|
  
TrueEdit Technical Reference
This appendix documents miscellaneous internal technical details
of TrueEdit which may be of interest to font developers.
TrueEdit source table formats
Much of the binary data in the mort, kern, and just tables is compiled from source-format tables. These source tables
have names of the form TRnn, where nn is a two-digit number. (The EDT0 and EDT1 tables are not used by TrueEdit.)
In general, source tables are introduced only when necessary;
whenever possible, the binary data is decompiled instead. One
kind of data that is very difficult (or impossible) to decompile
is a state table; therefore, for each kind of state table that
may be present in a font, a corresponding source table has been
defined.
The usage and formats of all the TRnn tables currently used by TrueEdit are documented below.
TR00: Required Ligatures
The TR00 table is used to compile the mort subtable that implements the Required Ligatures feature setting.
It has the following format:
| Type |
Name |
Description |
uint16 |
nLigs |
The number of ligatures in the table |
SrcLig |
ligArray[nLigs] |
The ligatures list, sorted in increasing dictionary order of the
component glyph ids. (By "dictionary order", what is meant is
the ordering of strings as in a dictionary, i.e. sorting on the
first glyph, then on the second, then on the third, and so on.)
|
Each SrcLig has the following format:
| Type |
Name |
Description |
uint16 |
ligGlyphId |
The glyph ID of the ligature |
uint16 |
nComps |
The number of component glyphs in the ligature (2 or 3) |
uint16 |
compGlyphIds[3] |
The glyph IDs of the component glyphs |
TR01: Common Ligatures
The TR01 table is used to compile the mort subtable that implements the Common Ligatures feature setting.
It has the same format as the TR00 table.
TR02: Rare Ligatures
The TR02 table is used to compile the mort subtable that implements the Rare Ligatures feature setting.
It has the same format as the TR00 table.
TR03: Logos
The TR03 table is used to compile the mort subtable that implements the Logos feature setting. It has the
same format as the TR00 table.
TR04: Rebus Pictures
The TR04 table is used to compile the mort subtable that implements the Rebus Pictures feature setting.
It has the same format as the TR00 table.
TR05: Cursive
The TR05 table is used to compile the mort subtable that implements the Cursive feature setting.
 At the moment, this table is capable of implementing only the
kind of cursive script found in Arabic text. It is not capable
of implementing Roman cursive text. At some point in the future
TrueEdit may support Roman cursive, at which time the format of
this table will have to change.
The TR05 table currently has the following format:
| Type |
Name |
Description |
int32 |
numQuadruples |
The number of cursive entries in the table |
int32 |
numKashidas |
The number of kashida glyphs in the table |
int32 |
numVowels |
The number of vowel glyphs in the table |
uint16 |
firstKashida |
The smallest kashida glyph ID in the table |
uint16 |
lastKashida |
The largest kashida glyph ID in the table |
uint16 |
firstVowel |
The smallest vowel glyph ID in the table |
uint16 |
lastVowel |
The largest vowel glyph ID in the table |
uint16 |
firstSole |
The smallest standalone glyph ID in the table |
uint16 |
lastSole |
The largest standalone glyph ID in the table |
uint16 |
firstFinal |
The smallest final glyph ID in the table |
uint16 |
lastFinal |
The largest final glyph ID in the table |
uint16 |
kashidas[numKashidas] |
The glyph IDs of the kashidas in the table, in increasing glyph
ID order |
uint16 |
vowels[numVowels] |
The glyph IDs of the vowels in the table, in increasing glyph
ID order |
QuadrupleRec |
quads[numQuadruples] |
The list of cursive entries in the table, sorted in increasing
order of the glyph ID of the standalone form of the glyph in each
cursive entry |
Each QuadrupleRec has the following format:
| Type |
Name |
Description |
uint16 |
sole |
The glyph ID of the standalone form of the glyph |
uint16 |
final |
The glyph ID of the final form of the glyph |
uint16 |
medial |
The glyph ID of the medial form of the glyph |
uint16 |
initial |
The glyph ID of the initial form of the glyph |
TR06: Show Diacritics
The TR06 table is used to compile the mort subtable that implements the Show Diacritics feature setting.
