TRE (regex matching library)
Lightweight, robust, and efficient POSIX compliant regexp matching library
Website: https://laurikari.net/tre/
Github: https://github.com/laurikari/tre/
Platforms supported: Windows, Linux, FreeBSD
Headers to include: tre/tre.bi, tre/regex.bi
Header version: 0.8.0
Examples: in examples/regex/TRE/
The Windows examples and their native data layouts were checked with the real TRE 0.9.0 SDK on Win32 and Win64. The binding's version remains 0.8.0. Use a library built for the executable's bitness.
A successful regcomp acquires resources in regex_t. Call regfree after matching, including when no match is found. regmatch_t offsets are zero-based byte offsets; an unmatched optional capture has negative offsets and must be checked before extracting a substring.
match.bas prints successive identifiers and waits for a key. replace.bas accepts numbered replacement captures such as \1, with \0 selecting the whole match. A doubled backslash inserts one literal backslash. Its demonstration prints Hi *you* strange *user* :D. subexpr.bas extracts filename components as a text example and does not access the filesystem.
The match and replacement examples use regnexec with explicit byte lengths. Empty matches advance by one byte and stop after a match at the end, which keeps iteration finite and preserves intervening subject bytes. These examples are intended for byte-oriented input; their global iteration has not been qualified for a multibyte locale. Empty results also represent compile or execution failures in the replacement helper.
Website: https://laurikari.net/tre/
Github: https://github.com/laurikari/tre/
Platforms supported: Windows, Linux, FreeBSD
Headers to include: tre/tre.bi, tre/regex.bi
Header version: 0.8.0
Examples: in examples/regex/TRE/
The Windows examples and their native data layouts were checked with the real TRE 0.9.0 SDK on Win32 and Win64. The binding's version remains 0.8.0. Use a library built for the executable's bitness.
A successful regcomp acquires resources in regex_t. Call regfree after matching, including when no match is found. regmatch_t offsets are zero-based byte offsets; an unmatched optional capture has negative offsets and must be checked before extracting a substring.
match.bas prints successive identifiers and waits for a key. replace.bas accepts numbered replacement captures such as \1, with \0 selecting the whole match. A doubled backslash inserts one literal backslash. Its demonstration prints Hi *you* strange *user* :D. subexpr.bas extracts filename components as a text example and does not access the filesystem.
The match and replacement examples use regnexec with explicit byte lengths. Empty matches advance by one byte and stop after a match at the end, which keeps iteration finite and preserves intervening subject bytes. These examples are intended for byte-oriented input; their global iteration has not been qualified for a multibyte locale. Empty results also represent compile or execution failures in the replacement helper.