GPOS Subtable Examples

The rest of this chapter describes examples of all the GPOS subtable formats, including each of the three formats available for contextual positioning. All the examples reflect unique parameters described below, but the samples provide a useful reference for building subtables specific to other situations.

All the examples have three columns showing hex data, source, and comments.

Example 1: GPOS Header Table

Example 1 shows a typical GPOS Header table definition with offsets to a ScriptList, FeatureList, and LookupList.

Example 1

Hex Data	Source	Comments
	GPOSHeader
	TheGPOSHeader	;GPOSHeader table definition
00010000	0x00010000	;Version
000A	TheScriptList	;offset to ScriptList table
001E	TheFeatureList	;offset to FeatureList table
002C	TheLookupList	;offset to LookupList table

Example 2: SinglePosFormat1 Subtable

Example 2 uses the SinglePosFormat1 subtable to lower the Y placement of subscript glyphs in a font. The LowerSubscriptsSubTable defines one Coverage table, called LowerSubscriptsCoverage, which lists one range of glyph indices for the numeral/numeric subscript glyphs. The subtable's ValueFormat setting indicates that the ValueRecord specifies only the YPlacement value, lowering each subscript glyph by 80 design units.

Example 2

Hex Data	Source	Comments
	SinglePosFormat1
	LowerSubscriptsSubTable	;SinglePos subtable definition
0001	1	;PosFormat
0008	LowerSubscriptsCoverage	;offset to Coverage table
0002	0x0002	;ValueFormat, YPlacement
		;Value[0], move Y position down
FFB0	-80

	CoverageFormat2
	LowerSubscriptsCoverage	;Coverage table definition
0002	2	;CoverageFormat
0001	1	;RangeCount
		;RangeRecord[0]
01B3	ZeroSubscriptGlyphID	;Start, first glyphID
01BC	NineSubscriptGlyphID	;End, last glyphID
0000	0	;StartCoverageIndex

Example 3: SinglePosFormat2 Subtable

This example uses a SinglePosFormat2 subtable to adjust the spacing of three dash glyphs by different amounts. The em dash spacing changes by 10 units, the en dash spacing changes by 25 units, and spacing of the standard dash changes by 50 units.

The DashSpacingSubTable contains one Coverage table with three dash glyph indices, plus an array of ValueRecords, one for each covered glyph. The ValueRecords use the same ValueFormat to modify the XPlacement and XAdvance values of each glyph. The ValueFormat bit setting of 0x0005 is produced by adding the XPlacement and XAdvance bit settings.

Example 3

Hex Data	Source	Comments

	SinglePosFormat2
	DashSpacingSubTable	;SinglePos subtable definition
0002	2	;PosFormat
0014	DashSpacingCoverage	;offset to Coverage table
0005	0x0005	;ValueFormat
		; for XPlacement and XAdvance
0003	3	;ValueCount
		;Value[0], for dash glyph
0032	50	;XPlacement
0032	50	;XAdvance
		;Value[1], for en dash glyph
0019	25	;XPlacement
0019	25	;XAdvance
		;Value[2], for em dash glyph
000A	10	;XPlacement
000A	10	;XAdvance

	CoverageFormat1
	DashSpacingCoverage	;Coverage table definition
0001	1	;CoverageFormat
0003	3	;GlyphCount
004F	DashGlyphID	;GlyphArray[0]
0125	EnDashGlyphID	;GlyphArray[1]
0129	EmDashGlyphID	;GlyphArray[2]

Example 4: PairPosFormat1 Subtable

Example 4 uses a PairPosFormat1 subtable to kern two glyph pairs — "Po" and "To" — by adjusting the XAdvance of the first glyph and the XPlacement of the second glyph. Two ValueFormats are defined, one for each glyph. The subtable contains a Coverage table that lists the index of the first glyph in each pair. It also contains an offset to a PairSet table for each covered glyph.

A PairSet table defines an array of PairValueRecords to specify all the glyph pairs that contain a covered glyph as their first component. In this example, the PPairSet table has one PairValueRecord that identifies the second glyph in the "Po" pair and two ValueRecords, one for the first glyph and one for the second. The TPairSet table also has one PairValueRecord that lists the second glyph in the "To" pair and two ValueRecords, one for each glyph.

