gxvalid: TrueType GX validator ============================== 1. What is this --------------- "gxvalid" is a module to validate TrueType GX tables: a collection of additional tables in TrueType font which is used by "QuickDraw GX Text", Apple Advanced Typography (AAT). In addition, gxvalid can validates "kern" table which had been extended for AAT. Like otvalid, gxvalid uses Freetype2's validator framework(ftvalid). You can link gxvalid with your program; before running your own layout engine, gxvalid validates a font file. As the result, you can reduce error-checking code from the layout engine. You can use gxvalid as a stand-alone font validator; ftvalid command included in ft2demo calls gxvalid internally. Stand-alone font validator may be useful for font developers. This documents contains following informations: - supported TrueType GX tables - validation limitation in principle - permissive error handling of broken GX tables - "kern" table issue. 2. Supported tables ------------------- Following GX tables are currently supported. bsln feat just kern(*) lcar mort morx opbd prop trak Following GX tables are currently unsupported. cvar fdsc fmtx fvar gvar Zapf Following GX tables won't be supported. acnt(**) hsty(***) Undocumented tables in TrueType fonts designed for Apple platform. CVTM TPNM addg umif *) "kern" validator includes both of classic kern (format supported by both of Microsoft and Apple platforms) and new kern (a format supported by Apple platform only). **) "acnt" tables is not supported by currently available Apple font tools. ***) There is one more Apple extension "hsty" but it is for Newton-OS, not GX (Newton-OS is a platform by Apple, but it can use sfnt- housed bitmap fonts only. Therefore, it should be excluded from "Apple platform" in the context of TrueType. gxvalid ignores it as Apple font tools do so. We have checked 183 fonts bundled to MacOS 9.1, MacOS 9.2, MacOS 10.0, MacOS X 10.1, MSIE for MacOS and AppleWorks 6.0. In addition, we have checked 67 Dynalab fonts (designed for MacOS) and 189 Ricoh fonts (designed for Windows and MacOS dual platforms). The number of fonts including TrueType GX tables are listed in following: bsln: 76 feat: 191 just: 84 kern: 59 lcar: 4 mort: 326 morx: 19 opbd: 4 prop: 114 trak: 16 Dynalab and Ricoh fonts didn't have GX tables except of feat and mort. 3. Validation limitations in principle -------------------------------------- TrueType GX provides layout information to font-rasterize/text-layout libraries. gxvalid can check whether layout information is stored as TrueType GX format specified by Apple. But gxvalid cannot check how QuickDraw GX/AAT renderer uses the stored information. 3-1. Validation of State Machine activity ----------------------------------------- QuickDraw GX/AAT has "State Machine" to provide "stateful" layout features, and TrueType GX stores the state transition diagram of "State Machine" in "StateTable" data structure. While State Machine receives a series of glyph ID, State Machine starts from "start of text" state, walks around various states and generates various layout informations to renderer, and finally reaches to "end of text". gxvalid can check essential errors like: - possibility of state transition to undefined states - existence of glyph ID that State Machine doesn't know how to handle it - State Machine cannot compute the layout information from given diagram these errors can be checked within finite steps, and without State Machine itself, because these are errors of "expression" of state transition diagram. There's no limitation about how long State Machine walks around, so validation of the algorithm in the state transition diagram requires infinite steps, even if we have State Machine in gxvalid. Therefore, following "errors" cannot be checked. - existence of states which State Machine never transits to. - possibility that State Machine never reaches to "end of text". - possibility of stack underflow/overflow in State Machine (in ligature and contextual glyph substitution, State Machine can store 16 glyphs onto its stack) In addition, gxvalid doesn't check "temporal glyph ID" used in the chained State Machines (in "mort" and "morx" tables). When a layout feature is implemented by single State Machine, glyph ID converted by State Machine is passed to the glyph renderer, thus it should not point to undefined glyph ID. But if a layout feature is implemented by chained State Machines, the component State Machine (if it is not final one) is permitted to generate undefined glyph ID for temporal use, because it is handled by next component State Machine, instead of the glyph renderer. To validate such temporal glyph ID, gxvalid must stack all undefined glyph IDs which is possible in the output of previous State Machine and search them in "ClassTable" of current State Machine. It is too complexed work to list all possible glyph IDs from StateTable, especially from ligature substitution table. 