PIKApp/app/text/pikafont.c

1336 lines
43 KiB
C

/* PIKA - Photo and Image Kooker Application
* a rebranding of The GNU Image Manipulation Program (created with heckimp)
* A derived work which may be trivial. However, any changes may be (C)2023 by Aldercone Studio
*
* Original copyright, applying to most contents (license remains unchanged):
* Copyright (C) 1995 Spencer Kimball and Peter Mattis
*
* pikafont.c
* Copyright (C) 2003 Michael Natterer <mitch@gimp.org>
* Sven Neumann <sven@gimp.org>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*/
#include "config.h"
#include <gdk-pixbuf/gdk-pixbuf.h>
#include <gegl.h>
#include <pango/pangocairo.h>
#include <hb.h>
#include <hb-ot.h>
#include <hb-ft.h>
#define PANGO_ENABLE_ENGINE 1 /* Argh */
#include <pango/pango-ot.h>
#include <ft2build.h>
#include FT_TRUETYPE_TABLES_H
#include "libpikabase/pikabase.h"
#include "libpikaconfig/pikaconfig.h"
#include "text-types.h"
#include "core/pikatempbuf.h"
#include "core/pika-memsize.h"
#include "core/pikacontainer.h"
#include "pikafont.h"
#include "pika-intl.h"
/* This is a so-called pangram; it's supposed to
contain all characters found in the alphabet. */
#define PIKA_TEXT_PANGRAM N_("Pack my box with\nfive dozen liquor jugs.")
#define PIKA_FONT_POPUP_SIZE (PANGO_SCALE * 30)
#define DEBUGPRINT(x) /* g_print x */
enum
{
PIKA_FONT_SYMBOL_FONTHASH,
PIKA_FONT_SYMBOL_FULLNAME,
PIKA_FONT_SYMBOL_FAMILY,
PIKA_FONT_SYMBOL_STYLE,
PIKA_FONT_SYMBOL_PSNAME,
PIKA_FONT_SYMBOL_INDEX,
PIKA_FONT_SYMBOL_WEIGHT,
PIKA_FONT_SYMBOL_SLANT,
PIKA_FONT_SYMBOL_WIDTH,
PIKA_FONT_SYMBOL_FONTVERSION
};
enum
{
PROP_0,
PROP_PANGO_CONTEXT
};
struct _PikaFont
{
PikaData parent_instance;
PangoContext *pango_context;
PangoLayout *popup_layout;
gint popup_width;
gint popup_height;
gchar *lookup_name;
/*properties for serialization*/
gchar *hash;
gchar *fullname;
gchar *family;
gchar *style;
gchar *psname;
gint weight;
gint width;
gint index;
gint slant;
gint fontversion;
/*for backward compatibility*/
gchar *desc;
};
struct _PikaFontClass
{
PikaDataClass parent_class;
PikaContainer *fontfactory;
};
static void pika_font_finalize (GObject *object);
static void pika_font_set_property (GObject *object,
guint property_id,
const GValue *value,
GParamSpec *pspec);
static void pika_font_get_preview_size (PikaViewable *viewable,
gint size,
gboolean popup,
gboolean dot_for_dot,
gint *width,
gint *height);
static gboolean pika_font_get_popup_size (PikaViewable *viewable,
gint width,
gint height,
gboolean dot_for_dot,
gint *popup_width,
gint *popup_height);
static PikaTempBuf * pika_font_get_new_preview (PikaViewable *viewable,
PikaContext *context,
gint width,
gint height);
static void pika_font_config_iface_init (PikaConfigInterface *iface);
static gboolean pika_font_serialize (PikaConfig *config,
PikaConfigWriter *writer,
gpointer data);
static PikaConfig * pika_font_deserialize_create (GType type,
GScanner *scanner,
gint nest_level,
gpointer data);
static gint64 pika_font_get_memsize (PikaObject *object,
gint64 *gui_size);
static inline gboolean pika_font_covers_string (PangoFont *font,
const gchar *sample);
static hb_tag_t get_hb_table_type (PangoOTTableType table_type);
static const gchar * pika_font_get_sample_string (PangoContext *context,
PangoFontDescription *font_desc);
static const gchar * pika_font_get_hash (PikaFont *font);
G_DEFINE_TYPE_WITH_CODE (PikaFont, pika_font, PIKA_TYPE_DATA,
G_IMPLEMENT_INTERFACE (PIKA_TYPE_CONFIG,
pika_font_config_iface_init))
#define parent_class pika_font_parent_class
static void
pika_font_config_iface_init (PikaConfigInterface *iface)
{
iface->serialize = pika_font_serialize;
iface->deserialize_create = pika_font_deserialize_create;
}
static gboolean
pika_font_serialize (PikaConfig *config,
PikaConfigWriter *writer,
gpointer data)
{
PikaFont *font;
g_return_val_if_fail (PIKA_IS_FONT (config), FALSE);
font = PIKA_FONT (config);
if (font == PIKA_FONT (pika_font_get_standard ()))
return TRUE;
pika_config_writer_open (writer, "fonthash");
pika_config_writer_string (writer, pika_font_get_hash (font));
pika_config_writer_close (writer);
pika_config_writer_open (writer, "fullname");