Example 4

Hex Data	Source	Comments
	PairPosFormat1
	PairKerningSubTable	;PairPos subtable definition
0001	1	;PosFormat
001E	PairKerningCoverage	;offset to Coverage table
0004	0x0004	;ValueFormat1
		; XAdvance only
0001	0x0001	;ValueFormat2
		; XPlacement only
0002	2	;PairSetCount
000E	PPairSetTable	;PairSet[0]
0016	TPairSetTable	;PairSet[1]

	PairSetTable
	PPairSetTable	;PairSet table definition
0001	1	;PairValueCount
		; one pair in set
		;PairValueRecord[0]
0059	LowercaseOGlyphID	;SecondGlyph
		;Value1
FFE2	-30	; XAdvance adjustment for first ; glyph
		;Value2
FFEC	-20	; XPlacement adjustment for
		; second glyph

	PairSetTable
	PairSetTable	;PairSet table definition
0001	1	;PairValueCount
		; one pair in set
		;PairValueRecord[0]
0059	LowercaseOGlyphID	;SecondGlyph
		;Value1
FFD8	-40	;XAdvance adjustment
		; for first glyph
		;Value2
FFE7	-25	;XPlacement adjustment
		; for second glyph

	CoverageFormat1
	PairKerningCoverage	;Coverage table definition
0001	1	;CoverageFormat
0002	2	;GlyphCount
002D	UppercasePGlyphID	;GlyphArray[0]
0031	UppercaseTGlyphID	;GlyphArray[1]

Example 5: PairPosFormat2 Subtable

The PairPosFormat2 subtable in this example defines pairs composed of two glyph classes. Two ClassDef tables are defined, one for each glyph class. The first glyph in each pair is in a class of lowercase glyphs with diagonal shapes (v, w, y), defined Class1 in the LowercaseClassDef table. The second glyph in each pair is in a class of punctuation glyphs (comma and period), defined in Class1 in the PunctuationClassDef table. The Coverage table only lists the indices of the glyphs in the LowercaseClassDef table since they occupy the first position in the pairs.

The subtable defines two Class1Records for the classes defined in LowecaseClassDef, including Class0. Each record, in turn, defines a Class2Record for each class defined in PunctuationClassDef, including Class0. The Class2Records specify the positioning adjustments for the glyphs.

The pairs are kerned by reducing the XAdvance of the first glyph by 50 design units. Because no positioning change applies to the second glyph, its ValueFormat2 is set to 0, to indicate that Value2 is empty for each pair.

Since no pairs begin with Class0 or Class2 glyphs, all the ValueRecords referenced in Class1Record[0] contain values of 0 or are empty. However, Class1Record[1] does define an XAdvance value in its Class2Record[1] for kerning all pairs that contain a Class1 glyph followed by a Class2 glyph.

Example 5

Hex Data	Source	Comments
	PairPosFormat2
	PunctKerningSubTable	;PairPos subtable definition
0002	2	;PosFormat
0018	PunctKerningCoverage	;offset to Coverage table
0004	0x0004	;ValueFormat1
		; XAdvance only
0000	0	;ValueFormat2
		; no ValueRecord for second
		; glyph
0022	LowercaseClassDef	;offset to ClassDef1 table
		; for first class in pair
0032	PunctuationClassDef	;offset to ClassDef2 table
		; for second class in pair
0002	2	;Class1Count
0002	2	;Class2Count
		;Class1Record[0]
		; no contexts begin with
		; Class0
		;Class2Record[0]
0000	0	;Value1
		; no change for first glyph
		;Value2
		; no ValueRecord for second
		; glyph
		;Class2Record[1]
0000	0	;Value1
		; no change for first glyph
		;Value2
		; no ValueRecord for second
		; glyph
		;Class1Record[1]
		; for contexts beginning with
		; Class1
		;Class2Record[0]
		; no contexts with Class0 as
		; second glyph
0000	0	;Value1
		; no change for first glyph
		;Value2
		; no ValueRecord for second
		; glyph
		;Class2Record[1]
		; contexts with Class1 as
		; second glyph
FFCE	-50	;Value1
		; move punctuation glyph left
		;Value2
		; no ValueRecord for second
		; glyph


	CoverageFormat1
	PunctKerningCoverage	;Coverage table definition
0001	1	;CoverageFormat, lists
0003	3	;GlyphCount
0046	LowercaseVGlyphID	;GlyphArray[0]
0047	LowercaseWGlyphID	;GlyphArray[1]
0049	LowercaseYGlyphID	;GlyphArray[2]

	ClassDefFormat2
	LowercaseClassDef	;ClassDef table definition
0002	2	;ClassFormat
0002	2	;ClassRangeCount
		;ClassRangeRecord[0]
0046	LowercaseVGlyphID	;Start
0047	LowercaseWGlyphID	;End
0001	1	;Class
		;ClassRangeRecord[1]
0049	LowercaseYGlyphID	;Start
0049	LowercaseYGlyphID	;End
0001	1	;Class

	ClassDefFormat2
	PunctuationClassDef	;ClassDef table definition
0002	2	;ClassFormat
0001	1	;ClassRangeCount
		;ClassRangeRecord[0]
006A	PeriodPunctGlyphID	;Start
006B	CommaPunctGlyphID	;End
0001	1	;Class

Example 6: CursivePosFormat1 Subtable

In Example 6, the Urdu language system uses a CursivePosFormat1 subtable to attach glyphs along a diagonal baseline that descends from right to left. Two glyphs are defined with attachment data and listed in the Coverage table—the Kaf and Ha glyphs. For each glyph, the subtable contains an EntryExitRecord that defines offsets to two Anchor tables, an entry attachment point, and an exit attachment point. Each Anchor table defines X and Y coordinate values. To render Urdu down and diagonally, the entry point's Y coordinate is above the baseline and the exit point's Y coordinate is located below the baseline.