3-2. Validation of relationship among multiple layout features -------------------------------------------------------------- gxvalid does not validate the relationship among multiple layout features at all. If multiple layout features are defined in TrueType GX tables, the interactivity, overriding, and conflict among layout features are defined in the font too. For example, there are several predefined spacing control features: - Text Spacing (Proportional/Monospace/Half-width/Normal) - Number Spacing (Monospaced-numbers/Proportional-numbers) - Kana Spacing (Full-width/Proportional) - Ideographic Spacing (Full-width/Proportional) - CJK Roman Spacing (Half-width/Proportional/Default-roman /Full-width-roman/Proportional) If all layout features are independently managed, we can set an inconsistent typographic rule, as like "Text Spacing=Monospace" and "Ideographic Spacing=Proportional", at the same time. The combination of each layout feature is managed by 32bit integer (1 bit for 1 selector setting), so we can define relationship among features up to 32 settings, theoretically. But if setting of a feature affects setting of another features, typographic priority of each layout feature is required to validate the relationship. TrueType GX format specification does not give such information even for predefined features. 4. Permissive error handling of broken GX tables ------------------------------------------------ When Apple's font rendering system finds an inconsistency, violation of specification or unspecified value in TrueType GX tables, they do not always return error. In most case, they silently ignore such wrong values or whole of table. In fact, MacOS is shipped with fonts including broken GX/AAT tables, but no harmful effects due to officially broken fonts are observed by end-users. gxvalid is designed to continue its validation as long as possible. When gxvalid find wrong value, gxvalid warns it at least, and take a fallback procedure if possible. The fallback procedure depends on the debug level. We used following 3 tools to refer Apple's error handling. - FontValidator (for MacOS 8.5 - 9.2) resource fork font - ftxvalidator (for MacOS X 10.1 -) dfont or naked-sfnt - ftxdumperfuser (for MacOS X 10.1 -) dfont or naked-sfnt However, all tests are on PowerPC based Macintosh, we have not tested on m68k-based Macintosh at all, at present. We checked 183 fonts bundled to MacOS 9.1, MacOS 9.2, MacOS 10.0, MacOS X 10.1, MSIE for MacOS and AppleWorks 6.0. These fonts are distributed officially, but many broken GX/AAT tables are found by Apple's font tools. In following, we list typical violation against GX specification, in Apple official fonts. At least, gxvalid warns them, and fallback method to continue 4-1. broken BinSrchHeader ( 19/183) ----------------------------------- BinSrchHeader is a header of data array, for m68k platform to access memory effectively. Although independent parameters for real use are only 2 (unitSize and nUnits), BinSrchHeader has 3 additional parameters which can be calculated from unitSize and nUnits, for fast setup. Apple font tools ignore them silently, so gxvalid warns inconsistency and always continues validation. The additional parameters are ignored regardless of the consistency. 19 fonts include inconsistent with calculated values all breaks are in BinSrchHeader of "kern" table. 4-2. too-short LookupTable ( 5/183) ----------------------------------- LookupTable format 0 is simple array to get a value from given GID, the index of array is GID. Therefore, the length of array is expected to be same with max GID defined in "maxp" table, but there is some fonts whose LookupTable format 0 is too short to cover all GID. FontValidator ignores this error silently, ftxvalidator and ftxdumperfuser warns and continues. Similar shortage is found in format 3 subtable of "kern". gxvalid warns always and abort at FT_VALIDATE_PARANOID. 5 fonts include too-short kern format 0 subtables. 1 font includes too-short kern format 3 subtable. 4-3. broken LookupTable format 2 ( 1/183) ----------------------------------------- LookupTable format 2, 4 covers GID space by collection of segments which specified by firstGlyph and lastGlyph. Some fonts stores firstGlyph and lastGlyph in reverse order, so segment specification is broken. Apple font tools ignores this error silently, broken segment is ignored as if it did not exist. gxvalid warns and normalize the segment at FT_VALIDATE_DEFAULT, or ignore the segment at FT_VALIDATE_TIGHT, or abort at FT_VALIDATE_PARANOID. 1 font includes broken LookupTable format 2, in "just" table. *) It seems that all fonts manufactured by ITC for AppleWorks have this error. 