Rendering

Rendering — Functions to run the rendering pipeline

Object Hierarchy

    GBoxed
    ╰── PangoItem
    GFlags
    ╰── PangoShapeFlags

Description

The Pango rendering pipeline takes a string of Unicode characters and converts it into glyphs. The functions described in this section accomplish various steps of this process.

Functions

pango_itemize ()

GList *
pango_itemize (PangoContext *context,
               const char *text,
               int start_index,
               int length,
               PangoAttrList *attrs,
               PangoAttrIterator *cached_iter);

Breaks a piece of text into segments with consistent directional level and shaping engine. Each byte of text will be contained in exactly one of the items in the returned list; the generated list of items will be in logical order (the start offsets of the items are ascending).

cached_iter should be an iterator over attrs currently positioned at a range before or containing start_index ; cached_iter will be advanced to the range covering the position just after start_index + length . (i.e. if itemizing in a loop, just keep passing in the same cached_iter ).

Parameters

context

a structure holding information that affects the itemization process.

 

text

the text to itemize. Must be valid UTF-8

 

start_index

first byte in text to process

 

length

the number of bytes (not characters) to process after start_index . This must be >= 0.

 

attrs

the set of attributes that apply to text .

 

cached_iter

Cached attribute iterator, or NULL.

[allow-none]

Returns

a GList of PangoItem structures. The items should be freed using pango_item_free() probably in combination with g_list_foreach(), and the list itself using g_list_free().

[transfer full][element-type Pango.Item]


pango_itemize_with_base_dir ()

GList *
pango_itemize_with_base_dir (PangoContext *context,
                             PangoDirection base_dir,
                             const char *text,
                             int start_index,
                             int length,
                             PangoAttrList *attrs,
                             PangoAttrIterator *cached_iter);

Like pango_itemize(), but the base direction to use when computing bidirectional levels (see pango_context_set_base_dir()), is specified explicitly rather than gotten from the PangoContext.

Parameters

context

a structure holding information that affects the itemization process.

 

base_dir

base direction to use for bidirectional processing

 

text

the text to itemize.

 

start_index

first byte in text to process

 

length

the number of bytes (not characters) to process after start_index . This must be >= 0.

 

attrs

the set of attributes that apply to text .

 

cached_iter

Cached attribute iterator, or NULL.

[allow-none]

Returns

a GList of PangoItem structures. The items should be freed using pango_item_free() probably in combination with g_list_foreach(), and the list itself using g_list_free().

[transfer full][element-type Pango.Item]

Since: 1.4


pango_item_free ()

void
pango_item_free (PangoItem *item);

Free a PangoItem and all associated memory.

Parameters

item

a PangoItem, may be NULL.

[nullable]

pango_item_copy ()

PangoItem *
pango_item_copy (PangoItem *item);

Copy an existing PangoItem structure.

Parameters

item

a PangoItem, may be NULL.

[nullable]

Returns

the newly allocated PangoItem, which should be freed with pango_item_free(), or NULL if item was NULL.

[nullable]


pango_item_new ()

PangoItem *
pango_item_new (void);

Creates a new PangoItem structure initialized to default values.

Returns

the newly allocated PangoItem, which should be freed with pango_item_free().


pango_item_split ()

PangoItem *
pango_item_split (PangoItem *orig,
                  int split_index,
                  int split_offset);

Modifies orig to cover only the text after split_index , and returns a new item that covers the text before split_index that used to be in orig . You can think of split_index as the length of the returned item. split_index may not be 0, and it may not be greater than or equal to the length of orig (that is, there must be at least one byte assigned to each item, you can't create a zero-length item). split_offset is the length of the first item in chars, and must be provided because the text used to generate the item isn't available, so pango_item_split() can't count the char length of the split items itself.

Parameters

orig

a PangoItem

 

split_index

byte index of position to split item, relative to the start of the item

 

split_offset

number of chars between start of orig and split_index

 

Returns

new item representing text before split_index , which should be freed with pango_item_free().


pango_item_apply_attrs ()

void
pango_item_apply_attrs (PangoItem *item,
                        PangoAttrIterator *iter);

Add attributes to a PangoItem. The idea is that you have attributes that don't affect itemization, such as font features, so you filter them out using pango_attr_list_filter(), itemize your text, then reapply the attributes to the resulting items using this function.