Example 6

Hex Data	Source	Comments

	CursivePosFormat1
	DiagonalWritingSubTable	;CursivePos subtable definition
0001	1	;PosFormat
000E	DiagonalWritingCoverage	;offset to Coverage table
0002	2	;EntryExitCount
		;EntryExitRecord[0]
		; for Kaf glyph
0016	KafEntryAnchor	;offset to EntryAnchor table
001C	KafExitAnchor	;offset to ExitAnchor table
		;EntryExitRecord[1]
		; for Ha glyph
0022	HaEntryAnchor	;offset to EntryAnchor table
0028	HaExitAnchor	;offset to ExitAnchor table

	CoverageFormat1
	DiagonalWritingCoverage	;Coverage table definition
0001	1	;CoverageFormat
0002	2	;GlyphCount
0203	KafGlyphID	;GlyphArray[0]
027E	HaGlyphID	;GlyphArray[1]

	AnchorFormat1
	KafEntryAnchor	;Anchor table definition
0001	1	;AnchorFormat
05DC	1500	;XCoordinate
002C	44	;YCoordinate

	AnchorFormat1
	KafExitAnchor	;Anchor table definition
0001	1	;AnchorFormat
0000	0	;XCoordinate
FFEC	-20	;YCoordinate

	AnchorFormat1
	HaEntryAnchor	;Anchor table definition
0001	1	;AnchorFormat
05DC	1500	;XCoordinate
002C	44	;YCoordinate

	AnchorFormat1
	HaExitAnchor	;Anchor table definition
0001	1	;AnchorFormat
0000	0	;XCoordinate
FFEC	-20	;Ycoordinate

Example 7: MarkBasePosFormat1 Subtable

The MarkBasePosFormat1 subtable in Example 7 defines one Arabic base glyph, Tah, and two Arabic mark glyphs: a fathatan mark above the base glyph, and a kasra mark below the base glyph. The BaseGlyphsCoverage table lists the base glyph, and the MarkGlyphsCoverage table lists the mark glyphs.

Each mark is also listed in the MarkArray, along with its attachment point data and a mark Class value. The MarkArray defines two mark classes: Class0 consists of marks located above base glyphs, and Class1 consists of marks located below base glyphs.

The BaseArray defines attachment data for base glyphs. In this array, one BaseRecord is defined for the Tah glyph with offsets to two BaseAnchor tables, one for each class of marks. AboveBaseAnchor defines an attachment point for marks placed above the Tah base glyph, and BelowBaseAnchor defines an attachment point for marks placed below it.

Example 7

Hex Data	Source	Comments
	MarkBasePosFormat1
	MarkBaseAttachSubTable	;MarkBasePos subtable definition
0001	1	;PosFormat
000C	MarkGlyphsCoverage	;offset to MarkCoverage table
0014	BaseGlyphsCoverage	;offset to BaseCoverage table
0002	2	;ClassCount
001A	MarkGlyphsArray	;offset to MarkArray table
0030	BaseGlyphsArray	;offset to BaseArray table

	CoverageFormat1
	MarkGlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0002	2	;GlyphCount
0333	fathatanMarkGlyphID	;GlyphArray[0]
033F	kasraMarkGlyphID	;GlyphArray[1]

	CoverageFormat1
	BaseGlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0001	1	;GlyphCount
0190	tahBaseGlyphID	;GlyphArray[0]

	MarkArray
	MarkGlyphsArray	;MarkArray table definition
0002	2	;MarkCount
		;MarkRecord[0]
		; in CoverageIndex order
0000	0	;Class, for marks over base
000A	fathatanMarkAnchor	;offset to Anchor table
		;MarkRecord[1]
0001	1	;Class, for marks under
0010	kasraMarkAnchor	;offset to Anchor table

	AnchorFormat1
	fathatanMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
015A	346	;XCoordinate
FF9E	-98	;YCoordinate

	AnchorFormat1
	kasraMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
0105	261	;XCoordinate
0058	88	;YCoordinate

	BaseArray
	BaseGlyphsArray	;BaseArray table definition
0001	1	;BaseCount
		;BaseRecord[0]
0006	AboveBaseAnchor	;BaseAnchor[0]
000C	BelowBaseAnchor	;BaseAnchor[1]

	AnchorFormat1
	AboveBaseAnchor	;Anchor table definition
0001	1	;AnchorFormat
033E	830	;XCoordinate
0640	1600	;YCoordinate

	AnchorFormat1
	BelowBaseAnchor	;Anchor table definition
0001	1	;AnchorFormat
033E	830	;XCoordinate
FFAD	-83	;Ycoordinate

Example 8: MarkLigPosFormat1 Subtable

Example 8 uses the MarkLigPosFormat1 subtable to attach marks to a ligature glyph in the Arabic script. The hypothetical ligature is composed of three glyph components: a Lam (initial form), a meem (medial form), and a jeem (medial form). Accent marks are defined for the first two components: the sukun accent is positioned above lam, and the kasratan accent is placed below meem.