It has the same format as the TR00 table.
TR07: Hide Diacritics
The TR07 table is used to compile the mort subtable that implements the Hide Diacritics feature setting.
It has the same format as the TR00 table.
TR08: Ordinals
The TR08 table is used to compile the mort subtable that implements the Ordinals feature setting. It has
the following format:
| Type |
Name |
Description |
uint16 |
numDigits |
The number of digit glyphs in the table |
uint16 |
numLetters |
The number of (letter, superior) entries in the table |
uint16 |
firstDigit |
The smallest digit glyph ID in the table |
uint16 |
lastDigit |
The largest digit glyph ID in the table |
uint16 |
firstLetter |
The smallest letter glyph ID in the table |
uint16 |
lastLetter |
The largest letter glyph ID in the table |
DigitRec |
digits[numDigits] |
The list of digits that can take ordinals, sorted in increasing
glyph ID order |
LetterRec |
letters[numLetters] |
The list of (letter, superior) entries, sorted in increasing order
of letter glyph ID |
Each DigitRec has the following format:
| Type |
Name |
Description |
uint16 |
digit |
The glyph ID of a digit |
Boolean |
selectionFlag |
A flag indicating whether the corresponding cell in the TrueEdit
edit window is currently selected (i.e., highlighted) |
Each LetterRec has the following format:
| Type |
Name |
Description |
uint16 |
letter |
The glyph ID of a letter |
uint16 |
superior |
The glyph ID of the superior form of the letter |
Boolean |
selectionFlag |
A flag indicating whether the corresponding cell in the TrueEdit
edit window is currently selected (i.e., highlighted) |
TR09: Vertical Fractions
The TR09 table is used to compile the mort subtable that implements the Vertical Fractions feature setting.
It has the same format as the TR00 table.
TR10: Diagonal Fractions
The TR10 table is used to compile the mort subtable that implements the Diagonal Fractions feature setting.
It has the following format:
| Type |
Name |
Description |
| uint16 |
numTriplets |
The number of (digit, numerator, denominator) entries in the table
|
uint16 |
numSlashes |
The number of slash glyphs in the table |
uint16 |
firstSlash |
The smallest slash glyph ID in the table |
uint16 |
lastSlash |
The largest slash glyph ID in the table |
uint16 |
firstDigit |
The smallest digit glyph ID in the table |
uint16 |
lastDigit |
The largest digit glyph ID in the table |
uint16 |
firstLower |
The smallest denominator glyph ID in the table |
uint16 |
lastLower |
The largest denominator glyph ID in the table |
SlashRec |
slashes[numSlashes] |
The list of slashes, sorted in increasing glyph ID order |
TripletRec |
triplets[numTriplets] |
The list of (digit, numerator, denominator) entries, sorted in
increasing order of the digit glyph IDs |
Each SlashRec has the following format:
| Type |
Name |
Description |
uint16 |
glyphId |
The glyph ID of the slash |
Boolean |
selectionFlag |
A flag indicating whether the corresponding cell in the TrueEdit
edit window is currently selected (i.e. highlighted) |
Each TripletRec has the following format:
| Type |
Name |
Description |
uint16 |
digit |
The glyph ID of a digit |
uint16 |
upper |
The glyph ID of the numerator form of that digit |
uint16 |
lower |
The glyph ID of the denominator form of that digit |
Boolean |
selectionFlag |
A flag indicating whether the corresponding cell in the TrueEdit
edit window is currently selected (i.e. highlighted) |
TR11: Contextual Kerning
The TR11 table is used to compile the kern subtable that implements Contextual Kerning. It has the following
format:
| Type |
Name |
Description |
KernOctalRec |
octals[] |
The list of contextual kerning entries. Each entry describes a
sequence of up to eight glyphs, as well as the kerning values
between these glyphs. The number of entries in the table is determined
by dividing the size of the table by the size of an entry |
Each KernOctalRec has the following format:
| Type |
Name |
Description |
unsigned short |
numGlyphs |
The number of glyphs in the contextual kerning string |
unsigned short |
glyphs[8] |
The glyph IDs of the glyphs composing the string |
int16 |
kernVals[7] |
The kerning values that apply between the glyphs in the string
|
Boolean |
selectionFlag |
A flag indicating whether the corresponding row in the TrueEdit
edit window is currently selected (i.e. highlighted) |
TR12: Class-Based Kerning
The TR12 table is used to compile the kern subtable that implements Class-Based Kerning. Its format is not
documented here because this table may be removed in future versions
of TrueEdit.
