2 <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.3//EN"
3 "http://www.oasis-open.org/docbook/xml/4.3/docbookx.dtd" [
4 <!ENTITY % local.common.attrib "xmlns:xi CDATA #FIXED 'http://www.w3.org/2003/XInclude'">
5 <!ENTITY version SYSTEM "version.xml">
7 <chapter id="shaping-and-shape-plans">
8 <title>Shaping and shape plans</title>
10 Once you have your face and font objects configured as desired and
11 your input buffer is filled with the characters you need to shape,
12 all you need to do is call <function>hb_shape()</function>.
15 HarfBuzz will return the shaped version of the text in the same
16 buffer that you provided, but it will be in output mode. At that
17 point, you can iterate through the glyphs in the buffer, drawing
18 each one at the specified position or handing them off to the
19 appropriate graphics library.
22 For the most part, HarfBuzz's shaping step is straightforward from
23 the outside. But that doesn't mean there will never be cases where
24 you want to look under the hood and see what is happening on the
25 inside. HarfBuzz provides facilities for doing that, too.
28 <section id="shaping-buffer-output">
29 <title>Shaping and buffer output</title>
31 The <function>hb_shape()</function> function call takes four arguments: the font
32 object to use, the buffer of characters to shape, an array of
33 user-specified features to apply, and the length of that feature
34 array. The feature array can be NULL, so for the sake of
35 simplicity we will start with that case.
38 Internally, HarfBuzz looks at the tables of the font file to
39 determine where glyph classes, substitutions, and positioning
40 are defined, using that information to decide which
41 <emphasis>shaper</emphasis> to use (<literal>ot</literal> for
42 OpenType fonts, <literal>aat</literal> for Apple Advanced
43 Typography fonts, and so on). It also looks at the direction,
44 script, and language properties of the segment to figure out
45 which script-specific shaping model is needed (at least, in
46 shapers that support multiple options).
49 If a font has a GDEF table, then that is used for
50 glyph classes; if not, HarfBuzz will fall back to Unicode
51 categorization by code point. If a font has an AAT <literal>morx</literal> table,
52 then it is used for substitutions; if not, but there is a GSUB
53 table, then the GSUB table is used. If the font has an AAT
54 <literal>kerx</literal> table, then it is used for positioning; if not, but
55 there is a GPOS table, then the GPOS table is used. If neither
56 table is found, but there is a <literal>kern</literal> table, then HarfBuzz will
57 use the <literal>kern</literal> table. If there is no <literal>kerx</literal>, no GPOS, and no
58 <literal>kern</literal>, HarfBuzz will fall back to positioning marks itself.
61 With a well-behaved OpenType font, you expect GDEF, GSUB, and
62 GPOS tables to all be applied. HarfBuzz implements the
63 script-specific shaping models in internal functions, rather
64 than in the public API.
68 used for shaping can be quite involved; HarfBuzz tries
69 to be compatible with the OpenType Layout specification
70 and, wherever there is any ambiguity, HarfBuzz attempts to replicate the
71 output of Microsoft's Uniscribe engine, to the extent that is feasible and desirable. See the <ulink
72 url="https://docs.microsoft.com/en-us/typography/script-development/standard">Microsoft
73 Typography pages</ulink> for more detail.
76 In general, though, all that you need to know is that
77 <function>hb_shape()</function> returns the results of shaping
78 in the same buffer that you provided. The buffer's content type
80 <literal>HB_BUFFER_CONTENT_TYPE_GLYPHS</literal>, indicating
81 that it contains shaped output, rather than input text. You can
82 now extract the glyph information and positioning arrays:
84 <programlisting language="C">
85 hb_glyph_info_t *glyph_info = hb_buffer_get_glyph_infos(buf, &glyph_count);
86 hb_glyph_position_t *glyph_pos = hb_buffer_get_glyph_positions(buf, &glyph_count);
89 The glyph information array holds a <type>hb_glyph_info_t</type>
90 for each output glyph, which has two fields:
91 <parameter>codepoint</parameter> and
92 <parameter>cluster</parameter>. Whereas, in the input buffer,
93 the <parameter>codepoint</parameter> field contained the Unicode
94 code point, it now contains the glyph ID of the corresponding
95 glyph in the font. The <parameter>cluster</parameter> field is
96 an integer that you can use to help identify when shaping has
97 reordered, split, or combined code points; we will say more
98 about that in the next chapter.