The LigGlyphsCoverage table lists the ligature glyph and the MarkGlyphsCoverage table lists the two accent marks. Each mark is also listed in the MarkArray, along with its attachment point data and a mark Class value. The MarkArray defines two mark classes: Class0 consists of marks located above base glyphs, and Class1 consists of marks located below base glyphs.

The LigGlyphsArray has an offset to one LigatureAttach table for the covered ligature glyph. This table, called LamWithMeemWithJeemLigAttach, defines a count and array of the component glyphs in the ligature. Each ComponentRecord defines offsets to two Anchor tables, one for each mark class.

In the example, the first glyph component, lam, specifies a high attachment point for positioning accents above, but does not specify a low attachment point for placing accents below. The second glyph component, meem, defines a low attachment point for placing accents below, but not above. The third component, jeem, has no attachment points since the example defines no accents for it.

Example 8

Hex Data	Source	Comments
	MarkLigPosFormat1
	MarkLigAttachSubTable	;MarkLigPos subtable definition
0001	1	;PosFormat
000C	MarkGlyphsCoverage	;offset to MarkCoverage table
0014	LigGlyphsCoverage	;offset to LigatureCoverage
		; table
0002	2	;ClassCount
001A	MarkGlyphsArray	;offset to MarkArray table
0030	LigGlyphsArray	;offset to LigatureArray table

	CoverageFormat1
	MarkGlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0002	2	;GlyphCount
033C	sukunMarkGlyphID	;GlyphArray[0]
033F	kasratanMarkGlyphID	;GlyphArray[1]

	CoverageFormat1
	LigGlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0001	1	;GlyphCount
0234	LamWithMeemWithJeemLigatureGlyphID ;GlyphArray[0]

	MarkArray
	MarkGlyphsArray	;MarkArray table definition
0002	2	;MarkCount
		;MarkRecord[0]
		; in CoverageIndex order
0000	0	;Class
		; for marks above components
000A	sukunMarkAnchor	;offset to Anchor table
		;MarkRecord[1]
0001	1	;Class, for marks below
		; components
0010	kasratanMarkAnchor	;offset to Anchor table

	AnchorFormat1
	sukunMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
015A	346	;XCoordinate
FF9E	-98	;YCoordinate

	AnchorFormat1
	kasratanMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
0105	261	;XCoordinate
01E8	488	;YCoordinate

	LigatureArray
	LigGlyphsArray	;LigatureArray table definition
0001	1	;LigatureCount
0004	LamWithMeemWithJeemLigAttach	;offset to LigatureAttach table

	LigatureAttach
	LamWithMeemWithJeemLigAttach	;LigatureAttach table definition
0003	3	;ComponentCount
		;ComponentRecord[0]
		; Right-to-Left text order
000E	AboveLamAnchor	;offset to LigatureAnchor table
		; ordered by mark class value
		; for Class0 marks (above)
0000	NULL	;offset to LigatureAnchor table
		; no attachment points for
		; Class1 marks
		;ComponentRecord[1]
0000	NULL	;offset to LigatureAnchor table
		; no attachment points for
		; Class0 marks
0014	BelowMeemAnchor	;offset to LigatureAnchor table
		; for Class1 marks (below)
		;ComponentRecord[2]
0000	NULL	;offset to LigatureAnchor table
		; no attachment points for
		; Class0 marks
0000	NULL	;offset to LigatureAnchor table
		; no attachment points for
		; Class1 marks

	AnchorFormat1
	AboveLamAnchor	;Anchor table definition
0001	1	;AnchorFormat
0271	625	;XCoordinate
0708	1800	;YCoordinate

	AnchorFormat1
	BelowMeemAnchor	;Anchor table definition
0001	1	;AnchorFormat
0178	376	;XCoordinate
FE90	-368	;Ycoordinate

Example 9: MarkMarkPosFormat1 Subtable

The MarkMarkPosFormat1 subtable in Example 9 defines two Arabic marks glyphs. The hanza mark, the base mark (Mark2), is identified in the Mark2GlyphsCoverage table. The damma mark, the attaching mark (Mark1), is defined in the Mark1GlyphsCoverage table.

Each Mark1 glyph is also listed in the Mark1Array, along with its attachment point data and a mark Class value. The Mark1GlyphsArray defines one mark class, Class0, that consists of marks located above Mark2 base glyphs. The Mark1GlyphsArray contains an offset to a dammaMarkAnchor table to specify the coordinate of the damma mark's attachment point.