![[Sticky!]](Sticky.gif)
TR13: Justification
The TR13 table is used to compile the just table, which implements Justification. Its format is not documented
here at this time because:
(1) It is quite complex
(2) It will probably change with future versions of the just editor in TrueEdit.
TR14: Smart Swashes
The TR14 table is used to compile the mort subtables that implement the various Smart Swashes feature settings
(Word Initial Swashes, Word Final Swashes, Line Initial Swashes,
Line Final Swashes, and Non-Final Swashes). It has the following
format:
| Type |
Name |
Description |
Boolean |
subtableType[5] |
Five flags indicating whether each of the corresponding five Smart
Swashes settings currently exists in the mort table |
int32 |
numSwash |
The number of swash entries in the table |
int32 |
numPunc |
The number of punctuation glyphs in the table |
int32 |
numSpec |
The number of special glyphs in the table |
uint16 |
firstGlyph |
The smallest normal, i.e. non- swash, glyph ID in the table |
uint16 |
lastGlyph |
The largest normal, i.e. non- swash, glyph ID in the table |
uint16 |
firstPunc |
The smallest punctuation glyph ID in the table |
uint16 |
lastPunc |
The largest punctuation glyph ID in the table |
uint16 |
firstSpec |
The smallest special glyph ID in the table |
uint16 |
lastSpec |
The largest special glyph ID in the table |
uint16 |
punc[numPunc] |
The glyph IDs of the punctuation glyphs, sorted in increasing
glyph ID order |
uint16 |
spec[numSpec] |
The glyph IDs of the special glyphs, sorted in increasing glyph
ID order |
MortSwashRec |
swashRecs[numSwash] |
The swash entries, sorted in increasing glyph ID order of the
normal, i.e. non-swash, form of the glyph in each swash entry
|
Each MortSwashRec record has the following format:
| Type |
Name |
Description |
long |
glyph |
The glyph ID of the normal, i.e. non-swash, form of the glyph
|
long |
wordInitial |
The glyph ID of the Word Initial swash form of the glyph |
long |
wordFinal |
The glyph ID of the Word Final swash form of the glyph |
long |
lineInitial |
The glyph ID of the Line Initial swash form of the glyph |
long |
lineFinal |
The glyph ID of the Line Final swash form of the glyph |
long |
nonFinal |
The glyph ID of the Non Final swash form of the glyph |
TR15: Diphthong Ligatures
The TR15 table is used to compile the mort subtable that implements the Diphthong Ligatures feature setting.
It has the same format as the TR00 table.
TR16: Prevent Overlap
The TR16 table is used to compile the mort subtable that implements the Prevent Overlap feature setting.
Information on its format will be added to a later version of
this document.
TR17: Hyphens to Em Dash
The TR17 table is used to compile the mort subtable that implements the Hyphens to Em Dash feature setting.
It has the following format:
| Type |
Name |
Description |
long |
hyphen |
The glyph ID of the hyphen |
long |
emDash |
The glyph ID of the em dash |
TR18: Hyphen to En Dash
The TR18 table is used to compile the mort subtable that implements the Hyphen to En Dash feature setting.
