World: r3wp

DeviceRGB is already added here. Is scale necessary?

Think so. :)

Yes, so your image is there, it's just tiny.

well, it still won't run the PS file.

Use scale 72x72 for 72 dpi.

back in 30 mins.

Henrik, I think, you need another factor. The one you have now, is 
set to 1. It's ok for your text. You have to put scale 72x72 (or 
something) when showing images.

I tried setting your factor to 72, then I can see the image, but 
the text is HUGE! :-)

BTW, if you don't mind: I've added a probe on the dialect code output 
in to-postscript.r. Then you can see if I'm writing something incorrect 
out.

interesting that you can run the file. I can't.

and I've already added scaling

I had a problem some time ago showing PS files, I produced. I think, 
it was some cache.

The dialect code, you produce, looks ok. I'm really not a shark with 
PS, so it doesn't mean, it's 100%, even if it looks ok to me. :)

Best way is to test, test, test, I'm afraid.

I can't get image! datatype to work in the postscript dialect (your 
suggestion). If I have it as a word, that is being "get", it works.

It's because, an image! datatype is the sequence
make image! [ ....

and that is being parsed.

I don't get The 'img variable is set to an image! type, just like 
it is when using load-image from a file or url (which I couldn't 
get to work either).

Sorry, should be:


I don't get it. The 'img variable is set to an image! type, just 
like it is when using load-image from a file! or url! (which I couldn't 
get to work either).

I'm very confused right now. Images now work, but are placed in the 
wrong corner, despite being given the same coordinates as text and 
boxes.

solved the positioning problem

Henrik, did you create a dialect that takes a VID layout and produces 
a printout in PS?

Yes... I know it's 1am for you!

yes, I did

How well does it work?

actually not a dialect, it just converts a View object tree to postscript. 
it's used in the same way as to-image is on a layout.

Kewl

Have you published it?

http://rebol.hmkdesign.dk/files/vid-postscript2.png

http://rebol.hmkdesign.dk/files/to-postscript.r

close ....

if you're doing B/W stuff and not worrying too much about text wrapping, 
then it should work.

since it works on a view object ... it should work on Rebgui as well

no bitmap support and no wrapped text support.

why not?

(which is why it says wednesday twice in one of the fields in the 
postscript output.

the ps dialect supports rebol image format

because that is handled at the View level. text wrapping information 
is not available in the layout tree.

Hope Vid+ doesn't have that limitation

The approach would be similar, but I don't know if that information 
can be extracted. Perhaps it can if Cyphre is pushed hard enough. 
:-)

Pehaps we need some hooks into the layout engine that allows us to 
extract this type of data

eg. we can submit an area to the layout engine and get a graphic 
back at a specified resolution

rich text has some bugs, which cyphre is able to fix once he gets 
time, so I hope this will be possible to hook into. i.e. it's not 
a closed part that we can't touch.

so, we can do the text formmating using VID+ and take it out as a 
graphic

probably better if we can get the xy of each character in the text 
string

that is possible

I'll get to know it more when we get to build a real rich text editor

Doesn't seem to work with layouts from rebgui

>> lo2: display "Test" [ text "hello" ]
>> unview/all
>> type? lo2
== object!
>> to-postscript lo2
** Script Error: Cannot use path on none! value
** Where: ps-face
** Near: if f/font/name [append font-name 'Helvetica]
if find

perhaps the layout tree is different for rebgui

all to-postscript.r does is go through the layout tree and renders 
each object as a postscript color box with a text box on top of it. 
face values are just mapped into postscript.

the first place it barfs in this simple layout is here

        if f/font/name [append font-name 'Helvetica]