Skip to content

README.md

Latest commit

 

History

History
74 lines (51 loc) · 2.89 KB

README.md

File metadata and controls

74 lines (51 loc) · 2.89 KB

UnicodeFontIndex - Speed up partial loading of Unicode fonts

Description

NOTE: Unlike other modifications, this modification is not standalone. It requires that you applied the UTF8Charset or UTF8CharsetLite modification, before this modification.

As Unicode fonts can grow quite big and often only very few glyph ranges are needed, the Fonts.Mod from the UTF8Charset modification loads glyphs in ranges of 64 character. However, on every load, the whole runs table and a large chunk of the boxes table need to be parsed to find the required offsets into the glyphs table. This modification introduces a Unicode Font Index file *.Scn.uFi which contains precalculated offsets for each range present in the file, indexed by a two-way indirect lookup table similar to the one used in memory when loading the font. UnicodeFontIndex.Build can be used to build such index files for Unicode fonts.

Index File Format

The file consists of a 20 byte header followed by 512-byte pages. The pages are 1-based, i. e. the start offset for page number i is i * 512 - 492.

The maximum size of a file (for a font that covers all of Unicode) is 16 pages (about 8 KiB). Page numbers in the file can therefore be encoded as a single byte.

The header consists of a magic value 01694675H followed by page numbers for the 16 lookup tables. If a lookup table is not present, the page number is zero.

Each lookup table consists of one page and stores for each of the 64 glyph ranges:

  • The number of runs to skip when loading the range (i.e. non-overlapping runs before the first overlapping run) [2 bytes]
  • The number of boxes to skip when loading the range [2 bytes]
  • Offset of the first glyph covered in this range [4 bytes]

In case a range does not contain any glyphs, the values are all -1.

When building the index, pages are allocated and everything except the glyph offsets are filled while parsing the runs. Then the glyph offsets will be dynamically updated while parsing the boxes.

Installation

  • Apply UnicodeFontIndex.patch to Fonts.Mod (or FontsU.Mod in case of UTF8CharsetLite modification)

  • Recompile the changed modules and their dependencies (including the compiler). In case of Fonts.Mod:

    ORP.Compile Fonts.Mod/s Texts.Mod/s ~
ORP.Compile Oberon.Mod/s ~
ORP.Compile MenuViewers.Mod/s ~
ORP.Compile TextFrames.Mod/s ~
ORP.Compile System.Mod/s ~
ORP.Compile Edit.Mod/s ~
ORP.Compile ORS.Mod/s ORB.Mod/s ~
ORP.Compile ORG.Mod/s ORP.Mod/s ~

    In case of FontsU.Mod:

    ORP.Compile FontsU.Mod/s TextFramesU.Mod/s EditU.Mod/s ~

  • Compile the index builder

    ORP.Compile UnicodeFontIndex.Mod/s ~

  • Build indexes for your fonts

    UnicodeFontIndex.Build Oberon10.Scn.Fnt ~

  • Restart the system.