Native FBInk: Another small tool to print crap on your screen!

[NiLuJe]

@PoP: Yep, not in a tagged release yet, hence why I said "master" .

There's a binary in this unrelated issue if you want to play with it early.

(I'm waiting for the dust to settle around the new Kobo Forma before a new release).

[PoP]

Thanks, not familiar with tag/master

Nothing to worry I can wait

On the K3:

[root@kindle bin]# pwd
/mnt/us/usbnet/bin
[root@kindle bin]# fbink
Segmentation fault

[NiLuJe]

Ah, yeah, that would be because it's a K4/K5 binary, oops .

Here's the full set .

[PoP]

Voilà!



...gives me all kind of crazy ideas for texturing fonts

[PoP]

Originally Posted by PoP

@NiLuJe... Oh wait I now see that you are working on an fbink API to draw an image from memory now write text using Open Type with @shermp resulting in ... small, scalable, proportional, and moreover supporting any font!!

Originally Posted by NiLuJe

Here's the full set .

Could future releases also include the static target device libfbink.a to play with?

[NiLuJe]

@PoP: The static lib? Wouldn't a shared one be more useful?

[PoP]

^
I (by incompetence) can't appreciate the pros or cons , OK libfbink.so then, just a binary one can link to without having to build.

[NiLuJe]

You'd still need to build *something* to make use of either, though .

(Or, I, guess, *link* something if you went the hacky way with static objects).

[PoP]

Just to play, I wanted to include:

cd /mnt/us/tcc
tcc -Iinclude -Iinclude/freetype2/freetype -Iinclude/freetype2 -Iinclude/fbink -L/mnt/us/linkfonts/lib/autohint -lfreetype -L/mnt/us/usbnet/bin -lfbink -o  fbinkpantspatrol fbinkpantspatrol.c

and see if I could call fbink to draw the FT glyphs,

#include <stdio.h>
#include <string.h>
#include <stdlib.h>

#include <ft2build.h>
#include FT_FREETYPE_H

#include <fbink.h>

  int main( )
  {
    FT_Library       font_library;
    FT_Face          font_face;
    FT_Bitmap        bitmap;
    FT_GlyphSlot     cur_glyph;
    FT_Glyph_Metrics glyph_metrics;

    int  glyph_ind;
    int  num_chars;
    char char_name[256];
    
    int x;

// What a FBInk config should look like. Perfectly sane when fully zero-initialized.
//struct FBIink_config {
// 	short int row;             // y axis (i.e., line), counts down from the bottom of the screen if negative
//	short int col;             // x axis (i.e., column), counts down from the right edge of the screen if negative
//	uint8_t   fontmult;        // Font scaling multiplier (i.e., 4 -> x4), 0 means automatic.
//	uint8_t   fontname;        // Request a specific font (c.f., FONT_INDEX_T enum)
// 	bool      is_inverted;     // Invert colors
//	bool      is_flashing;     // Request a black flash on refresh
//	bool      is_cleared;      // Clear the screen beforehand (honors is_inverted)
// 	bool      is_centered;     // Center the text (horizontally)
//	short int hoffset;         // Horizontal offset (in pixels) for text position
// 	short int voffset;         // Vertical offset (in pixels) for text position
//	bool      is_halfway;      // Vertically center the text, honoring row offsets
// 	bool      is_padded;       // Pad the text with blanks (on the left, or on both sides if is_centered)
// 	uint8_t   fg_color;        // Requested foreground color for text (c.f., FG_COLOR_INDEX_T enum)
// 	uint8_t   bg_color;        // Requested background color for text (c.f., BG_COLOR_INDEX_T enum)
//	bool      is_overlay;      // Don't draw bg, use inverse of fb's underlying pixel as pen fg color
//	bool      is_bgless;       // Don't draw bg (mutually exclusive with is_overlay)
// 	bool      is_fgless;       // Don't draw fg (takes precendence over is_overlay/is_bgless)
// 	bool      is_quiet;        // Hide fbink_init()'s hardware setup info (sent to stderr)
//	bool      ignore_alpha;    // Ignore any potential alpha channel in source image (i.e., flatten the image)
//	uint8_t   halign;    // Horizontal alignment of images (NONE/LEFT, CENTER, EDGE/RIGHT; c.f., ALIGN_INDEX_T enum)
// 	uint8_t   valign;    // Vertical alignment of images (NONE/TOP, CENTER, EDGE/BOTTOM; c.f., ALIGN_INDEX_T enum)
//};

FBInkConfig fbink_config = { 0 };
int fbfd = FBFD_AUTO;

