The TrueType post table

The post table holds what a PostScript printer needs: the italic angle, the underline, whether the font is monospaced, memory estimates, and, in version 2.0, a PostScript name for every glyph. Windows reads the italic angle but needs no glyph names. Fonts with TrueType outlines may use any version; fonts with CFF outlines must use 3.0, as their glyph names are in the CFF data.

Header

32 bytes, the same in every version.

Offset	Type	Name	Description
`0`	`Version16Dot16`	`version`	`0x00010000` 1.0, `0x00020000` 2.0, `0x00025000` 2.5 (deprecated), `0x00030000` 3.0; Apple also defines `0x00040000` 4.0, which OpenType does not support.
`4`	`Fixed`	`italicAngle`	Degrees counter-clockwise from vertical: 0 for upright, negative for text leaning right. −12° is `0xFFF40000`.
`8`	`FWORD`	`underlinePosition`	The suggested y of the top of the underline. (PostScript’s `UnderlinePosition` key means its centre; `post` does not use that meaning.)
`10`	`FWORD`	`underlineThickness`	The suggested underline thickness, normally that of the underscore and of the strikeout in `OS/2`.
`12`	`uint32`	`isFixedPitch`	0 if proportional, any other value if monospaced.
`16`	`uint32`	`minMemType42`	Minimum printer memory used when the font is downloaded as a Type 42 font; 0 if unknown.
`20`	`uint32`	`maxMemType42`	Maximum memory, as a Type 42 font: the minimum plus the most the rasteriser may need at run time.
`24`	`uint32`	`minMemType1`	Minimum memory when downloaded as a Type 1 font.
`28`	`uint32`	`maxMemType1`	Maximum memory when downloaded as a Type 1 font.

In a variable font, the underline can vary through MVAR with the tags undo (position) and unds (thickness); the italic angle comes from the slnt axis.

Versions

1.0: the standard Macintosh glyph set

Nothing follows the header. The font has exactly the 258 glyphs of the standard Macintosh set, in the standard order, and takes their names from that list.

2.0: names for every glyph

Type	Name	Description
`uint16`	`numGlyphs`	The number of glyphs; should equal `numGlyphs` in `maxp`.
`uint16` × `numGlyphs`	`glyphNameIndex`	For each glyph, 0 to 257 for a standard Macintosh name, or 258 and up for string number (index − 258) in `stringData`.
`uint8`[ ]	`stringData`	Pascal strings to the end of the table: a length byte, then that many ASCII characters.

A name may use only A–Z, a–z, 0–9, full stop and underscore, and be at most 63 characters; some older software assumes 31. A glyph that should have no name gets index 0, .notdef. To find string n, a reader skips the n strings before it one by one. Names are normally those of the Adobe Glyph List, or uni followed by four hex digits for other characters.

2.5: reordered standard set (deprecated)

Type	Name	Description
`uint16`	`numGlyphs`	The number of glyphs.
`int8` × `numGlyphs`	`offset`	For each glyph, its standard Macintosh index minus its glyph ID.

For fonts whose glyphs are all in the standard set but some are missing or out of order. A font of just A, B and C as glyphs 0, 1 and 2 would store +36, +36, +36, since they are standard glyphs 36, 37 and 38.

3.0: no glyph names

Nothing follows the header, and the font gives no glyph names. A PostScript driver may print nothing or make up names, but must not fail. This keeps fonts with many glyphs, such as CJK fonts, small.

4.0: Apple’s character codes

Defined only by Apple, for fonts that map to composite fonts on Chinese, Japanese or Korean printers. The header is followed by one uint16 per glyph: the character code for that glyph, or 0xFFFF for none. The driver names each glyph a followed by the code in hex. Apple says such tables are no longer needed and should be avoided.

The 258 standard Macintosh glyph names

Used by version 1.0, and by version 2.0 for indices 0 to 257. Glyph 1, .null, and glyph 2, nonmarkingreturn, are the empty glyphs old Mac fonts carried for the null and carriage return characters.