The iter should be positioned before the range of the item, and will be advanced past it. This function is meant to be called in a loop over the items resulting from itemization, while passing the iter to each call.

Parameters

item

a PangoItem

 

iter

a PangoAttrIterator

 

Since: 1.44


pango_reorder_items ()

GList *
pango_reorder_items (GList *logical_items);

From a list of items in logical order and the associated directional levels, produce a list in visual order. The original list is unmodified.

Parameters

logical_items

a GList of PangoItem in logical order.

[element-type Pango.Item]

Returns

a GList of PangoItem structures in visual order.

(Please open a bug if you use this function. It is not a particularly convenient interface, and the code is duplicated elsewhere in Pango for that reason.).

[transfer full][element-type Pango.Item]


pango_break ()

void
pango_break (const gchar *text,
             int length,
             PangoAnalysis *analysis,
             PangoLogAttr *attrs,
             int attrs_len);

pango_break has been deprecated since version 1.44 and should not be used in newly-written code.

Use pango_default_break() and pango_tailor_break()

Determines possible line, word, and character breaks for a string of Unicode text with a single analysis. For most purposes you may want to use pango_get_log_attrs().

Parameters

text

the text to process. Must be valid UTF-8

 

length

length of text in bytes (may be -1 if text is nul-terminated)

 

analysis

PangoAnalysis structure from pango_itemize()

 

attrs

an array to store character information in.

[array length=attrs_len]

attrs_len

size of the array passed as attrs

 

pango_get_log_attrs ()

void
pango_get_log_attrs (const char *text,
                     int length,
                     int level,
                     PangoLanguage *language,
                     PangoLogAttr *log_attrs,
                     int attrs_len);

Computes a PangoLogAttr for each character in text . The log_attrs array must have one PangoLogAttr for each position in text ; if text contains N characters, it has N+1 positions, including the last position at the end of the text. text should be an entire paragraph; logical attributes can't be computed without context (for example you need to see spaces on either side of a word to know the word is a word).

Parameters

text

text to process. Must be valid UTF-8

 

length

length in bytes of text

 

level

embedding level, or -1 if unknown

 

language

language tag

 

log_attrs

array with one PangoLogAttr per character in text , plus one extra, to be filled in.

[array length=attrs_len]

attrs_len

length of log_attrs array

 

pango_find_paragraph_boundary ()

void
pango_find_paragraph_boundary (const gchar *text,
                               gint length,
                               gint *paragraph_delimiter_index,
                               gint *next_paragraph_start);

Locates a paragraph boundary in text . A boundary is caused by delimiter characters, such as a newline, carriage return, carriage return-newline pair, or Unicode paragraph separator character. The index of the run of delimiters is returned in paragraph_delimiter_index . The index of the start of the paragraph (index after all delimiters) is stored in next_paragraph_start .

If no delimiters are found, both paragraph_delimiter_index and next_paragraph_start are filled with the length of text (an index one off the end).

Parameters

text

UTF-8 text

 

length

length of text in bytes, or -1 if nul-terminated

 

paragraph_delimiter_index

return location for index of delimiter.

[out]

next_paragraph_start

return location for start of next paragraph.

[out]

pango_default_break ()

void
pango_default_break (const gchar *text,
                     int length,
                     PangoAnalysis *analysis,
                     PangoLogAttr *attrs,
                     int attrs_len);

This is the default break algorithm. It applies Unicode rules without language-specific tailoring, therefore the analyis argument is unused and can be NULL.

See pango_tailor_break() for language-specific breaks.

Parameters

text

text to break. Must be valid UTF-8

 

length

length of text in bytes (may be -1 if text is nul-terminated)

 

analysis

a PangoAnalysis for the text .

[nullable]

attrs

logical attributes to fill in

 

attrs_len

size of the array passed as attrs

 

pango_tailor_break ()

void
pango_tailor_break (const char *text,
                    int length,
                    PangoAnalysis *analysis,
                    int offset,
                    PangoLogAttr *log_attrs,
                    int log_attrs_len);

Apply language-specific tailoring to the breaks in log_attrs , which are assumed to have been produced by pango_default_break().

If offset is not -1, it is used to apply attributes from analysis that are relevant to line breaking.