pika_config_writer_string (writer, font->fullname);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "family");
pika_config_writer_string (writer, font->family);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "style");
pika_config_writer_string (writer, font->style);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "psname");
pika_config_writer_string (writer, font->psname);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "index");
pika_config_writer_printf (writer, "%d", font->index);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "weight");
pika_config_writer_printf (writer, "%d", font->weight);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "slant");
pika_config_writer_printf (writer, "%d", font->slant);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "width");
pika_config_writer_printf (writer, "%d", font->width);
pika_config_writer_close (writer);
pika_config_writer_open (writer, "fontversion");
pika_config_writer_printf (writer, "%d", font->fontversion);
pika_config_writer_close (writer);
return TRUE;
}
static PikaConfig *
pika_font_deserialize_create (GType type,
GScanner *scanner,
gint nest_level,
gpointer data)
{
PikaFont *font;
PikaContainer *fonts_container = PIKA_FONT_CLASS (g_type_class_peek (PIKA_TYPE_FONT))->fontfactory;
gint most_similar_font_index = -1;
gint font_count = pika_container_get_n_children (fonts_container);
gint largest_similarity = 0;
GList *similar_fonts = NULL;
GList *iter;
gint i;
gchar *fonthash = NULL;
gchar *fullname = NULL;
gchar *family = NULL;
gchar *psname = NULL;
gchar *style = NULL;
gint index = -1;
gint weight = -1;
gint slant = -1;
gint width = -1;
gint fontversion = -1;
guint scope_id;
guint old_scope_id;
/* This is for backward compatibility with older xcf files.
* The font used to be serialized as a string containing
* its name.
*/
if (g_scanner_peek_next_token (scanner) == G_TOKEN_STRING)
{
gchar* font_name = NULL;
pika_scanner_parse_string (scanner, &font_name);
for (i = 0; i < font_count; i++)
{
font = PIKA_FONT (pika_container_get_child_by_index (fonts_container, i));
if (!g_strcmp0 (font->desc, font_name))
break;
font = NULL;
}
if (font == NULL)
font = PIKA_FONT (pika_font_get_standard ());
else
g_object_ref (font);
g_free (font_name);
return PIKA_CONFIG (PIKA_FONT (font));
}
if (g_scanner_peek_next_token (scanner) == G_TOKEN_RIGHT_PAREN)
return PIKA_CONFIG (PIKA_FONT (pika_font_get_standard ()));
scope_id = g_type_qname (type);
old_scope_id = g_scanner_set_scope (scanner, scope_id);
g_scanner_scope_add_symbol (scanner, scope_id, "fonthash",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_FONTHASH));
g_scanner_scope_add_symbol (scanner, scope_id, "fullname",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_FULLNAME));
g_scanner_scope_add_symbol (scanner, scope_id, "family",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_FAMILY));
g_scanner_scope_add_symbol (scanner, scope_id, "style",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_STYLE));
g_scanner_scope_add_symbol (scanner, scope_id, "psname",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_PSNAME));
g_scanner_scope_add_symbol (scanner, scope_id, "index",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_INDEX));
g_scanner_scope_add_symbol (scanner, scope_id, "weight",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_WEIGHT));
g_scanner_scope_add_symbol (scanner, scope_id, "slant",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_SLANT));
g_scanner_scope_add_symbol (scanner, scope_id, "width",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_WIDTH));
g_scanner_scope_add_symbol (scanner, scope_id, "fontversion",
GINT_TO_POINTER (PIKA_FONT_SYMBOL_FONTVERSION));
while (g_scanner_peek_next_token (scanner) == G_TOKEN_LEFT_PAREN)
{
GTokenType token;
g_scanner_get_next_token (scanner); /* ( */
token = g_scanner_get_next_token (scanner);
if (token != G_TOKEN_SYMBOL)
{
/* When encountering an unknown symbol, we simply ignore its contents
* (up to the next closing parenthese). This would allow us to load
* future XCF files in case we add new fields to recognize fonts
* without having to bump the XCF version (just ignoring unknown
* fields). We still output a small message on stderr.