The Mark2GlyphsArray table defines a count and an array of Mark2Records, one for each covered Mark2 base glyph. Each record contains an offset to a Mark2Anchor table for each Mark1 class. One Anchor table, AboveMark2Anchor, specifies a coordinate value for attaching the damma mark above the hanza base mark.

Example 9

Hex Data	Source	Comments
	MarkMarkPosFormat1
	MarkMarkAttachSubTable	;MarkBasePos subtable definition
0001	1	;PosFormat
000C	Mark1GlyphsCoverage	;offset to Mark1Coverage table
0012	Mark2GlyphsCoverage	;offset to Mark2Coverage table
0001	1	;ClassCount
0018	Mark1GlyphsArray	;offset to Mark1Array table
0024	Mark2GlyphsArray	;offset to Mark2Array table

	CoverageFormat1
	Mark1GlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0001	1	;GlyphCount
0296	dammaMarkGlyphID	;GlyphArray[0]

	CoverageFormat1
	Mark2GlyphsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0001	1	;GlyphCount
0289	hanzaMarkGlyphID	;GlyphArray[1]

	MarkArray
	Mark1GlyphsArray	;MarkArray table definition
0001	1	;MarkCount
		;MarkRecord[0]
		; in CoverageIndex order
0000	0	;Class
		; for marks above base mark
0006	dammaMarkAnchor	;offset to Anchor table

	AnchorFormat1
	dammaMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
00BD	189	;XCoordinate
FF99	-103	;YCoordinate

	Mark2Array
	Mark2GlyphsArray	;Mark2Array table definition
0001	1	;Mark2Count
		;Mark2Record[0]
0004	AboveMark2Anchor	;offset to Anchor table[0]

	AnchorFormat1
	AboveMark2Anchor	;Anchor table definition
0001	1	;AnchorFormat
00DD	221	;XCoordinate
012D	301	;Ycoordinate

Example 10: ContextPosFormat1 Subtable and PosLookupRecord

Example 10 uses a ContextPosFormat1 subtable to adjust the spacing between three Arabic glyphs in a word. The context is the glyph sequence (from right to left): heh (initial form), thal (final form), and heh (isolated form). In the rendered word, the first two glyphs are connected, but the last glyph (the isolated form of heh), is separate. This subtable reduces the amount of space between the last glyph and the rest of the word.

The subtable contains a WordCoverage table that lists the first glyph in the word, heh (initial), and one PosRuleSet table, called WordPosRuleSet, that defines all contexts beginning with this covered glyph.

The WordPosRuleSet contains one PosRule that describes a word context of three glyphs and identifies the second and third glyphs (the first glyph is identified by the WordPosRuleSet). When a text-processing client locates this context in text, it applies a SinglePos lookup (not shown in the example) at position 3 to reduce the spacing between the glyphs.

Example 10

Hex Data	Source	Comments
	ContextPosFormat1
	MoveHehInSubtable	;ContextPos subtable definition
0001	1	;PosFormat
0008	WordCoverage	;offset to Coverage table
0001	1	;PosRuleSetCount
000E	WordPosRuleSet	;offset to PosRuleSet[0] table

	CoverageFormat1
	WordCoverage	;Coverage table offset
0001	1	;CoverageFormat
0001	1	;GlyphCount
02A6	hehInitialGlyphID	;GlyphArray[0]

	PosRuleSet
	WordPosRuleSet	;PosRuleSet table definition
0001	1	;PosRuleCount
0004	WordPosRule	;Offset to PosRule[0] table

	PosRule
	WordPosRule	;PosRule table definition
0003	3	;GlyphCount
0001	1	;PosCount
02DD	thalFinalGlyphID	;Input[1]
02C6	hehIsolatedGlyphID	;Input[0]
		;PosLookupRecord[0]
0003	3	;SequenceIndex
0001	1	;LookupListIndex

Example 11: ContextPosFormat2 Subtable

The ContextPosFormat2 subtable in Example 11 defines context strings for five glyph classes: Class1 consists of uppercase glyphs that overhang and create a wide open space on their right side; Class2 consists of uppercase glyphs that overhang and create a narrow space on their right side; Class3 contains lowercase x-height vowels; and Class4 contains accent glyphs placed over the lowercase vowels. The rest of the glyphs in the font fall into Class0.

The MoveAccentsSubtable defines two similar context strings. The first consists of a Class1 uppercase glyph followed by a Class3 lowercase vowel glyph with a Class4 accent glyph over the vowel. When this context is found in the text, the client lowers the accent glyph over the vowel so that it does not collide with the overhanging glyph shape. The second context consists of a Class2 uppercase glyph, followed by a Class3 lowercase vowel glyph with a Class4 accent glyph over the vowel. When this context is found in the text, the client increases the advance width of the uppercase glyph to expand the space between it and the accented vowel.