Parameters

text

text to process. Must be valid UTF-8

 

length

length in bytes of text

 

analysis

PangoAnalysis structure from pango_itemize() for text

 

offset

Byte offset of text from the beginning of the paragraph, or -1 to ignore attributes from analysis

 

log_attrs

array with one PangoLogAttr per character in text , plus one extra, to be filled in.

[array length=log_attrs_len]

log_attrs_len

length of log_attrs array

 

Since: 1.44


pango_shape ()

void
pango_shape (const char *text,
             int length,
             const PangoAnalysis *analysis,
             PangoGlyphString *glyphs);

Given a segment of text and the corresponding PangoAnalysis structure returned from pango_itemize(), convert the characters into glyphs. You may also pass in only a substring of the item from pango_itemize().

It is recommended that you use pango_shape_full() instead, since that API allows for shaping interaction happening across text item boundaries.

Parameters

text

the text to process

 

length

the length (in bytes) of text

 

analysis

PangoAnalysis structure from pango_itemize()

 

glyphs

glyph string in which to store results

 

pango_shape_full ()

void
pango_shape_full (const char *item_text,
                  int item_length,
                  const char *paragraph_text,
                  int paragraph_length,
                  const PangoAnalysis *analysis,
                  PangoGlyphString *glyphs);

Given a segment of text and the corresponding PangoAnalysis structure returned from pango_itemize(), convert the characters into glyphs. You may also pass in only a substring of the item from pango_itemize().

This is similar to pango_shape(), except it also can optionally take the full paragraph text as input, which will then be used to perform certain cross-item shaping interactions. If you have access to the broader text of which item_text is part of, provide the broader text as paragraph_text . If paragraph_text is NULL, item text is used instead.

Parameters

item_text

valid UTF-8 text to shape.

 

item_length

the length (in bytes) of item_text . -1 means nul-terminated text.

 

paragraph_text

text of the paragraph (see details). May be NULL.

[allow-none]

paragraph_length

the length (in bytes) of paragraph_text . -1 means nul-terminated text.

 

analysis

PangoAnalysis structure from pango_itemize().

 

glyphs

glyph string in which to store results.

 

Since: 1.32


pango_shape_with_flags ()

void
pango_shape_with_flags (const char *item_text,
                        int item_length,
                        const char *paragraph_text,
                        int paragraph_length,
                        const PangoAnalysis *analysis,
                        PangoGlyphString *glyphs,
                        PangoShapeFlags flags);

Given a segment of text and the corresponding PangoAnalysis structure returned from pango_itemize(), convert the characters into glyphs. You may also pass in only a substring of the item from pango_itemize().

This is similar to pango_shape_full(), except it also takes flags that can influence the shaping process.

Parameters

item_text

valid UTF-8 text to shape

 

item_length

the length (in bytes) of item_text . -1 means nul-terminated text.

 

paragraph_text

text of the paragraph (see details). May be NULL.

[allow-none]

paragraph_length

the length (in bytes) of paragraph_text . -1 means nul-terminated text.

 

analysis

PangoAnalysis structure from pango_itemize()

 

glyphs

glyph string in which to store results

 

flags

flags influencing the shaping process

 

Since: 1.44

Types and Values

struct PangoItem

struct PangoItem {
  gint offset;
  gint length;
  gint num_chars;
  PangoAnalysis analysis;
};

The PangoItem structure stores information about a segment of text.

Members

gint offset;

byte offset of the start of this item in text.

 

gint length;

length of this item in bytes.

 

gint num_chars;

number of Unicode characters in the item.

 

PangoAnalysis analysis;

analysis results for the item.

 

struct PangoAnalysis

struct PangoAnalysis {
  PangoEngineShape *shape_engine;
  PangoEngineLang  *lang_engine;
  PangoFont *font;

  guint8 level;
  guint8 gravity;
  guint8 flags;

  guint8 script;
  PangoLanguage *language;

  GSList *extra_attrs;
};

The PangoAnalysis structure stores information about the properties of a segment of text.

Members

PangoEngineShape *shape_engine;

unused

 

PangoEngineLang *lang_engine;

unused

 

PangoFont *font;

the font for this segment.

 

guint8 level;

the bidirectional level for this segment.

 

guint8 gravity;

the glyph orientation for this segment (A PangoGravity).