*/
if (token == G_TOKEN_IDENTIFIER)
g_printerr ("%s: ignoring unknown symbol '%s'.\n", G_STRFUNC, scanner->value.v_string);
else
g_printerr ("%s: ignoring unknown token %d.\n", G_STRFUNC, token);
while ((token = g_scanner_get_next_token (scanner)) != G_TOKEN_EOF)
{
if (token == G_TOKEN_RIGHT_PAREN)
break;
}
continue;
}
switch (GPOINTER_TO_INT (scanner->value.v_symbol))
{
case PIKA_FONT_SYMBOL_FONTHASH:
pika_scanner_parse_string (scanner, &fonthash);
break;
case PIKA_FONT_SYMBOL_FULLNAME:
pika_scanner_parse_string (scanner, &fullname);
break;
case PIKA_FONT_SYMBOL_FAMILY:
pika_scanner_parse_string (scanner, &family);
break;
case PIKA_FONT_SYMBOL_STYLE:
pika_scanner_parse_string (scanner, &style);
break;
case PIKA_FONT_SYMBOL_PSNAME:
pika_scanner_parse_string (scanner, &psname);
break;
case PIKA_FONT_SYMBOL_INDEX:
pika_scanner_parse_int (scanner, &index);
break;
case PIKA_FONT_SYMBOL_WEIGHT:
pika_scanner_parse_int (scanner, &weight);
break;
case PIKA_FONT_SYMBOL_SLANT:
pika_scanner_parse_int (scanner, &slant);
break;
case PIKA_FONT_SYMBOL_WIDTH:
pika_scanner_parse_int (scanner, &width);
break;
case PIKA_FONT_SYMBOL_FONTVERSION:
pika_scanner_parse_int (scanner, &fontversion);
break;
default:
break;
}
if (g_scanner_get_next_token (scanner) != G_TOKEN_RIGHT_PAREN)
break;
}
g_scanner_scope_remove_symbol (scanner, scope_id, "fonthash");
g_scanner_scope_remove_symbol (scanner, scope_id, "fullname");
g_scanner_scope_remove_symbol (scanner, scope_id, "family");
g_scanner_scope_remove_symbol (scanner, scope_id, "style");
g_scanner_scope_remove_symbol (scanner, scope_id, "psname");
g_scanner_scope_remove_symbol (scanner, scope_id, "index");
g_scanner_scope_remove_symbol (scanner, scope_id, "weight");
g_scanner_scope_remove_symbol (scanner, scope_id, "slant");
g_scanner_scope_remove_symbol (scanner, scope_id, "width");
g_scanner_scope_remove_symbol (scanner, scope_id, "fontversion");
g_scanner_set_scope (scanner, old_scope_id);
for (i = 0; i < font_count; i++)
{
gint current_font_similarity = 0;
font = PIKA_FONT (pika_container_get_child_by_index (fonts_container, i));
/* Some attrs are more identifying than others,
* hence their higher importance in measuring similarity.
*/
if (fullname != NULL && !g_strcmp0 (font->fullname, fullname))
current_font_similarity += 5;
if (family != NULL && !g_strcmp0 (font->family, family))
current_font_similarity += 5;
if (psname != NULL && font->psname != NULL && !g_strcmp0 (font->psname, psname))
current_font_similarity += 5;
if (current_font_similarity < 5)
continue;
if (style != NULL && font->style != NULL && !g_strcmp0 (font->style, style))
current_font_similarity++;
if (font->weight != -1 && font->weight == weight)
current_font_similarity++;
if (font->width != -1 && font->width == width)
current_font_similarity++;
if (font->slant != -1 && font->slant == slant)
current_font_similarity++;
if (font->index != -1 && font->index == index)
current_font_similarity++;
if (font->fontversion != -1 && font->fontversion == fontversion)
current_font_similarity++;
if (current_font_similarity > largest_similarity)
{
largest_similarity = current_font_similarity;
most_similar_font_index = i;
g_clear_pointer (&similar_fonts, g_list_free);
similar_fonts = g_list_prepend (similar_fonts, GINT_TO_POINTER (i));
}
else if (current_font_similarity == largest_similarity)
{
similar_fonts = g_list_prepend (similar_fonts, GINT_TO_POINTER (i));
}
}
/* In case there are multiple font with identical info,
* the font file hash should be used for comparison.