The MoveAccents subtable defines a MoveAccentsCoverage table that identifies the first glyphs in the two contexts and offsets to five PosClassSet tables, one for each class defined in the ClassDef table. Since no contexts begin with Class0, Class3, or Class4 glyphs, the offsets to the PosClassSet tables for these classes are NULL. PosClassSet[1] defines all contexts beginning with Class1 glyphs; it is called UCWideOverhangPosClass1Set. PosClassSet[2] defines all contexts beginning with Class2 glyphs, and it is called UCNarrowOverhangPosClass1Set.

Each PosClassSet defines one PosClassRule. The UCWideOverhangPosClass1Set uses the UCWideOverhangPosClassRule to specify the first context. The first class in this context string is identified by the PosClassSet that includes a PosClassRule, in this case Class1. The PosClassRule table lists the second and third classes in the context as Class3 and Class4. A SinglePos Lookup (not shown) lowers the accent glyph in position 3 in the context string.

The UCNarrowOverhangPosClass1Set defines the UCNarrowOverhangPosClassRule for the second context. This PosClassRule is identical to the UCWideOverhangPosClassRule, except that the first class in the context string is a Class2 lowercase glyph. A SinglePos Lookup (not shown) increases the advance width of the overhanging uppercase glyph in position 0 in the context string.

Example 11

Hex Data	Source	Comments

	ContextPosFormat2
	MoveAccentsSubtable	;ContextPos subtable definition
0002	2	;PosFormat
0012	MoveAccentsCoverage	;Offset to Coverage table
0020	MoveAccentsClassDef	;Offset to ClassDef
0005	5	;PosClassSetCnt
0000	NULL	;PosClassSet[0]
		; no contexts begin with Class0
		; glyphs
0060	UCWideOverhangPosClass1Set	;PosClassSet[1]
		; contexts beginning with Class1
		; glyphs
0070	UCNarrowOverhangPosClass2Set	;PosClassSet[2]
		; context beginning with Class2
		; glyphs
0000	NULL	;PosClassSet[3]
		; no contexts begin with Class3
		; glyphs
0000	NULL	;PosClassSet[4]
		; no contexts begin with Class4
		; glyphs

	CoverageFormat1
	MoveAccentsCoverage	;Coverage table definition
0001	1	;CoverageFormat
0005	5	;GlyphCount
0029	UppercaseFGlyphID	;GlyphArray[0]
0033	UppercasePGlyphID	;GlyphArray[1]
0037	UppercaseTGlyphID	;GlyphArray[2]
0039	UppercaseVGlyphID	;GlyphArray[3]
003A	UppercaseWGlyphID	;GlyphArray[4]

	ClassDefFormat2
	MoveAccentsClassDef	;ClassDef table definition
		; defines five classes = 0 (all
		; else), 1 (T, V, W: UCUnderhang),
		; 2 (F, P: UCOverhang), 3 (a, e,
		; I, o, u: LCVowels), 4 (tilde,
		; umlaut)
0002	2	;ClassFormat, ranges
000A	10	;ClassRangeCount
		;ClassRangeRecord[0]
0029	UppercaseFGlyphID	;Start
0029	UppercaseFGlyphID	;End
0002	2	;Class
		;ClassRangeRecord[1]
0033	UppercasePGlyphID	;Start
0033	UppercasePGlyphID	;End
0002	2	;Class
		;ClassRangeRecord[2]
0037	UppercaseTGlyphID	;Start
0037	UppercaseTGlyphID	;End
0001	1	;Class
		;ClassRangeRecord[3]
0039	UppercaseVGlyphID	;Start
003A	UppercaseWGlyphID	;End
0001	1	;Class
		;ClassRangeRecord[4]
0042	LowercaseAGlyphID	;Start
0042	LowercaseAGlyphID	;End
0003	3	;Class
		;ClassRangeRecord[5]
0046	LowercaseEGlyphID	;Start
0046	LowercaseEGlyphID	;End
0003	3	;Class
		;ClassRangeRecord[6]
004A	LowercaseIGlyphID	;Start
004A	LowercaseIGlyphID	;End
0003	3	;Class
		;ClassRangeRecord[7]
0051	LowercaseOGlyphID	;Start
0051	LowercaseOGlyphID	;End
0003	3	;Class
		;ClassRangeRecord[8]
0056	LowercaseUGlyphID	;Start
0056	LowercaseUGlyphID	;End
0003	3	;Class
		;ClassRangeRecord[9]
00F5	TildeAccentGlyphID	;Start
00F6	UmlautAccentGlyphID	;End
0004	4	;Class

	PosClassSet
	UCWideOverhangPosClass1Set	;PosClassSet table definition
0001	1	;PosClassRuleCnt
0004	UCWideOverhangPosClassRule	;PosClassRule[0]

	PosClassRule
	UCWideOverhangPosClassRule	;PosClassRule table definition
0003	3	;GlyphCount
0001	1	;PosCount
0003	3	;Class[1]
		; lowercase vowel
0004	4	;Class[2]
		; accent
		;PosLookupRecord[0]
0002	2	;SequenceIndex
0001	1	;LookupListIndex
		; lower the accent