4-4. bad bracketing in glyph property ( 14/183) ----------------------------------------------- GX/AAT defines bracketing property of the glyphs by "prop" table, to control layout functionalities for string closed in brackets and out of brackets. Some fonts give inappropriate bracket properties to glyphs. Apple font tools warn this error. gxvalid warns always and abort at FT_VALIDATE_PARANOID. 14 fonts include wrong bracket properties. 4-5. invalid feature number (117/183) ------------------------------------- GX/AAT extension can include 255 different features for layout, but popular layout features are predefined (see http://developer.apple.com/fonts/Registry/index.html). Some fonts include feature number which is incompatible with predefined feature registry. In our survey, there are 140 fonts including "feat" table. a) 67 fonts uses feature number which should not be used. b) 117 fonts set wrong feature range (nSetting). this infraction is found in mort/morx. Apple font tools gives no warning, although they cannot recognize what the feature is. At FT_VALIDATE_DEFAULT, gxvalid warns but continues in both cases (a, b). At FT_VALIDATE_TIGHT, gxvalid warns and aborts for (a), but continues for (b). At FT_VALIDATE_PARANOID, gxvalid warns and aborts in both cases (a, b). 4-6. invalid prop version ( 10/183) ----------------------------------- As most TrueType GX tables, prop table must start with 32bit version: 0x00010000, 0x00020000 or 0x00030000. But some fonts store nonsense binary data in it. When Apple font tools find them, they abort the processing at once, and following data are unhandled. gxvalid does same always. 10 fonts include broken prop version. All of these fonts are classic TrueType for Japanese script, manufactured by Apple. 4-7. unknown resource name ( 2/183) ------------------------------------ NOTE: THIS IS NOT TRUETYPE GX ERROR When TrueType font is stored in resource fork or dfont format, the data must be tagged as "sfnt" in resource fork index, to invoke TrueType font handler for the data. But the TrueType font data in "Keyboard.dfont" is tagged as "kbd", and that in "LastResort.dfont" is tagged as "lst". Apple font tools can detect the data is of TrueType and successfully validate them. Possibly this because they are known to be dfont. Current implementation of resource fork driver of FreeType cannot do that, thus gxvalid cannot validate them. 2 fonts use unknown tag for TrueType font resource. 5. "kern" table issue --------------------- In common terminology of TrueType, "kern" is classified to basic and platform-independent table. But there are Apple extensions of kern, and there is an extension which requires GX state machine for contextual kerning. Therefore, gxvalid includes validator for kern. Unfortunately, there is no exact algorithm to check Apple's extension, so gxvalid includes pragmatic detector of data format and validator for all possible data formats, including data format for Microsoft. By calling classic_kern_validate() instead of gxv_validate(), you can specify available "kern" format explicitly. However, current FreeType2 uses Microsoft "kern" format only, others are ignored. 5-1. History ------------ Original 16bit version of "kern" had been designed by Apple in pre- GX era, and it was also approved by Microsoft. Afterwards, Apple has designed new 32bit version "kern". Apple has noted as the difference between 16bit and 32bit version is only the size of variables in "kern" header. In following, we call the original 16bit version as "classic", and 32bit version as "new". 5-2. Versions and dialects which should be discriminated -------------------------------------------------------- The "kern" table consists of the table header and several subtables. The version "classic" or "new" is explicitly written in the table header, but there are undocumented difference of font parser between Microsoft and Apple. It is called as "dialect" in following. There are 3 cases which should be discriminated: new Apple-dialect, classic Apple-dialect, and classic Microsoft-dialect. Analysis and auto detection algorithm of gxvalid is described in following. 5-2-1. Version detection: classic and new kern ---------------------------------------------- According to Apple TrueType specification, the clarified difference between classic and new version are only 2: - "kern" table header starts with the version number. The classic version starts with 0x0000 (16bit), the new version starts with 0x00010000 (32bit). - In the "kern" table header, the number of subtables follows to the version number. In the classic version, it is stored in 16bit variable. In the new version, it is stored in 32bit variable. From Apple font tool's output (DumpKERN is also tested in addition to 3 Apple font tools in above), there is another undocumented difference. In new version, the subtable header includes a 16bit variable named "tupleIndex" which does not exist in the classic version. New version can store all subtable formats (0, 1, 2 and 3), but Apple TrueType specification does not mention about subtable formats available in classic version. 