Errors in pango_parse_markup

2008-03-07 Thread Jürgen Ladstätter
Hi there,

i’m currently working with pango to render fonts to outlines to ps files.
Dont ask me why i do this, i get paid for it :)
I have a rich text field where the user can write something and edit the
style of this text. Then i transform it to markup language. E.g.: span
underline=single weight=bold style=italic foreground=#FF
font_desc=Algerian 24test/span

In die documentation it says: possible values of underline: none,
single, double, low, error
No matter what value I use i get the following error:
'single' is not a valid value for the 'underline' attribute on span tag,
line 1; valid values are 
After are  nothing is appended so i dont know what would be valid! If i
dont set the underline markup tag everything works well.
Do you have any ideas? I'm working with pango 1.18.4

Also I'm facing the problem that the foreground color is never used! When i
render everything via cairo to a png file every single character is black.
No matter what foreground is. So hopefully you have any idea what could bet
he problem.

Here is the src i use:

PangoFontDescription *pFontDescr;
PangoAttrList *pAttrList;
PangoAttribute *pAttr;
GError * pError = 0;
char * pText = 0;
pango_parse_markup ( here the string -1, 0, pAttrList,
pText, NULL, pError) )


/**/
cairo_t *cr;
cairo_surface_t *surface;
PangoLayout *layout;
PangoFontDescription *desc;  

surface = cairo_image_surface_create (CAIRO_FORMAT_ARGB32, 1000,
500);
cr = cairo_create (surface);

//
layout = pango_cairo_create_layout( cr );
pango_layout_set_text( layout, pText, -1 );

cairo_move_to( cr, 100, 300 );  
pango_layout_set_attributes( layout, pAttrList );

pango_cairo_layout_line_path( cr, pango_layout_get_line(layout, 0)
);
cairo_fill( cr );
g_object_unref( layout );
//

cairo_surface_write_to_png (surface, pango_multiple_markup.png);
cairo_destroy( cr );

hope you can find the problem :( i cant.

Kind regards, juergen

___
gtk-app-devel-list mailing list
gtk-app-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list


Re: Gtk demo look and feel

2008-03-07 Thread Tim Janik
On Fri, 7 Mar 2008, Alberto Ruiz wrote:

 Hi all,
 now that we have a new Gtk+ logo and we follow the Tango guidelines,
 wouldn't be a good time to replace the so '90s-ish images from the demo?

 On my last blogpost[0] I demoed the Gdi+ pixbuf loader animation support
 with new images using the Gtk+ logo and tango icons.

 Would it be okay to commit this stuff into trunk?

can you please file a bug report with the patch attached, so
this can go through normal review like everything else?

 [0] http://aruiz.typepad.com/siliconisland/2008/03/gtk-foo.html
 --
 Cheers,
 Alberto Ruiz


---
ciaoTJ
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Scene graph (aka canvas) proposal

2008-03-07 Thread Havoc Pennington
Hi,

Have posted some notes over the last couple years discussing a canvas
widget, based on working on HippoCanvas, and looking over GooCanvas.
Recently I've also been hacking on Clutter, and prototyping a
specialized compositing manager based on Clutter.

I think there's a practical, attainable, and useful path forward for
GTK+ that would give us a mainstream solution for the core problems
these canvas APIs are addressing, and rationalize the relationship
between them and GTK+.

Before the hackfest I thought I'd write up how I think this could be
done. It has implications for GDK, Cairo, etc. in addition to GTK+.

I did the writeup in slide form, these are text-heavy and intended to
be read not presented:

   http://docs.google.com/Presentation?id=dgn6j4pg_50dw7wh6vt

(note: you do not have to create a Google account to read these. Click
open in new window and it will just show the document to you.)

I'll also attach the presentation document exported as a text doc, but
it is not formatted well.

It looks like I can't be at the hackfest though I made a last-minute
effort to figure out how to be there. I am hoping some of the relevant
people will talk about this scene graph idea, though.