*/
if (g_list_length (similar_fonts) == 1)
g_clear_pointer (&similar_fonts, g_list_free);
for (iter = similar_fonts; iter; iter = iter->next)
{
i = GPOINTER_TO_INT (iter->data);
font = PIKA_FONT (pika_container_get_child_by_index (fonts_container, i));
if (g_strcmp0 (pika_font_get_hash (font), fonthash) == 0)
{
most_similar_font_index = i;
break;
}
}
if (most_similar_font_index > -1)
{
font = PIKA_FONT (pika_container_get_child_by_index (fonts_container, most_similar_font_index));
g_object_ref (font);
}
else
{
font = PIKA_FONT (pika_font_get_standard ());
}
g_list_free (similar_fonts);
g_free (fonthash);
g_free (fullname);
g_free (family);
g_free (psname);
g_free (style);
return PIKA_CONFIG (font);
}
void
pika_font_class_set_font_factory (PikaContainer *factory)
{
PikaFontClass *klass = PIKA_FONT_CLASS (g_type_class_peek (PIKA_TYPE_FONT));
klass->fontfactory = factory;
}
void
pika_font_set_font_info (PikaFont *font,
gpointer font_info[])
{
font->fullname = g_strdup ((gchar*)font_info[PROP_FULLNAME]);
font->family = g_strdup ((gchar*)font_info[PROP_FAMILY]);
font->style = g_strdup ((gchar*)font_info[PROP_STYLE]);
font->psname = g_strdup ((gchar*)font_info[PROP_PSNAME]);
font->desc = g_strdup ((gchar*)font_info[PROP_DESC]);
font->weight = *(gint*)font_info[PROP_WEIGHT];
font->width = *(gint*)font_info[PROP_WIDTH];
font->index = *(gint*)font_info[PROP_INDEX];
font->slant = *(gint*)font_info[PROP_SLANT];
font->fontversion = *(gint*)font_info[PROP_FONTVERSION];
}
void
pika_font_set_lookup_name (PikaFont *font,
gchar *name)
{
font->lookup_name = name;
}
gboolean
pika_font_match_by_lookup_name (PikaFont *font,
const gchar *name)
{
return !g_strcmp0 (font->lookup_name, name);
}
const gchar*
pika_font_get_lookup_name (PikaFont *font)
{
return font->lookup_name;
}
static void
pika_font_class_init (PikaFontClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
PikaViewableClass *viewable_class = PIKA_VIEWABLE_CLASS (klass);
PikaObjectClass *pika_object_class = PIKA_OBJECT_CLASS (klass);
object_class->finalize = pika_font_finalize;
object_class->set_property = pika_font_set_property;
pika_object_class->get_memsize = pika_font_get_memsize;
viewable_class->get_preview_size = pika_font_get_preview_size;
viewable_class->get_popup_size = pika_font_get_popup_size;
viewable_class->get_new_preview = pika_font_get_new_preview;
viewable_class->default_icon_name = "gtk-select-font";
g_object_class_install_property (object_class, PROP_PANGO_CONTEXT,
g_param_spec_object ("pango-context",
NULL, NULL,
PANGO_TYPE_CONTEXT,
PIKA_PARAM_WRITABLE));
}
static void
pika_font_init (PikaFont *font)
{
}
static void
pika_font_finalize (GObject *object)
{
PikaFont *font = PIKA_FONT (object);
g_clear_object (&font->pango_context);
g_clear_object (&font->popup_layout);
g_free (font->lookup_name);
g_free (font->hash);
g_free (font->fullname);
g_free (font->family);
g_free (font->style);
g_free (font->psname);
g_free (font->desc);
G_OBJECT_CLASS (parent_class)->finalize (object);
}
static void
pika_font_set_property (GObject *object,
guint property_id,
const GValue *value,
GParamSpec *pspec)
{
PikaFont *font = PIKA_FONT (object);
switch (property_id)
{
case PROP_PANGO_CONTEXT:
if (font->pango_context)
g_object_unref (font->pango_context);
font->pango_context = g_value_dup_object (value);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, property_id, pspec);
break;
}
}
static gint64
pika_font_get_memsize (PikaObject *object,
gint64 *gui_size)
{
PikaFont *font = PIKA_FONT (object);
gint64 memsize = 0;
memsize += pika_string_get_memsize (font->lookup_name);
memsize += pika_string_get_memsize (font->hash);
memsize += pika_string_get_memsize (font->fullname);
memsize += pika_string_get_memsize (font->family);
memsize += pika_string_get_memsize (font->style);
memsize += pika_string_get_memsize (font->psname);
memsize += pika_string_get_memsize (font->desc);
return memsize + PIKA_OBJECT_CLASS (parent_class)->get_memsize (object,
gui_size);
}
static void
pika_font_get_preview_size (PikaViewable *viewable,
gint size,
gboolean popup,
gboolean dot_for_dot,
gint *width,
gint *height)
{
*width = size;
*height = size;
}
static gboolean
pika_font_get_popup_size (PikaViewable *viewable,
gint width,
gint height,
gboolean dot_for_dot,
gint *popup_width,
gint *popup_height)
{
PikaFont *font = PIKA_FONT (viewable);
PangoFontDescription *font_desc;
PangoRectangle ink;
PangoRectangle logical;
const gchar *name;
if (! font->pango_context)
return FALSE;
name = font->lookup_name;