It has the following format:
| Type |
Name |
Description |
long |
hyphen |
The glyph ID of the hyphen |
long |
enDash |
The glyph ID of the en dash |
int16 |
numSpaces |
The number of space glyphs in the table |
int16 |
numDigits |
The number of digit glyphs in the table |
uint16 |
firstSpace |
The smallest space glyph ID in the table |
uint16 |
lastSpace |
The largest space glyph ID in the table |
uint16 |
firstDigit |
The smallest digit glyph ID in the table |
uint16 |
lastDigit |
The largest digit glyph ID in the table |
uint16 |
spaces[numSpaces] |
The glyph IDs of the space glyphs, sorted in increasing glyph
ID order |
uint16 |
digits[numDigits] |
The glyph IDs of the digit glyphs, sorted in increasing glyph
ID order |
TR19: Asterisk to Multiply
The TR19 table is used to compile the mort subtable that implements the Asterisk to Multiply feature setting.
It has the following format:
| Type |
Name |
Description |
long |
star |
The glyph ID of the asterisk |
long |
multSign |
The glyph ID of the multiplication sign |
int16 |
numSpaces |
The number of space glyphs in the table |
uint16 |
firstSpace |
The smallest space glyph ID in the table |
uint16 |
lastSpace |
The largest space glyph ID in the table |
uint16 |
spaces[numSpaces] |
The glyph IDs of the space glyphs, sorted in increasing glyph
ID order |
TR20: Exponent
The TR20 table is used to compile the mort subtable that implements the Exponent feature setting. It has
the following format:
| Type |
Name |
Description |
long |
caret |
The glyph ID of the caret |
int16 |
numDigits |
The number of digit glyphs in the table |
uint16 |
firstDigit |
The smallest digit glyph ID in the table |
uint16 |
lastDigit |
The largest digit glyph ID in the table |
MortExpDigitRec |
expDigitRecs[numDigits] |
The list of (digit, exponent) entries in the table, sorted in
increasing glyph ID order of the digits |
Each MortExpDigitRec has the following format:
| Type |
Name |
Description |
long |
digit |
The glyph ID of the digit |
long |
superiorForm |
The glyph ID of the corresponding exponent |
TR21: Hyphen to Minus
The TR21 table is used to compile the mort subtable that implements the Hyphen to Minus feature setting.
It has the following format:
| Type |
Name |
Description |
long |
hyphen |
The glyph ID of the hyphen |
long |
minusSign |
The glyph ID of the minus sign |
int16 |
numSpaces |
The number of space glyphs in the table |
uint16 |
firstSpace |
The smallest space glyph ID in the table |
uint16 |
lastSpace |
The largest space glyph ID in the table |
uint16 |
spaces[numSpaces] |
The glyph IDs of the space glyphs, sorted in increasing glyph
ID order |
TR22: Slash to Divide
The TR22 table is used to compile the mort subtable that implements the Slash to Divide feature setting.
It has the following format:
| Type |
Name |
Description |
long |
slash |
The glyph ID of the slash |
long |
divSymbol |
The glyph ID of the divide sign |
int16 |
numSpaces |
The number of space glyphs in the table |
uint16 |
firstSpace |
The smallest space glyph ID in the table |
uint16 |
lastSpace |
The largest space glyph ID in the table |
uint16 |
spaces[numSpaces] |
The glyph IDs of the space glyphs, sorted in increasing glyph
ID order |
TR23: Inequality Ligatures
The TR23 table is used to compile the mort subtable that implements the Inequality Ligatures feature setting.