Appreciate any feedback or discussion on the list as well. If there's
some good discussion at the hackfest hopefully someone will take some
notes to post.

Havoc

GTK+ Graphics

The Future! Or some ideas anyway
Very rough draft notes

Where do people seek GTK+ alternatives?

Custom visual design of desktop applications (or parts of them)

WebKit

GtkDrawingArea

HippoCanvas

Devices with custom UI (phones, etc.)

Hacked-up GTK+ (theming, other patches)

Clutter

Compositing Manager

Compiz object system

Clutter

Stuff GTK+ makes hard

Pixel-precise control of visual design

Animations and transitions

3D effects: shader effects, rotate in/out, etc.

Nonrectangular or overlapping elements

Embedding in a nonstandard context (GTK+ insists all widgets are inside a 
GtkWindow, not some other canvas or composite overlay window, for example)

Goals

Allow custom visual design, whether for desktop apps or embedded devices; trust 
people to use tool responsibly

Support convenient retained-mode graphics, classic example is Fifteen Game

Support smooth animated transitions and effects, including fast transform of 
images and windows

3D effects including shaders and rotation

Support hardware acceleration for both 2D and 3D in the same layout

Cleanly make use of existing widgets (Entry, Button, TextView, TreeView, etc.) 
and don't create a need to reinvent these

Non-goals

API should support basic 3D effects and use of 3D APIs, but in a somewhat 
2D-centric context; not designed for virtual reality type of UI

Model-view should be for special cases (GtkTextBuffer, GtkTreeModel) but 
generic model-view support is not needed

Not necessary to support untrusted API users (as with X server or Compiz 
objects); too hard/specialized, GtkWidget does not support this anyhow

Canvas APIs also add some conveniences and cleanups such as 
border/padding/alignment for all items, height-for-width, IF_FITS packing, 
etc.; table these features for later work and focus on more fundamental reasons 
people are using a canvas rather than GTK

Proposal

Create a new scene graph API. Make widgets one kind of scene graph object.

Scene graph seems like a less confusing way to describe this than canvas ?

For this presentation, let's call the scene graph objects actors for short, 
rather than canvas item

Scene graph and actors intended to emphasize that the new API will have 
graphics primitives, not interactive UI primitives - GTK+ will continue to 
provide widgets

What's in the scene graph API?

An Actor interface

paint() method which can use either 2D or 3D operations

allocate() method allocates a 2D area

Per-actor transform

Includes scale, rotate, translate

Includes z-axis transformation

Transform is post-allocate/layout but pre-paint

Event picking and bubbling

A Container interface defining the tree of Actor

Layout-and-paint operation: Request, Allocate, Apply Transforms, Paint to 
Buffer, Wait for VSync, Swap Buffer

Replacement for GdkWindow - clipping, scrolling, events

Fixed-position layout manager and interface to get min and natural size request

Effect on GDK

What stays

Toplevel GdkWindow and related operations

Event queue and dispatch of same

Display and Screen objects

What becomes obsolete, replaced by scene graph

Child GdkWindows

Double-buffer approach used by GTK+ may change

What remains obsolete

X11 leakage (colormaps, etc.)

non-Cairo drawing API

Effect on GtkWidget

Implements the Actor interface

GtkContainer becomes somewhat obsolete; walking hierarchy with GtkContainer 
skips all non-widget actors, as it does in HippoCanvas

GtkContainer modified to implement the scene graph's Container interface as well

child GdkWindow deprecated; back compat is complicated, but ideally there's 
some way for stock widgets 

Re: Scene graph (aka canvas) proposal

2008-03-07 Thread Havoc Pennington
Hi,

Some people asked about Pigment; I thought it was all in Python so had
not looked at it. But it turns out the core is in C. So, this is
something else to look at. I did not leave it out on purpose, just
didn't realize it had the C part.

Here is the home page: https://code.fluendo.com/pigment/trac/

Havoc
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list