles Forsyth]

>Since the default ghostscript on most 
>systems doesn't support plan9 bitmaps, I have it output windows bitmaps 
>and pipe them through bmp(1) (messing with ghostscript isn't worth the 

it supports an "inferno" driver out of the box on ubuntu.
it looks as though it is rgbv.

Re: [9fans] Page ported to plan9port

[Kris Maglione]

On Fri, Feb 09, 2007 at 09:24:55AM +0000, Charles Forsyth wrote:
>>Since the default ghostscript on most 
>>systems doesn't support plan9 bitmaps, I have it output windows bitmaps 
>>and pipe them through bmp(1) (messing with ghostscript isn't worth the 
>
>it supports an "inferno" driver out of the box on ubuntu.
>it looks as though it is rgbv.

Yes, but when I've used the inferno driver, the results have always been 
grainy, for some reason.

Re: [9fans] Page ported to plan9port

[Russ Cox]

> >it supports an "inferno" driver out of the box on ubuntu.
> >it looks as though it is rgbv.
>
> Yes, but when I've used the inferno driver, the results have always been
> grainy, for some reason.

The plan9 driver (in the Plan 9 sources) was tweaked to accomodate
the new full-color model, and in the process I made alpha-blended
output work a bit better.  I suspect that if you added these two
lines from gdevplan9.c to gdevifno.c, then the inferno driver would
look as good for black-and-white text.

	if(pdev->color_info.anti_alias.graphics_bits ||
pdev->color_info.anti_alias.text_bits)
		if(ldepth < 2)
			ldepth = 2;

All that said, using some more standard format is probably the
right answer for page -- I would have chosen png, but bmp is okay.

Russ

Re: [9fans] Page ported to plan9port

[Kris Maglione]

On Fri, Feb 09, 2007 at 12:21:10PM -0500, Russ Cox wrote:
>All that said, using some more standard format is probably the
>right answer for page -- I would have chosen png, but bmp is okay.

I initially chose png, but it turned out that bmp was orders of 
magnitude faster, so I decided to use that instead.

Uriel asks that I suggest that you bug the ghostscript maintainers to 
include gdevplan9.c in their main distros. He'd do it himself, but he's 
still banned.

Re: [9fans] Page ported to plan9port

[Russ Cox]

> Uriel asks that I suggest that you bug the ghostscript maintainers to
> include gdevplan9.c in their main distros. He'd do it himself, but he's
> still banned.

I'd be happy to send it to them for inclusion if
someone else would confirm that it compiles
against the latest gs sources.

Russ

Re: [9fans] Page ported to plan9port

[Kris Maglione]

On Fri, Feb 09, 2007 at 12:52:14PM -0500, Russ Cox wrote:
>I'd be happy to send it to them for inclusion if
>someone else would confirm that it compiles
>against the latest gs sources.

Yes, it does, against AFPL, at least. Uriel tried it with GNU and I seem 
to recall that it compiled, but he got fed up with the build system.

Re: [9fans] Page ported to plan9port

[Russ Cox]

On 2/8/07, Kris Maglione <bsdaemon@comcast.net> wrote:
> I've ported page(1) to plan9port.

I've added this to p9p.

Russ