font_desc = pango_font_description_from_string (name);
g_return_val_if_fail (font_desc != NULL, FALSE);
pango_font_description_set_size (font_desc, PIKA_FONT_POPUP_SIZE);
if (font->popup_layout)
g_object_unref (font->popup_layout);
font->popup_layout = pango_layout_new (font->pango_context);
pango_layout_set_font_description (font->popup_layout, font_desc);
pango_font_description_free (font_desc);
pango_layout_set_text (font->popup_layout, gettext (PIKA_TEXT_PANGRAM), -1);
pango_layout_get_pixel_extents (font->popup_layout, &ink, &logical);
*popup_width = MAX (ink.width, logical.width) + 6;
*popup_height = MAX (ink.height, logical.height) + 6;
*popup_width = cairo_format_stride_for_width (CAIRO_FORMAT_A8, *popup_width);
font->popup_width = *popup_width;
font->popup_height = *popup_height;
return TRUE;
}
static PikaTempBuf *
pika_font_get_new_preview (PikaViewable *viewable,
PikaContext *context,
gint width,
gint height)
{
PikaFont *font = PIKA_FONT (viewable);
PangoLayout *layout;
PangoRectangle ink;
PangoRectangle logical;
gint layout_width;
gint layout_height;
gint layout_x;
gint layout_y;
PikaTempBuf *temp_buf;
cairo_t *cr;
cairo_surface_t *surface;
if (! font->pango_context)
return NULL;
if (! font->popup_layout ||
font->popup_width != width || font->popup_height != height)
{
PangoFontDescription *font_desc;
const gchar *name;
name = font->lookup_name;
DEBUGPRINT (("%s: ", name));
font_desc = pango_font_description_from_string (name);
g_return_val_if_fail (font_desc != NULL, NULL);
pango_font_description_set_size (font_desc,
PANGO_SCALE * height * 2.0 / 3.0);
layout = pango_layout_new (font->pango_context);
pango_layout_set_font_description (layout, font_desc);
pango_layout_set_text (layout,
pika_font_get_sample_string (font->pango_context,
font_desc),
-1);
pango_font_description_free (font_desc);
}
else
{
layout = g_object_ref (font->popup_layout);
}
width = cairo_format_stride_for_width (CAIRO_FORMAT_A8, width);
temp_buf = pika_temp_buf_new (width, height, babl_format ("Y' u8"));
memset (pika_temp_buf_get_data (temp_buf), 255, width * height);
surface = cairo_image_surface_create_for_data (pika_temp_buf_get_data (temp_buf),
CAIRO_FORMAT_A8,
width, height, width);
pango_layout_get_pixel_extents (layout, &ink, &logical);
layout_width = MAX (ink.width, logical.width);
layout_height = MAX (ink.height, logical.height);
layout_x = (width - layout_width) / 2;
layout_y = (height - layout_height) / 2;
if (ink.x < logical.x)
layout_x += logical.x - ink.x;
if (ink.y < logical.y)
layout_y += logical.y - ink.y;
cr = cairo_create (surface);
cairo_translate (cr, layout_x, layout_y);
cairo_set_operator (cr, CAIRO_OPERATOR_CLEAR);
pango_cairo_show_layout (cr, layout);
cairo_destroy (cr);
cairo_surface_destroy (surface);
g_object_unref (layout);
return temp_buf;
}
PikaData *
pika_font_get_standard (void)
{
static PikaData *standard_font = NULL;
if (! standard_font)
{
g_set_weak_pointer (&standard_font,
g_object_new (PIKA_TYPE_FONT,
"name", "",
NULL));;
/* pango_font_description_from_string doesn't accept NULL */
PIKA_FONT (standard_font)->lookup_name = g_strdup ("");
pika_data_clean (standard_font);
pika_data_make_internal (standard_font, "pika-font-standard");
}
return standard_font;
}
/* Private functions */
static inline gboolean
pika_font_covers_string (PangoFont *font,
const gchar *sample)
{
const gchar *p;
for (p = sample; *p; p = g_utf8_next_char (p))
{
if (! pango_font_has_char (font, g_utf8_get_char (p)))
return FALSE;
}
return TRUE;
}
/* This function was picked up from Pango's pango-ot-info.c. Until there
* is a better way to get the tag, we use this.
*/
static hb_tag_t
get_hb_table_type (PangoOTTableType table_type)
{
switch (table_type)
{
case PANGO_OT_TABLE_GSUB:
return HB_OT_TAG_GSUB;
case PANGO_OT_TABLE_GPOS:
return HB_OT_TAG_GPOS;
default:
return HB_TAG_NONE;
}
}
/* Guess a suitable short sample string for the font. */
static const gchar *
pika_font_get_sample_string (PangoContext *context,
PangoFontDescription *font_desc)
{
PangoFont *font;
hb_face_t *hb_face;
hb_font_t *hb_font;
FT_Face face;
TT_OS2 *os2;
PangoOTTableType tt;
gint i;
/* This is a table of scripts and corresponding short sample strings
* to be used instead of the Latin sample string Aa. The script
* codes are as in ISO15924 (see
* https://www.unicode.org/iso15924/iso15924-codes.html), but in
* lower case. The Unicode subrange bit numbers, as used in TrueType
* so-called OS/2 tables, are from
* https://www.microsoft.com/typography/otspec/os2.htm#ur .