It has the following format:
| Type |
Name |
Description |
long |
greaterThan |
The glyph ID of the "greater than" sign |
long |
lessThan |
The glyph ID of the "less than" sign |
long |
equal |
The glyph ID of the equal sign |
long |
greaterThanOrEqualTo |
The glyph ID of the "greater than or equal to" sign |
long |
lessThanOrEqualTo |
The glyph ID of the "less than or equal to" sign |
TR24: Initial Caps
The TR24 table is used to compile the mort subtable that implements the Initial Caps feature setting. It
has the following format:
| Type |
Name |
Description |
int16 |
numPairs |
The number of (lower case, upper case) glyph pairs in the table
|
int16 |
numSpaces |
The number of space glyphs in the table |
uint16 |
firstLower |
The smallest lower case glyph ID in the table |
uint16 |
lastLower |
The largest lower case glyph ID in the table |
uint16 |
firstSpace |
The smallest space glyph ID in the table |
uint16 |
lastSpace |
The largest space glyph ID in the table |
MortUpperPairRec |
pairRecs[numPairs] |
The list of (lower case, upper case) glyph pairs, sorted in increasing
order of the lower case glyph ids |
uint16 |
spaces[numSpaces] |
The glyph IDs of the spaces, in increasing glyph ID order |
Each MortUpperPairRec has the following format:
| Type |
Name |
Description |
long |
lower |
The glyph ID of a lower case glyph |
long |
upper |
The glyph ID of the corresponding upper case glyph |
TR25
Information on this table will be added to a later version of
this document.
TR26: Smart Quotes
Information on this topic will be added to a later version of
this document.
TrueEdit and the mort table
The format of the mort table is specified in QuickDraw GX Font Formats. In addition, TrueEdit makes some assumptions about how the mort table is laid out. If these assumptions are not valid for a particular
font, then TrueEdit will not be able to open the mort table properly.
TrueEdit at present does not check the layout of the mort table
first; it just goes ahead and tries to open it, and typically
hangs or bombs if the layout of the table is not as expected.
The following are the assumptions made by TrueEdit regarding the
overall layout of the mort table:
- The
mort table consists of a single chain, i.e. contains at most 32 feature
settings.
- The order of the feature settings in the Feature List following
the Chain Header is the order in which they will be displayed
by TrueEdit when it opens the
mort table main window.
- The order of the subtables in the
mort table is the same as the order, in the Feature List, of the feature settings which
they implement. If more than one subtable is needed to implement
a particular feature setting, all those subtables will be contiguous,
and, as a group, they will be in Feature List order.
Currently, the only feature setting that requires more than one
subtable to implement is Diagonal Fractions. It requires two subtables:
one to scan the run in the forward direction, and one to scan
the run in the reverse direction.
TrueEdit makes no assumptions about how many subtables are used
to implement each feature setting; it can deduce this information
from the Subfeature Flags of the subtables, and the Enable Flags
of the Feature List entries (more on this below).
- The Enable Flags of each "Active" feature setting (i.e. each feature
setting not of the "Disable" variety; more on this below) consist
of a single "on" bit and 31 "off" bits. The bit which is "on"
will be different for each feature setting. The Subfeature Flags
for each subtable that participates in implementing a particular
feature setting are equal to the Enable Flags for that feature
setting.
- The Feature List consists of the "Active" feature settings, followed
by the "Disable" feature settings, and ends with an "All Typographic
Features: Off" feature setting. The "Active" feature settings
are those for which subtables exist to implement them. The "Disable"
feature settings consist of the following:
- The "None" setting for each radio button feature type that has
such a setting (see attached Feature Registry summary), and that
has at least one setting in the "Active" part of the Feature List.
- The "Off" setting corresponding to each "On" check-box setting
in the "Active" part of the Feature List.
- The Enable Flags and Disable Flags for each feature setting in
the Feature List have the following values:
- "Active" Radio Button feature setting:
- Enable Flags = a single, unique "On" bit
- Disable Flags = such that they disable all other settings of that
feature type in the Feature List (including the "Disable" setting,
if any)
- "Disable" Radio Button feature setting:
- Enable Flags = 0
- Disable Flags = such that they disable all other settings of that
feature type in the Feature List
- "Active" Check Box feature setting:
- Enable Flags = a single, unique "On" bit
- Disable Flags = 0
- "Disable" Check Box feature setting:
- Enable Flags = 0
- Disable Flags = complement of the Enable Flags for the corresponding
"On" setting
- "All Typographic Features: Off" feature setting:
- Enable Flags = 0
- Disable Flags = 0
The CHMP resource
The "Build & Add" option available in TrueEdit for adding a cmap subtable is useful because it allows the user to add a complete
subtable, without having to specify any of the new subtable's
entries. The entries in the new subtable are determined by the
platform, script, and language IDs specified by the user. TrueEdit
determines the format (0, 4, or 6) of the new subtable based on
its contents.
