GENDICT(1)

💖 GENDICT(1) — the ICU 78.3 manual era 💖

okay bestie, sit down, hydrate, because we're about to compile a word list into an ICU string trie dictionary and it's gonna be cute 🎀

---

✨ NAME ✨

gendict — compiles a word list into an ICU string trie dictionary 🦋

---

🌸 SYNOPSIS 🌸

gendict [ --uchars | --bytes --transform transform ] [ -h, -?, --help ]
[ -V, --version ] [ -c, --copyright ] [ -v, --verbose ] [ -i,
--icudatadir directory ]  input-file  output-file

---

🫶 DESCRIPTION 🫶

so gendict literally reads the word list from dictionary-file and creates a string trie dictionary file. no cap. normally that data file has the .dict extension, that's just the vibe 💅

here's the tea on how words work: words begin at the beginning of a line and are terminated by the first whitespace. and lines that begin with whitespace? ignored. not invited to the party. bye 🙈

---

⭐ OPTIONS ⭐

• -h, -?, --help — prints help about usage and dips out. that's it, that's the whole job 💖
• -V, --version — prints the version of gendict and then exits. iconic behaviour honestly 👑
• -c, --copyright — embeds the standard ICU copyright straight into the output-file. legal girlies stay winning 🎀
• -v, --verbose — spills extra informative messages while it's running. she's chatty and we love that for her 🦋
• -i, --icudatadir directory — looks for any necessary ICU data files in directory. for example, the file pnames.icu must be located when ICU's data is not built as a shared library. the default ICU data directory is specified by the environment variable ICU_DATA. most configurations of ICU do not require this argument, so don't stress bestie 🌸
• --uchars — sets the output trie type to UChar. mutually exclusive with --bytes — you can't have both, pick a lane 💀
• --bytes — sets the output trie type to Bytes. mutually exclusive with --uchars. same energy, opposite direction 🔥
• --transform — sets the transform type. should only be specified with --bytes. currently supported transforms are: offset-<hex-number>, which specifies an offset to subtract from all input characters. and listen — it should be noted that the offset transform also maps U+200D to 0xFF and U+200C to 0xFE, in order to offer compatibility to languages that require these characters. a transform must be specified for a bytes trie, and when applied to the non-value characters in the input-file it must produce output between 0x00 and 0xFF. no shortcuts, the math has to math ✨
• input-file — the source file to read. your raw material, your muse 💜
• output-file — the file to write the output dictionary to. the glow-up destination 🌈

---

😭 CAVEATS 😭

okay this is the part where we get serious for a sec, bestie, because these are real:

• the input-file is assumed to be encoded in UTF-8. don't come at her with anything else 🫶
• the integers in the input-file that are used as values must be made up of ASCII digits. they may be specified either in hex, by using a 0x prefix, or in decimal.
• either --bytes or --uchars must be specified. like, mandatory mandatory. no trie type, no dictionary, periodt 💅

---

🌈 ENVIRONMENT 🌈

• ICU_DATA — specifies the directory containing ICU data. defaults to ${prefix}/share/icu/78.3/. ⚠️ some tools in ICU depend on the presence of the trailing slash. it is thus important to make sure that it is present if ICU_DATA is set. like, do not remove that slash, she notices 🎀

fr fr, that trailing slash is not decorative. keep it.

---

👑 AUTHORS 👑

Maxime Serrano

---

💜 VERSION 💜

1.0

---

🎀 COPYRIGHT 🎀

Copyright (C) 2012 International Business Machines Corporation and others

---

🦋 SEE ALSO 🦋

http://www.icu-project.org/userguide/boundaryAnalysis.html

---

and that's the whole page bestie!! this lil' legend lives in the ICU MANPAGE, version 1.0, dated 1 June 2012, page 1 of the GENDICT(1) manual. go compile something gorgeous 💖✨🌸