*
* The table is mostly ordered by Unicode order. But as there are
* fonts that support several of these scripts, the ordering is
* be modified so that the script which such a font is more likely
* to be actually designed for comes first and matches.
*
* These sample strings are mostly just guesswork as for their
* usefulness. Usually they contain what I assume is the first
* letter in the corresponding alphabet, or two first letters if the
* first one happens to look too "trivial" to be recognizable by
* itself.
*
* This table is used to determine the primary script a font has
* been designed for.
*
* Very useful link: https://www.wazu.jp/index.html
*/
static const struct
{
const gchar script[4];
gint bit;
const gchar *sample;
} scripts[] = {
/* CJK are first because fonts that support these characters
* presumably are primarily designed for it.
*
* Though not a universal rules, most fonts targeting Korean will
* also have the ideographs as they were used historically (and
* still in some cases). As for Japanese, ideographs are still
* definitely used daily. After all the block is called "*CJK*
* (Chinese-Japanese-Korean) Unified Ideographs". So let's give the
* Chinese representation less priority towards finding
* Korean/Japanese targeting.
* Then we prioritize Korean because many fonts for Korean also
* design Hiragana/Katakana blocks. On the other hand, I could see
* very few (none so far) Japanese fonts designing also Hangul
* characters. So if we see both Hiragana and Hangul, we can assume
* this is a font for Korean market, whereas Hiragana only is for
* Japanese.
* XXX: of course we could probably come with a better algorithm to
* determine the target language. Probably looking at more than the
* existence of a single block would help (since some languages are
* spread on several blocks).
*/
{
"hang", /* Korean alphabet (Hangul) */
56, /* Hangul Syllables block */
"\355\225\234" /* U+D55C HANGUL SYLLABLE HAN */
},
{
"japa", /* Japanese */
49, /* Hiragana block */
"\343\201\202" /* U+3042 HIRAGANA LETTER A */
},
{
"hani", /* Han Ideographic */
59, /* CJK Unified Ideographs */
"\346\260\270" /* U+6C38 "forever". Ed Trager says
* this is a "pan-stroke" often used
* in teaching Chinese calligraphy. It
* contains the eight basic Chinese
* stroke forms.
*/
},
{
"copt", /* Coptic */
7,
"\316\221\316\261" /* U+0410 GREEK CAPITAL LETTER ALPHA
U+0430 GREEK SMALL LETTER ALPHA
*/
},
{
"grek", /* Greek */
7,
"\316\221\316\261" /* U+0410 GREEK CAPITAL LETTER ALPHA
U+0430 GREEK SMALL LETTER ALPHA
*/
},
{
"cyrl", /* Cyrillic */
9,
"\320\220\325\260" /* U+0410 CYRILLIC CAPITAL LETTER A
U+0430 CYRILLIC SMALL LETTER A
*/
},
{
"armn", /* Armenian */
10,
"\324\261\325\241" /* U+0531 ARMENIAN CAPITAL LETTER AYB
U+0561 ARMENIAN SMALL LETTER AYB
*/
},
{
"hebr", /* Hebrew */
11,
"\327\220" /* U+05D0 HEBREW LETTER ALEF */
},
{
"arab", /* Arabic */
13,
"\330\247\330\250" /* U+0627 ARABIC LETTER ALEF
* U+0628 ARABIC LETTER BEH
*/
},
{
"syrc", /* Syriac */
71,
"\334\220\334\222" /* U+0710 SYRIAC LETTER ALAPH
* U+0712 SYRIAC LETTER BETH
*/
},
{
"thaa", /* Thaana */
72,
"\336\200\336\201" /* U+0780 THAANA LETTER HAA
* U+0781 THAANA LETTER SHAVIYANI
*/
},
/* Should really use some better sample strings for the complex
* scripts. Something that shows how well the font handles the
* complex processing for each script.