//fbink_init API
//fbink_init(int fbfd, const FBInkConfig* fbink_config)
fbink_init(fbfd, &fbink_config);


    if ( FT_Init_FreeType( &font_library ) )
      exit( 1 );
    if ( FT_New_Face( font_library, "pantspatrol.ttf", 0 , &font_face ) )
      exit( 1 );
    if ( FT_Set_Char_Size( font_face , 0 , 768 , 300 , 300 ) )
      exit( 1 );

    num_chars = (int)font_face->num_glyphs;
    FT_Set_Transform( font_face , NULL , NULL );

    for ( glyph_ind = 0 ; glyph_ind < num_chars; glyph_ind++ )
    {
      if ( FT_Load_Glyph( font_face, glyph_ind, FT_LOAD_DEFAULT ) )
        exit( 1 );
      cur_glyph = font_face->glyph;
      if ( cur_glyph->format != FT_GLYPH_FORMAT_BITMAP )
        if ( FT_Render_Glyph( font_face->glyph, FT_RENDER_MODE_MONO ) )
          exit( 1 );
      if ( FT_Get_Glyph_Name( font_face, glyph_ind, char_name, 16 ) )
        exit( 1 );

      bitmap = cur_glyph->bitmap;
      x=glyph_metrics.horiBearingX / 64;

//  fbink_print_raw_data API
//    fbink_print_raw_data(int fbfd,
//			 unsigned char* data,
//			 const int w,
//			 const int h ,
//			 const size_t len,
//			 short int x_off,
//			 short int y_off,
//			 const FBInkConfig* fbink_config)

      fbink_print_raw_data(fbfd, &bitmap, bitmap.width, bitmap.rows, bitmap.width*bitmap.rows, x, -glyph_metrics. horiBearingY / 64, &fbink_config);
      
      x=x+glyph_metrics.horiAdvance / 64;
      glyph_metrics = cur_glyph->metrics;

      printf( "Glyph %d  name %s %ld %ld %ld %d %d\n",
              glyph_ind,
              char_name,
              glyph_metrics.horiBearingX / 64,
              glyph_metrics.horiBearingY / 64,
              glyph_metrics.horiAdvance / 64,
              bitmap.width , bitmap.rows );
    }

    return 0;
  }

at the risk of exposing my limited coding ability

[NiLuJe]

Yeah, that makes sense . I'll tweak the release script later tonight!

[jackie_w]

@NiLuJe,

I'm probably flogging a dead horse here but nothing ventured, nothing gained ...

Following on from the Display time when reading thread is there any way that fbink can detect whether the Kobo is currently in inverted colour mode and reverse the background/foreground colours automatically?

[NiLuJe]

@jackie_w: Nnnnnnyes, but not really .

(Technically possible, although potentially not accurately depending on the approach chosen, but in practice a terrible idea).

That's much better left to whatever's using FBInk, and toggling its own invert flag (-h) based on whether it finds NM enabled or not .

[NiLuJe]

@jackie_w: In this case, after taking a look at the nightmode hack, I'd say something like...

if grep -q "state:invertActive = yes" "/mnt/onboard/.kobo/nightmode.ini" ; then
    fbink -h ....
else
    fbink ....
fi

[NiLuJe]

@PoP: Updated the package in #93 .

[jackie_w]

Originally Posted by NiLuJe

@jackie_w: In this case, after taking a look at the nightmode hack, I'd say something like...

if grep -q "state:invertActive = yes" "/mnt/onboard/.kobo/nightmode.ini" ; then
    fbink -h ....
else
    fbink ....
fi

Thanks for looking I tried, but no joy. No matter how many times I toggled NightMode on/off the miniclock was always displayed black-on-white.
miniclock.cfg file has default settings

fg_color=BLACK
bg_color=WHITE

Part of the problem may be that I don't think the nightmode.ini file's [state]invertActive parameter is dynamically updated as the device toggles in/out of NightMode. The timestamp on the .ini file remains unchanged. I don't know how/where the actual current state is stored.

But I don't think this can be the whole problem. Even if I manually set the .ini to invertActive = yes, before rebooting, I still always get black-on-white. It doesn't look like the -h parameter is doing anything.

However, in case my practically non-existent shell scripting skills are at fault, I changed miniclock.sh from

Spoiler:

update() {
    sleep 0.1
    fbink -X "$cfg_offset_x" -Y "$cfg_offset_y" -F "$cfg_font" -S "$cfg_size" \
          -C "$cfg_fg_color" -B "$cfg_bg_color" \
          "$(date +"$cfg_format")"
}

to

Spoiler:

update() {
    sleep 0.1
    if grep -q "state:invertActive = yes" "/mnt/onboard/.kobo/nightmode.ini" ; then
        fbink -h -X "$cfg_offset_x" -Y "$cfg_offset_y" -F "$cfg_font" -S "$cfg_size" \
          -C "$cfg_fg_color" -B "$cfg_bg_color" \
          "$(date +"$cfg_format")"
    else
        fbink -X "$cfg_offset_x" -Y "$cfg_offset_y" -F "$cfg_font" -S "$cfg_size" \
          -C "$cfg_fg_color" -B "$cfg_bg_color" \
          "$(date +"$cfg_format")"
    fi
}

and I made the changes to frostschutz's Miniclock update dated today.

I suspect I'm outstaying my welcome on this annoying little issue, but if you do get any more bright ideas I'll be happy to test them out ... as long as it can be done with a simple Windows PC