Install tre with Homebrew, apk, dnf, MacPorts, Nix, pacman, apt, zypper

Project links

• Homepage

Commands and aliases

• agrep

Project history

The official README presents TRE as a POSIX-compliant regexp matcher focused on robustness, predictable performance, and features uncommon in free POSIX implementations. It advertises approximate matching, strict standard conformance, linear worst-case matching time in the searched text, modest memory use, multibyte and wide-character support, binary data support, thread safety, and portability.

TRE's approximate matching uses edit distance, allowing insertions, deletions, and substitutions with configurable costs. The README notes that TRE's included `agrep` differs from older agrep implementations by allowing full regexps of any length, any number of errors, and non-uniform insertion, deletion, and substitution costs.

The NEWS file records a long release line. Earlier releases added approximate matching syntax, pkg-config support, non-greedy repetition, system ABI support, Unicode and wide-character improvements, Python bindings, command-line `agrep` options, Cygwin and MinGW builds, and license changes. Later releases changed the license to a BSD-style license in 0.7.6, introduced a `tre_`-prefixed ABI in 0.8.0, and added raw byte-vector API variants in 0.9.0.

Adoption history

TRE's adoption history is mostly library-driven: C programs and Unix-like systems package it for POSIX regex compatibility plus fuzzy matching. The batch input records package-manager names such as Homebrew `tre`, Debian and Ubuntu `libtre-dev`, MacPorts `tre`, Nix `tre`, and `agrep` packages in apk, dnf, pacman, and zypper.

The split between library package names and `agrep` executable names explains why package catalogs can look inconsistent. Some ecosystems expose the library as TRE, while others emphasize the user-facing approximate grep command.

How it is used

Library users compile and link against TRE for POSIX regex APIs with stronger matching guarantees, approximate matching, multibyte support, and thread-safe execution. The README documents the usual autotools build path through `./utils/autogen.sh`, `./configure`, `make`, `make check`, and `make install`.

Command-line users encounter TRE through `agrep`, which searches approximately rather than requiring exact matches. That makes it useful for typo-tolerant text search, OCR cleanup, fuzzy log investigation, and cases where edit-distance matching is more important than plain regular-expression grep.

Why package nerds care

For package nerds, TRE is a compact example of a C library that survives because it offers a hard-to-replace semantic feature: POSIX-like regular expressions with approximate matching and predictable execution behavior.

The Homebrew formula's significance is partly in the executable it exposes. Installing `tre` gives users `agrep`, a niche but memorable command for fuzzy grep workflows, while also supplying headers and libraries that downstream software can build against.

Timeline

• 0.5.x: NEWS records pkg-config support, system ABI work, minimal repetition operators, and best-match `agrep` mode.
• 0.7.6: NEWS records license change from LGPL to BSD-style and removal of gnulib due to potential license conflicts.
• 0.8.0: NEWS records `tre_` prefixes for exported libtre functions and an ABI change.
• 0.9.0: NEWS records raw byte-vector API variants and updated Visual Studio project files.

Related projects

• Related to POSIX regex implementations such as Rx, Regex++, and AT&T ast regex, which the README mentions in its discussion of subpattern correctness.
• Related to earlier agrep implementations by Sun Wu and Udi Manber, which the README contrasts with TRE's included `agrep`.
• Related to grep-like command-line search tools, but distinguished by edit-distance matching.

Sources

• Batch input fields: source_facts.description, source_facts.executables, source_facts.package-manager
• Official NEWS file: https://github.com/laurikari/tre/blob/master/NEWS
• Official README: https://github.com/laurikari/tre#readme