*/
{
"deva", /* Devanagari */
15,
"\340\244\205" /* U+0905 DEVANAGARI LETTER A*/
},
{
"beng", /* Bengali */
16,
"\340\246\205" /* U+0985 BENGALI LETTER A */
},
{
"guru", /* Gurmukhi */
17,
"\340\250\205" /* U+0A05 GURMUKHI LETTER A */
},
{
"gujr", /* Gujarati */
18,
"\340\252\205" /* U+0A85 GUJARATI LETTER A */
},
{
"orya", /* Oriya */
19,
"\340\254\205" /* U+0B05 ORIYA LETTER A */
},
{
"taml", /* Tamil */
20,
"\340\256\205" /* U+0B85 TAMIL LETTER A */
},
{
"telu", /* Telugu */
21,
"\340\260\205" /* U+0C05 TELUGU LETTER A */
},
{
"knda", /* Kannada */
22,
"\340\262\205" /* U+0C85 KANNADA LETTER A */
},
{
"mylm", /* Malayalam */
23,
"\340\264\205" /* U+0D05 MALAYALAM LETTER A */
},
{
"sinh", /* Sinhala */
73,
"\340\266\205" /* U+0D85 SINHALA LETTER AYANNA */
},
{
"thai", /* Thai */
24,
"\340\270\201\340\270\264"/* U+0E01 THAI CHARACTER KO KAI
* U+0E34 THAI CHARACTER SARA I
*/
},
{
"laoo", /* Lao */
25,
"\340\272\201\340\272\264"/* U+0E81 LAO LETTER KO
* U+0EB4 LAO VOWEL SIGN I
*/
},
{
"tibt", /* Tibetan */
70,
"\340\274\200" /* U+0F00 TIBETAN SYLLABLE OM */
},
{
"mymr", /* Myanmar */
74,
"\341\200\200" /* U+1000 MYANMAR LETTER KA */
},
{
"geor", /* Georgian */
26,
"\341\202\240\341\203\200" /* U+10A0 GEORGIAN CAPITAL LETTER AN
* U+10D0 GEORGIAN LETTER AN
*/
},
{
"ethi", /* Ethiopic */
75,
"\341\210\200" /* U+1200 ETHIOPIC SYLLABLE HA */
},
{
"cher", /* Cherokee */
76,
"\341\216\243" /* U+13A3 CHEROKEE LETTER O */
},
{
"cans", /* Unified Canadian Aboriginal Syllabics */
77,
"\341\220\201" /* U+1401 CANADIAN SYLLABICS E */
},
{
"ogam", /* Ogham */
78,
"\341\232\201" /* U+1681 OGHAM LETTER BEITH */
},
{
"runr", /* Runic */
79,
"\341\232\240" /* U+16A0 RUNIC LETTER FEHU FEOH FE F */
},
{
"tglg", /* Tagalog */
84,
"\341\234\200" /* U+1700 TAGALOG LETTER A */
},
{
"hano", /* Hanunoo */
-1,
"\341\234\240" /* U+1720 HANUNOO LETTER A */
},
{
"buhd", /* Buhid */
-1,
"\341\235\200" /* U+1740 BUHID LETTER A */
},
{
"tagb", /* Tagbanwa */
-1,
"\341\235\240" /* U+1760 TAGBANWA LETTER A */
},
{
"khmr", /* Khmer */
80,
"\341\236\201\341\237\222\341\236\211\341\236\273\341\237\206"
/* U+1781 KHMER LETTER KHA
* U+17D2 KHMER LETTER SIGN COENG
* U+1789 KHMER LETTER NYO
* U+17BB KHMER VOWEL SIGN U
* U+17C6 KHMER SIGN NIKAHIT
* A common word meaning "I" that contains
* lots of complex processing.
*/
},
{
"mong", /* Mongolian */
81,
"\341\240\240" /* U+1820 MONGOLIAN LETTER A */
},
{
"limb", /* Limbu */
-1,
"\341\244\201" /* U+1901 LIMBU LETTER KA */
},
{
"tale", /* Tai Le */
-1,
"\341\245\220" /* U+1950 TAI LE LETTER KA */
},
{
"latn", /* Latin */
0,
"Aa"
}
};
gint ot_alts[4];
gint n_ot_alts = 0;
gint sr_alts[20];
gint n_sr_alts = 0;
font = pango_context_load_font (context, font_desc);
g_return_val_if_fail (PANGO_IS_FC_FONT (font), "Aa");
hb_font = pango_font_get_hb_font (font);
g_return_val_if_fail (hb_font != NULL, "Aa");
/* These are needed to set hb_font to the right internal format, which
* can only work if the font is not immutable. Hence we make a copy.
*/
hb_font = hb_font_create_sub_font (hb_font);
hb_ft_font_set_funcs (hb_font);
/* TODO: use hb_ft_font_lock_face/hb_ft_font_unlock_face() when we
* bump to harfbuzz >= 2.6.5.
*/
face = hb_ft_font_get_face (hb_font);
/* Are there actual cases where this function could return NULL while
* it's not a bug in the code?
* For instance if the font file is broken, we don't want to return a
* CRITICAL, but properly address the issue (removing the font with a
* warning or whatnot).
* See #5922.
*/
g_return_val_if_fail (face != NULL, "Aa");
hb_face = hb_ft_face_create (face, NULL);
/* First check what script(s), if any, the font has GSUB or GPOS
* OpenType layout tables for.