Index	Names
0 to 9	`.notdef` `.null` `nonmarkingreturn` `space` `exclam` `quotedbl` `numbersign` `dollar` `percent` `ampersand`
10 to 19	`quotesingle` `parenleft` `parenright` `asterisk` `plus` `comma` `hyphen` `period` `slash` `zero`
20 to 29	`one` `two` `three` `four` `five` `six` `seven` `eight` `nine` `colon`
30 to 39	`semicolon` `less` `equal` `greater` `question` `at` `A` `B` `C` `D`
40 to 49	`E` `F` `G` `H` `I` `J` `K` `L` `M` `N`
50 to 59	`O` `P` `Q` `R` `S` `T` `U` `V` `W` `X`
60 to 69	`Y` `Z` `bracketleft` `backslash` `bracketright` `asciicircum` `underscore` `grave` `a` `b`
70 to 79	`c` `d` `e` `f` `g` `h` `i` `j` `k` `l`
80 to 89	`m` `n` `o` `p` `q` `r` `s` `t` `u` `v`
90 to 99	`w` `x` `y` `z` `braceleft` `bar` `braceright` `asciitilde` `Adieresis` `Aring`
100 to 109	`Ccedilla` `Eacute` `Ntilde` `Odieresis` `Udieresis` `aacute` `agrave` `acircumflex` `adieresis` `atilde`
110 to 119	`aring` `ccedilla` `eacute` `egrave` `ecircumflex` `edieresis` `iacute` `igrave` `icircumflex` `idieresis`
120 to 129	`ntilde` `oacute` `ograve` `ocircumflex` `odieresis` `otilde` `uacute` `ugrave` `ucircumflex` `udieresis`
130 to 139	`dagger` `degree` `cent` `sterling` `section` `bullet` `paragraph` `germandbls` `registered` `copyright`
140 to 149	`trademark` `acute` `dieresis` `notequal` `AE` `Oslash` `infinity` `plusminus` `lessequal` `greaterequal`
150 to 159	`yen` `mu` `partialdiff` `summation` `product` `pi` `integral` `ordfeminine` `ordmasculine` `Omega`
160 to 169	`ae` `oslash` `questiondown` `exclamdown` `logicalnot` `radical` `florin` `approxequal` `Delta` `guillemotleft`
170 to 179	`guillemotright` `ellipsis` `nonbreakingspace` `Agrave` `Atilde` `Otilde` `OE` `oe` `endash` `emdash`
180 to 189	`quotedblleft` `quotedblright` `quoteleft` `quoteright` `divide` `lozenge` `ydieresis` `Ydieresis` `fraction` `currency`
190 to 199	`guilsinglleft` `guilsinglright` `fi` `fl` `daggerdbl` `periodcentered` `quotesinglbase` `quotedblbase` `perthousand` `Acircumflex`
200 to 209	`Ecircumflex` `Aacute` `Edieresis` `Egrave` `Iacute` `Icircumflex` `Idieresis` `Igrave` `Oacute` `Ocircumflex`
210 to 219	`apple` `Ograve` `Uacute` `Ucircumflex` `Ugrave` `dotlessi` `circumflex` `tilde` `macron` `breve`
220 to 229	`dotaccent` `ring` `cedilla` `hungarumlaut` `ogonek` `caron` `Lslash` `lslash` `Scaron` `scaron`
230 to 239	`Zcaron` `zcaron` `brokenbar` `Eth` `eth` `Yacute` `yacute` `Thorn` `thorn` `minus`
240 to 249	`multiply` `onesuperior` `twosuperior` `threesuperior` `onehalf` `onequarter` `threequarters` `franc` `Gbreve` `gbreve`
250 to 257	`Idotaccent` `Scedilla` `scedilla` `Cacute` `cacute` `Ccaron` `ccaron` `dcroat`

Example

The 292-byte version 2.0 post table of KaTeX’s Size2 Regular font:

00000000  00 02 00 00 00 00 00 00  FF 83 00 32 00 00 00 00
00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00
00000020  00 28 00 00 00 01 00 02  00 03 00 0B 00 0C 00 12
00000030  00 3E 00 3F 00 40 00 5E  00 60 01 02 00 D8 00 D9
00000040  01 03 01 04 00 9A 01 05  00 99 00 A5 00 9C 01 06
…
00000070  01 17 07 75 6E 69 30 30  41 30 07 75 6E 69 30 33

Version 2.0, upright (italic angle 0), underline top at −125 and 50 thick, proportional, no memory estimates.
40 glyph name indices. Glyphs 0 to 3 are standard names 0 to 3, .notdef to space; glyph 4 is 11, parenleft; glyph 17 is 154, product, and glyph 19 is 153, summation, the glyphs the cmap gives for ∏ and ∑.
Glyph 12 has index 258 (01 02): the first string, 07 then uni00A0, the no-break space. Indices 258 to 279 name 22 glyphs this way, such as tildecomb and uni2210.

.TTF File Format – post