 

guint8 flags;

boolean flags for this segment (Since: 1.16).

 

guint8 script;

the detected script for this segment (A PangoScript) (Since: 1.18).

 

PangoLanguage *language;

the detected language for this segment.

 

GSList *extra_attrs;

extra attributes for this segment.

 

PANGO_ANALYSIS_FLAG_CENTERED_BASELINE

#define PANGO_ANALYSIS_FLAG_CENTERED_BASELINE (1 << 0)

Whether the segment should be shifted to center around the baseline. Used in vertical writing directions mostly.

Since: 1.16


PANGO_ANALYSIS_FLAG_IS_ELLIPSIS

#define PANGO_ANALYSIS_FLAG_IS_ELLIPSIS (1 << 1)

This flag is used to mark runs that hold ellipsized text, in an ellipsized layout.

Since: 1.36.7


PANGO_ANALYSIS_FLAG_NEED_HYPHEN

#define PANGO_ANALYSIS_FLAG_NEED_HYPHEN (1 << 2)

This flag tells Pango to add a hyphen at the end of the run during shaping.

Since: 1.44


PangoLogAttr

typedef struct {
  guint is_line_break               : 1;
  guint is_mandatory_break          : 1;
  guint is_char_break               : 1;
  guint is_white                    : 1;
  guint is_cursor_position          : 1;
  guint is_word_start               : 1;
  guint is_word_end                 : 1;
  guint is_sentence_boundary        : 1;
  guint is_sentence_start           : 1;
  guint is_sentence_end             : 1;
  guint backspace_deletes_character : 1;
  guint is_expandable_space         : 1;
  guint is_word_boundary            : 1;
} PangoLogAttr;

The PangoLogAttr structure stores information about the attributes of a single character.

Members

guint is_line_break : 1;

if set, can break line in front of character

 

guint is_mandatory_break : 1;

if set, must break line in front of character

 

guint is_char_break : 1;

if set, can break here when doing character wrapping

 

guint is_white : 1;

is whitespace character

 

guint is_cursor_position : 1;

if set, cursor can appear in front of character. i.e. this is a grapheme boundary, or the first character in the text. This flag implements Unicode's

Grapheme Cluster Boundaries semantics.
 

guint is_word_start : 1;

is first character in a word

 

guint is_word_end : 1;

is first non-word char after a word Note that in degenerate cases, you could have both is_word_start and is_word_end set for some character.

 

guint is_sentence_boundary : 1;

is a sentence boundary. There are two ways to divide sentences. The first assigns all inter-sentence whitespace/control/format chars to some sentence, so all chars are in some sentence; is_sentence_boundary denotes the boundaries there. The second way doesn't assign between-sentence spaces, etc. to any sentence, so is_sentence_start /is_sentence_end mark the boundaries of those sentences.

 

guint is_sentence_start : 1;

is first character in a sentence

 

guint is_sentence_end : 1;

is first char after a sentence. Note that in degenerate cases, you could have both is_sentence_start and is_sentence_end set for some character. (e.g. no space after a period, so the next sentence starts right away)

 

guint backspace_deletes_character : 1;

if set, backspace deletes one character rather than the entire grapheme cluster. This field is only meaningful on grapheme boundaries (where is_cursor_position is set). In some languages, the full grapheme (e.g. letter + diacritics) is considered a unit, while in others, each decomposed character in the grapheme is a unit. In the default implementation of pango_break(), this bit is set on all grapheme boundaries except those following Latin, Cyrillic or Greek base characters.

 

guint is_expandable_space : 1;

is a whitespace character that can possibly be expanded for justification purposes. (Since: 1.18)

 

guint is_word_boundary : 1;

is a word boundary, as defined by UAX29. More specifically, means that this is not a position in the middle of a word. For example, both sides of a punctuation mark are considered word boundaries. This flag is particularly useful when selecting text word-by-word. This flag implements Unicode's

Word Boundaries semantics. (Since: 1.22)
 

enum PangoShapeFlags

Flags influencing the shaping process. These can be passed to pango_shape_with_flags().

Members

PANGO_SHAPE_NONE

Default value.

 

PANGO_SHAPE_ROUND_POSITIONS

Round glyph positions and widths to whole device units. This option should be set if the target renderer can't do subpixel positioning of glyphs.