101 The glyph positions array holds a corresponding
102 <type>hb_glyph_position_t</type> for each output glyph,
103 containing four fields: <parameter>x_advance</parameter>,
104 <parameter>y_advance</parameter>,
105 <parameter>x_offset</parameter>, and
106 <parameter>y_offset</parameter>. The advances tell you how far
107 you need to move the drawing point after drawing this glyph,
108 depending on whether you are setting horizontal text (in which
109 case you will have x advances) or vertical text (for which you
110 will have y advances). The x and y offsets tell you where to
111 move to start drawing the glyph; usually you will have both and
112 x and a y offset, regardless of the text direction.
115 Most of the time, you will rely on a font-rendering library or
116 other graphics library to do the actual drawing of glyphs, so
117 you will need to iterate through the glyphs in the buffer and
118 pass the corresponding values off.
122 <section id="shaping-opentype-features">
123 <title>OpenType features</title>
125 OpenType features enable fonts to include smart behavior,
126 implemented as "lookup" rules stored in the GSUB and GPOS
127 tables. The OpenType specification defines a long list of
128 standard features that fonts can use for these behaviors; each
129 feature has a four-character reserved name and a well-defined
133 Some OpenType features are defined for the purpose of supporting
134 script-specific shaping, and are automatically activated, but
135 only when a buffer's script property is set to a script that the
139 Other features are more generic and can apply to several (or
140 any) script, and shaping engines are expected to implement
141 them. By default, HarfBuzz activates several of these features
142 on every text run. They include <literal>abvm</literal>,
143 <literal>blwm</literal>, <literal>ccmp</literal>,
144 <literal>locl</literal>, <literal>mark</literal>,
145 <literal>mkmk</literal>, and <literal>rlig</literal>.
148 In addition, if the text direction is horizontal, HarfBuzz
149 also applies the <literal>calt</literal>,
150 <literal>clig</literal>, <literal>curs</literal>,
151 <literal>dist</literal>, <literal>kern</literal>,
152 <literal>liga</literal> and <literal>rclt</literal>, features.
155 Additionally, when HarfBuzz encounters a fraction slash
156 (<literal>U+2044</literal>), it looks backward and forward for decimal
157 digits (Unicode General Category = Nd), and enables features
158 <literal>numr</literal> on the sequence before the fraction slash,
159 <literal>dnom</literal> on the sequence after the fraction slash,
160 and <literal>frac</literal> on the whole sequence including the fraction
164 Some script-specific shaping models
165 (see <xref linkend="opentype-shaping-models" />) disable some of the
166 features listed above:
171 Hangul: <literal>calt</literal>
176 Indic: <literal>liga</literal>
181 Khmer: <literal>liga</literal>
186 If the text direction is vertical, HarfBuzz applies
187 the <literal>vert</literal> feature by default.
190 Still other features are designed to be purely optional and left
191 up to the application or the end user to enable or disable as desired.
194 You can adjust the set of features that HarfBuzz applies to a
195 buffer by supplying an array of <type>hb_feature_t</type>
196 features as the third argument to
197 <function>hb_shape()</function>. For a simple case, let's just
198 enable the <literal>dlig</literal> feature, which turns on any
199 "discretionary" ligatures in the font:
201 <programlisting language="C">
202 hb_feature_t userfeatures[1];
203 userfeatures[0].tag = HB_TAG('d','l','i','g');
204 userfeatures[0].value = 1;
205 userfeatures[0].start = HB_FEATURE_GLOBAL_START;
206 userfeatures[0].end = HB_FEATURE_GLOBAL_END;
209 <literal>HB_FEATURE_GLOBAL_END</literal> and
210 <literal>HB_FEATURE_GLOBAL_END</literal> are macros we can use
211 to indicate that the features will be applied to the entire
212 buffer. We could also have used a literal <literal>0</literal>
213 for the start and a <literal>-1</literal> to indicate the end of
214 the buffer (or have selected other start and end positions, if needed).