*/
for (tt = PANGO_OT_TABLE_GSUB;
n_ot_alts < G_N_ELEMENTS (ot_alts) && tt <= PANGO_OT_TABLE_GPOS;
tt++)
{
hb_tag_t tag;
unsigned int count;
PangoOTTag *slist;
tag = get_hb_table_type (tt);
count = hb_ot_layout_table_get_script_tags (hb_face, tag, 0, NULL, NULL);
slist = g_new (PangoOTTag, count + 1);
hb_ot_layout_table_get_script_tags (hb_face, tag, 0, &count, slist);
slist[count] = 0;
for (i = 0;
n_ot_alts < G_N_ELEMENTS (ot_alts) && i < G_N_ELEMENTS (scripts);
i++)
{
gint j, k;
for (k = 0; k < n_ot_alts; k++)
if (ot_alts[k] == i)
break;
if (k == n_ot_alts)
{
for (j = 0;
n_ot_alts < G_N_ELEMENTS (ot_alts) && slist[j];
j++)
{
#define TAG(s) FT_MAKE_TAG (s[0], s[1], s[2], s[3])
if (slist[j] == TAG (scripts[i].script) &&
pika_font_covers_string (font, scripts[i].sample))
{
ot_alts[n_ot_alts++] = i;
DEBUGPRINT (("%.4s ", scripts[i].script));
}
#undef TAG
}
}
}
g_free (slist);
}
hb_face_destroy (hb_face);
DEBUGPRINT (("; OS/2: "));
/* Next check the OS/2 table for Unicode ranges the font claims
* to cover.
*/
os2 = FT_Get_Sfnt_Table (face, ft_sfnt_os2);
if (os2)
{
for (i = 0;
n_sr_alts < G_N_ELEMENTS (sr_alts) && i < G_N_ELEMENTS (scripts);
i++)
{
if (scripts[i].bit >= 0 &&
(&os2->ulUnicodeRange1)[scripts[i].bit/32] & (1 << (scripts[i].bit % 32)) &&
pika_font_covers_string (font, scripts[i].sample))
{
sr_alts[n_sr_alts++] = i;
DEBUGPRINT (("%.4s ", scripts[i].script));
}
}
}
hb_font_destroy (hb_font);
g_object_unref (font);
if (n_ot_alts > 2)
{
/* The font has OpenType tables for several scripts. If it
* support Basic Latin as well, use Aa.
*/
gint i;
for (i = 0; i < n_sr_alts; i++)
if (scripts[sr_alts[i]].bit == 0)
{
DEBUGPRINT (("=> several OT, also latin, use Aa\n"));
return "Aa";
}
}
if (n_ot_alts > 0 && n_sr_alts >= n_ot_alts + 3)
{
/* At least one script with an OpenType table, but many more
* subranges than such scripts. If it supports Basic Latin,
* use Aa, else the highest priority subrange.
*/
gint i;
for (i = 0; i < n_sr_alts; i++)
if (scripts[sr_alts[i]].bit == 0)
{
DEBUGPRINT (("=> several SR, also latin, use Aa\n"));
return "Aa";
}
DEBUGPRINT (("=> several SR, use %.4s\n", scripts[sr_alts[0]].script));
return scripts[sr_alts[0]].sample;
}
if (n_ot_alts > 0)
{
/* OpenType tables for at least one script, use the
* highest priority one
*/
DEBUGPRINT (("=> at least one OT, use %.4s\n",
scripts[sr_alts[0]].script));
return scripts[ot_alts[0]].sample;
}
if (n_sr_alts > 0)
{
/* Use the highest priority subrange. This means that a
* font that supports Greek, Cyrillic and Latin (quite
* common), will get the Greek sample string. That is
* capital and lowercase alpha, which looks like capital A
* and lowercase alpha, so it's actually quite nice, and
* doesn't give a too strong impression that the font would
* be for Greek only.
*/
DEBUGPRINT (("=> at least one SR, use %.4s\n",
scripts[sr_alts[0]].script));
return scripts[sr_alts[0]].sample;
}
/* Final fallback */
DEBUGPRINT (("=> fallback, use Aa\n"));
return "Aa";
}
static const gchar *
pika_font_get_hash (PikaFont *font)
{
/* Computing the hash is expensive, so it's only done in a few case, such as
* when serializing, when a similar fonts at deserialization looks like it
* could be a good identity candidate.
*/
if (font->hash == NULL)
{
PangoFontDescription *pfd = pango_font_description_from_string (font->lookup_name);
PangoFcFont *pango_font = PANGO_FC_FONT (pango_context_load_font (font->pango_context, pfd));
GChecksum *checksum = g_checksum_new (G_CHECKSUM_SHA256);
gchar *file;
hb_blob_t *hb_blob;
guint length;
const char *hb_data;
FcPatternGetString (pango_fc_font_get_pattern (pango_font), FC_FILE, 0, (FcChar8 **) &file);
hb_blob = hb_blob_create_from_file (file);
hb_data = hb_blob_get_data (hb_blob, &length);
g_checksum_update (checksum, (const guchar*) hb_data, length);
font->hash = g_strdup (g_checksum_get_string (checksum));
pango_font_description_free (pfd);
g_object_unref (pango_font);
hb_blob_destroy (hb_blob);
g_checksum_free (checksum);
}
return font->hash;
}