Source: triehash
Section: devel
Priority: optional
Maintainer: Julian Andres Klode <jak@debian.org>
Rules-Requires-Root: no
Build-Depends: debhelper (>= 11)
Standards-Version: 4.3.0
Homepage: https://github.com/julian-klode/triehash
Vcs-Git: https://github.com/julian-klode/triehash -b debian/master
Vcs-Browser: https://github.com/julian-klode/triehash

Package: triehash
Architecture: all
Depends: ${misc:Depends}, ${perl:Depends}
Multi-Arch: foreign
Description: Generates perfect hash functions as native machine code
 TrieHash generates perfect hash functions as C code which then gets
 compiled into optimal machine code as part of the usual program compilation.
 .
 TrieHash works by translating a list of strings to a trie, and then converting
 the trie to a set of recursive switch statements; first switching by length,
 and then switching by bytes.
 .
 TrieHash has various optimizations such as processing multiple bytes at once
 (on GNU C), and shortcuts for reducing the complexity of case-insensitive
 matching (ASCII only). Generated code performs substantially faster than
 gperf, but is larger.
 .
 TrieHash was written for use in APT.