217 When we pass the <varname>userfeatures</varname> array to
218 <function>hb_shape()</function>, any discretionary ligature
219 substitutions from our font that match the text in our buffer
222 <programlisting language="C">
223 hb_shape(font, buf, userfeatures, num_features);
226 Just like we enabled the <literal>dlig</literal> feature by
227 setting its <parameter>value</parameter> to
228 <literal>1</literal>, you would disable a feature by setting its
229 <parameter>value</parameter> to <literal>0</literal>. Some
230 features can take other <parameter>value</parameter> settings;
231 be sure you read the full specification of each feature tag to
232 understand what it does and how to control it.
236 <section id="shaping-shaper-selection">
237 <title>Shaper selection</title>
239 The basic version of <function>hb_shape()</function> determines
240 its shaping strategy based on examining the capabilities of the
241 font file. OpenType font tables cause HarfBuzz to try the
242 <literal>ot</literal> shaper, while AAT font tables cause HarfBuzz to try the
243 <literal>aat</literal> shaper.
246 In the real world, however, a font might include some unusual
247 mix of tables, or one of the tables might simply be broken for
248 the script you need to shape. So, sometimes, you might not
249 want to rely on HarfBuzz's process for deciding what to do, and
250 just tell <function>hb_shape()</function> what you want it to try.
253 <function>hb_shape_full()</function> is an alternate shaping
254 function that lets you supply a list of shapers for HarfBuzz to
255 try, in order, when shaping your buffer. For example, if you
256 have determined that HarfBuzz's attempts to work around broken
257 tables gives you better results than the AAT shaper itself does,
258 you might move the AAT shaper to the end of your list of
259 preferences and call <function>hb_shape_full()</function>
261 <programlisting language="C">
262 char *shaperprefs[3] = {"ot", "default", "aat"};
264 hb_shape_full(font, buf, userfeatures, num_features, shaperprefs);
267 to get results you are happier with.
270 You may also want to call
271 <function>hb_shape_list_shapers()</function> to get a list of
272 the shapers that were built at compile time in your copy of HarfBuzz.
276 <section id="shaping-plans-and-caching">
277 <title>Plans and caching</title>
279 Internally, HarfBuzz uses a structure called a shape plan to
280 track its decisions about how to shape the contents of a
281 buffer. The <function>hb_shape()</function> function builds up the shape plan by
282 examining segment properties and by inspecting the contents of
286 This process can involve some decision-making and
287 trade-offs — for example, HarfBuzz inspects the GSUB and GPOS
288 lookups for the script and language tags set on the segment
289 properties, but it falls back on the lookups under the
290 <literal>DFLT</literal> tag (and sometimes other common tags)
291 if there are actually no lookups for the tag requested.
294 HarfBuzz also includes some work-arounds for
295 handling well-known older font conventions that do not follow
296 OpenType or Unicode specifications, for buggy system fonts, and for
297 peculiarities of Microsoft Uniscribe. All of that means that a
298 shape plan, while not something that you should edit directly in
299 client code, still might be an object that you want to
300 inspect. Furthermore, if resources are tight, you might want to
301 cache the shape plan that HarfBuzz builds for your buffer and
302 font, so that you do not have to rebuild it for every shaping call.
305 You can create a cacheable shape plan with
306 <function>hb_shape_plan_create_cached(face, props,
307 user_features, num_user_features, shaper_list)</function>, where
308 <parameter>face</parameter> is a face object (not a font object,
309 notably), <parameter>props</parameter> is an
310 <type>hb_segment_properties_t</type>,
311 <parameter>user_features</parameter> is an array of
312 <type>hb_feature_t</type>s (with length
313 <parameter>num_user_features</parameter>), and
314 <parameter>shaper_list</parameter> is a list of shapers to try.
317 Shape plans are objects in HarfBuzz, so there are
318 reference-counting functions and user-data attachment functions
320 use. <function>hb_shape_plan_reference(shape_plan)</function>
321 increases the reference count on a shape plan, while
322 <function>hb_shape_plan_destroy(shape_plan)</function> decreases
323 the reference count, destroying the shape plan when the last
324 reference is dropped.
327 You can attach user data to a shaper (with a key) using the
328 <function>hb_shape_plan_set_user_data(shape_plan,key,data,destroy,replace)</function>
329 function, optionally supplying a <function>destroy</function>
330 callback to use. You can then fetch the user data attached to a
332 <function>hb_shape_plan_get_user_data(shape_plan, key)</function>.