	PosClassSet
	UCNarrowOverhangPosClass2Set	;PosClassSet table definition
0001	1	;PosClassRuleCnt
0004	UCNarrowOverhangPosClassRule	;PosClassRule[0]

	PosClassRule
	UCNarrowOverhangPosClassRule	;PosClassRule table definition
0003	3	;GlyphCount
0001	1	;PosCount
0003	3	;Class[1]
		; lowercase vowel
0004	4	;Class[2]
		; accent
		;PosLookupRecord[0]
0000	0	;SequenceIndex
0002	2	;LookupListIndex
		; increase overhang advance
		; width

Example 12: ContextPosFormat3 Subtable

Example 12 uses a ContextPosFormat3 subtable to lower the position of math signs in math equations consisting of a lowercase descender or x-height glyph, a math sign glyph, and any lowercase glyph. Format3 is better to use for this context than the class-based Format2 because the sets of covered glyphs for positions 0 and 2 overlap.

The LowerMathSignsSubtable contains offsets to three Coverage tables (XhtDescLCCoverage, MathSignCoverage, and LCCoverage), one for each position in the context glyph string. When the client finds the context in the text stream, it applies the PosLookupRecord data at position 1 and repositions the math sign.

Example 12

Hex Data	Source	Comments
	ContextPosFormat3
	LowerMathSignsSubtable	;ContextPos subtable definition
0003	3	;PosFormat
0003	3	;GlyphCount
0001	1	;PosLookup
0010	XhtDescLCCoverage	;Offset to Coverage[0] table
003C	MathSignCoverage	;Offset to Coverage[1] table
0044	LCCoverage	;Offset to Coverage[2] table
		;PosLookupRecord[0]
0001	1	;SequenceIndex
0001	1	;LookupListIndex

	CoverageFormat1
	XhtDescLCCoverage	;Coverage table definition
0001	1	;CoverageFormat
0014	20	;GlyphCount
0033	LCaGlyphID	;GlyphArray[0]
0035	LCcGlyphID	;GlyphArray[1]
0037	LCeGlyphID	;GlyphArray[2]
0039	LCgGlyphID	;GlyphArray[3]
003B	LCiGlyphID	;GlyphArray[4]
003C	LCjGlyphID	;GlyphArray[5]
003F	LCmGlyphID	;GlyphArray[6]
0040	LCnGlyphID	;GlyphArray[7]
0041	LCoGlyphID	;GlyphArray[8]
0042	LCpGlyphID	;GlyphArray[9]
0043	LCqGlyphID	;GlyphArray[10]
0044	LCrGlyphID	;GlyphArray[11]
0045	LCsGlyphID	;GlyphArray[12]
0046	LCtGlyphID	;GlyphArray[13]
0047	LCuGlyphID	;GlyphArray[14]
0048	LCvGlyphID	;GlyphArray[15]
0049	LCwGlyphID	;GlyphArray[16]
004A	LCxGlyphID	;GlyphArray[17]
004B	LCyGlyphID	;GlyphArray[18]
004C	LCzGlyphID	;GlyphArray[19]

	CoverageFormat1
	MathSignCoverage	;Coverage table definition
0001	1	;CoverageFormat
0002	2	;GlyphCount
011E	EqualsSignGlyphID	;GlyphArray[0]
012D	PlusSignGlyphID	;GlyphArray[1]


	CoverageFormat2
	LCCoverage	;Coverage table definition
0002	2	;CoverageFormat
0001	1	;RangeCount
		;RangeRecord[0]
0033	LCaGlyphID	;Start
004C	LCzGlyphID	;End
0000	0	;StartCoverageIndex

Example 13: PosLookupRecord

The PosLookupRecord in Example 13 identifies a lookup to apply at the second glyph position in a context glyph string.

Example 13

Hex Data	Source	Comments
	PosLookupRecord
	PosLookupRecord[0]	;PosLookupRecord definition
0001	1	;SequenceIndex
		; for second glyph position
0001	1	;LookupListIndex
		; apply this lookup to second
		; glyph position

Example 14: ValueFormat Table and ValueRecord

Example 14 demonstrates how to specify positioning values in the GPOS table. Here, a SinglePosFormat1 subtable defines the ValueFormat and ValueRecord. The ValueFormat bit setting of 0x0099 says that the corresponding ValueRecord contains values for a glyph's XPlacement and YAdvance. Device tables specify pixel adjustments for these values at font sizes from 11 ppem to 15 ppem.