5-2-2. Avaibale subtable format in classic version -------------------------------------------------- Although Apple TrueType specification recommends to use classic version in the case if the font is designed for both of Apple and Microsoft platforms, it does not note about the available subtable formats in classic version. According to Microsoft TrueType specification, the subtable format assured for Windows & OS/2 support is only subtable format 0. Also Microsoft TrueType specification describes the subtable format 2, but does not mention about which platforms support it. About subtable format 1, 3 and later are noted as reserved for future use. Therefore, the classic version can store subtable formats 0 and 2, at least. ttfdump.exe, a font tool provided by Microsoft ignores the subtable format written in the subtable header, and parse as if all subtables are in format 0. kern subtable format 1 uses StateTable, so it cannot be utilized without GX State Machine. Therefore, it is reasonable to assume format 1 (and 3) is introduced after Apple have introduced GX and moved to new 32bit version. 5-2-3. Apple and Microsoft dialects ----------------------------------- The kern subtable has 16bit "coverage" to describe kerning attributions, but bit-interpretations by Apple and Microsoft are reverse ordered: e.g. Apple-dialect writes subtable format from 0x000F bit range, Microsoft-dialect writes subtable format from 0x0F00 bit range). In addition, from the outputs of DumpKERN and FontValidator, Apple's bit-interpretations of coverage in classic and new version are incompatible. In summary, there are 3 dialects: classic Apple- dialect, classic Microsoft-dialect, and new Apple-dialect. The classic Microsoft-dialect and new Apple-dialect are documented by each vendors' TrueType font specification, but the document for classic Apple-dialect had been lost. For example, in new Apple-dialect, the bit 0x8000 is documented as "set to 1 when the kerning is vertical". On the other hand, in classic Microsoft-dialect, the bit 0x0001 is documented as "set to 1 when the kerning is horizontal". From the outputs of DumpKERN and FontValidator, classic Apple-dialect recognizes the bit 0x8000 as "set to 1 when the kerning is horizontal". From the results of similar experiments, classic Apple-dialect is ein ndian-reverse of classic Microsoft-dialect. It must be noted: no font tool can sense classic Apple-dialect or classic-Microsoft dialect automatically. 5-2-4. gxvalid auto dialect detection algorithm ----------------------------------------------- The first 16bit of kern table is enough to sense the version: - if first 16bit is 0x0000, kern table is in classic Apple-dialect or classic Microsoft-dialect, - if first 16bit is 0x0001, and next 16bit is 0x0000, kern table is in new Apple-dialect. If kern table is classic version, 16bit coverage is checked for in next. For first, the coverage is decoded by classic Apple-dialect as following (it is based on DumpKERN output): 0x8000: 1=horizontal, 0=vertical 0x4000: not used 0x2000: 1=cross-stream, 0=normal 0x1FF0: reserved 0x000F: subtable format If any of reserved bits are set or subtable format is interpreted as 1 or 3, we take it as "impossible in classic Apple-dialect", and retry by classic Microsoft-dialect. The most popular coverage in new Apple-dialect: 0x8000, The most popular coverage in classic Apple-dialect: 0x0000, The most popular coverage in classic Microsoft dialect: 0x0001. 5-3. Tested fonts ----------------- We checked 59 fonts bundled to MacOS which includes kern, and 38 fonts bundled to Windows which includes kern. - fonts bundled to MacOS * new Apple-dialect format 0: 18 format 2: 1 format 3: 1 * classic Apple-dialect format 0: 14 * classic Microsoft-dialect format 0: 15 - fonts bundled to Windows * classic Microsoft-dialect format 0: 38 It looks strange that classic Microsoft-dialect fonts are bundled to MacOS: they come from MSIE for MacOS, except of MarkerFelt.dfont. ACKNOWLEDGEMENT --------------- Some part of gxvalid is derived from both gxlayout module and otvalid module. Development of gxlayout was support of Information-technology Promotion Agency(IPA), Japan. The detailed analysis of undefined glyph ID utilization in mort, morx is provided by George Williams. ------------------------------------------------------------------------ Copyright 2004, 2005 by suzuki toshiya, Masatake YAMATO, Red hat K.K., David Turner, Robert Wilhelm, and Werner Lemberg. This file is part of the FreeType project, and may only be used, modified, and distributed under the terms of the FreeType project license, LICENSE.TXT. By continuing to use, modify, or distribute this file you indicate that you have read the license and understand and accept it fully. --- end of README ---