The new subtable is built by using a CHMP resource (from TrueEdit's resource file) in conjunction with
an existing cmap subtable in the font. Details of the CHMP resource follow.
How it works
The CHMP resource maps character codes in a source (src) encoding to those
in a destination (dest) encoding. For example, one CHMP might map Mac-Roman-English character codes to their Unicode-Roman-English
equivalents, another might map Mac-Roman-Dutch codes to Microsoft-Roman-Dutch
codes, etc. There may be one CHMP resource for each mapping from one set of encoding IDs to another.
If CHMP resources with duplicate encoding IDs exist, the first applicable
CHMP will be used.
The following diagram shows how an existing cmap subtable in an sfnt and a CHMP which shares a similar set of encoding IDs (differing only in
platform) can be used to generate a new cmap subtable. The new cmap subtable maps each destination character code in the CHMP to the glyph index which is mapped to in the existing cmap subtable by the corresponding source (Mac) character code in
the CHMP.
In the Add cmap Subtable dialog, TrueEdit allows the user to specify the platform,
script, and language IDs of the cmap subtable to be added. The Build & Add button in this dialog will
be highlighted, enabling a cmap subtable to be built from a CHMP resource and an existing cmap subtable, if the following conditions hold:
- The encoding IDs of the subtable to be added do not duplicate
those of an existing subtable.
- There is an existing subtable in the
sfnt with the same script and language IDs as the subtable to be added
but with "Macintosh" as a platform.
- There is a
CHMP resource in TrueEdit's resource file which meets the following
requirements:
- source platform ID = "Macintosh"
- destination platform ID = platform ID of subtable to be added
- source/destination script IDs = script ID of subtable to be added
- source/destination language IDs = language ID of subtable to be
added
Format
ResEdit should be used to create or modify CHMP resources.
If a CHMP is being created outside of the TrueEdit resource file, create
a new resource file or open an existing one in ResEdit and choose
"Create New Resource" from the Resource menu. Enter CHMP in the text field, and a new window (a hex window)
will appear.
If a resource file which already contains CHMPs is being used, double-click CHMP in the list and choose "Create New Resource".
In either case, the new CHMP can be built once a hex window has been created for it in ResEdit.
The following is a representation of the format of the CHMP resource.
Note that:
- All values in a
CHMP resource are in hexadecimal.
- On the Macintosh, the source platform should always be "Macintosh".
- All
CHMP character codes are two bytes in length.
- If there is a source character code which does not map to a destination
character code, it can either be excluded from the
CHMP or mapped to character code 0xFFFF.
Below is a screen snapshot taken in ResEdit. It shows an actual
CHMP with its components labeled. This particular CHMP maps 256 Macintosh-Roman-English character codes to the equivalent
256 Unicode-Roman-English codes. Notice the block of FFFFs, which denote those codes which exist in the source encoding
(Mac-Roman-English) but not in the destination encoding (Unicode-Roman-English).
It is simple, though a bit tedious, to create CHMPs. The benefit of spending the time creating a CHMP is that TrueEdit users will not have to spend time on painful
and repetitive entry of character codes and their associated glyph
indices; this process is automated by the presence of CHMP resources.
For more information
Consult the documents TrueType Font Format Specification and QuickDraw GX Font Formats for more information on cmaps. The latter document offers an in-depth explanation of the different
cmap formats, including format 2, which might be useful in working
with cmaps.
The Apple Fonts Group (applefonts@apple.com)
|