Example 14

Hex Data	Source	Comments
	SinglePosFormat1
	OnesSubtable	;SinglePos subtable definition
0001	1	;PosFormat
000E	Cov	;Offset to Coverage table
0099	0x0099	;ValueFormat
		; for XPlacement, YAdvance, XPlaDevice, YAdvaDevice
		;Value
0050	80	; Xplacement value
00D2	210	; Yadvance value
0018	XPlaDeviceTable	;Offset to XPlaDevice table
0020	YAdvDeviceTable	;Offset to YAdvDevice table

	CoverageFormat2
	Cov	;Coverage table definition
0002	2	;CoverageFormat
0001	1	;RangeCount
		;RangeRecord[0]
00C8	200	;Start
		; first glyph ID in range
00D1	209	;End
		; last glyph ID in range
0000	0	;StartCoverageIndex


	DeviceTableFormat1
	XPlaDeviceTable	;Device Table definition
000B	11	;StartSize
000F	15	;EndSize
0001	1	;DeltaFormat
	1	; increase 11ppem by 1 pixel
	1	; increase 12ppem by 1 pixel
	1	; increase 13ppem by 1 pixel
	1	; increase 14ppem by 1 pixel
5540	1	; increase 15ppem by 1 pixel

	DeviceTableFormat1
	YAdvDeviceTable	;Device Table definition
000B	11	;StartSize
000F	15	;EndSize
0001	1	;DeltaFormat
	1	; increase 11ppem by 1 pixel
	1	; increase 12ppem by 1 pixel
	1	; increase 13ppem by 1 pixel
	1	; increase 14ppem by 1 pixel
5540	1	; increase 15ppem by 1 pixel

Example 15: AnchorFormat1 Table

Example 15 illustrates an Anchor table for the damma mark glyph in the Arabic script. Format1 is used to specify X and Y coordinate values in design units.

Example 15

Hex Data	Source	Comments
	AnchorFormat1
	dammaMarkAnchor	;Anchor table definition
0001	1	;AnchorFormat
00BD	189	;XCoordinate
FF99	-103	;YCoordinate

Example 16: AnchorFormat2 Table

Example 16 shows an AnchorFormat2 table for an attachment point placed above a base glyph. With this format, the coordinate value for the Anchor depends on the final position of a specific contour point on the base glyph after hinting. The coordinates are specified in design units.

Example 16

Hex Data	Source	Comments
	AnchorFormat2
	AboveBaseAnchor	;Anchor table definition
0002	2	;AnchorFormat
0142	322	;XCoordinate
0384	900	;Ycoordinate
000D	13	;AnchorPoint
		; glyph contour point index

Example 17: AnchorFormat3 Table

Example 17 shows an AnchorFormat3 table that specifies an attachment point above a base glyph. Device tables modify the X and Y coordinates of the Anchor for the point size and resolution of the output font. Here, the Device tables define pixel adjustments for font sizes from 12 ppem to 17 ppem.

Example 17

Hex Data	Source	Comments
	AnchorFormat3
	AboveBaseAnchor	;Anchor table definition
0003	3	;AnchorFormat
0117	279	;XCoordinate
0515	1301	;YCoordinate
000A	XDevice	;offset to DeviceTable
		; for X coordinate
0014	YDevice	;offset to Device table
		; for Y coordinate

	DeviceTableFormat2
	XDevice	;Device Table definition
000C	12	;StartSize
0011	17	;EndSize
0002	2	;DeltaFormat
	1	; increase 12ppem by 1 pixel
	1	; increase 13ppem by 1 pixel
	1	; increase 14ppem by 1 pixel
1111	1	; increase 15ppem by 1 pixel
	2	; increase 16ppem by 1 pixel
2200	2	; increase 17ppem by 1 pixel

	DeviceTableFormat2
	YDevice	;Device Table definition
000C	12	;StartSize
0011	17	;EndSize
0002	2	;DeltaFormat
	1	; increase 12ppem by 1 pixel
	1	; increase 13ppem by 1 pixel
	1	; increase 14ppem by 1 pixel
1111	1	; increase 15ppem by 1 pixel
	2	; increase 16ppem by 1 pixel
2200	2	; increase 17ppem by 1 pixel

Example 18: MarkArray Table and MarkRecord

Example 18 shows a MarkArray table with class and attachment point data for two accent marks, a grave and a cedilla. Two MarkRecords are defined, one for each covered mark glyph. The first MarkRecord assigns a mark class value of 0 to accents placed above base glyphs, such as the grave, and has an offset to a graveMarkAnchor table. The second MarkRecord assigns a mark class value of 1 for all accents positioned below base glyphs, such as the cedilla, and has an offset to a cedillaMarkAnchor table.

Example 18

Hex Data	Source	Comments
	MarkArray
	MarkGlyphsArray	;MarkArray table definition
0002	2	;MarkCount
		;MarkRecord[0]
		;for first mark in
		; MarkCoverage table,
		; grave
0000	0	;Class, for marks placed above
		; base glyphs
000A	graveMarkAnchor	;offset to Anchor table
		;MarkRecord[1]
		;for second mark in
		; MarkCoverage table =
		; cedilla
0001	1	;Class, for marks placed below
		; base glyphs
0010	cedillaMarkAnchor	;offset to Anchor table