File is a utility that classifies arbitrary files of any type mainly by their contents. It is considered standard on most Unix distributions. It can output either human-readable concise descriptions or MIME types for machine handling.
security: Misc buffer overruns and missing buffer size tests in cdf parsing. Enforce limit of 8K on regex searches that have no limits. Allow the l modifier for regex to mean line count. Default to byte count. If line count is specified, assume a max of 80 characters per line to limit the byte count. Don't allow conversions to be used for dates, allowing the mask field to be used as an offset. Make the range operator limit the length of the regex search. Windows fixes. Hangul word processor recognition. Encoding irregularities in text files. Fix uninitialized title in CDF files. Fix compilation of empty files. Fix integer formats: We don't specify 'l' or 'h' and 'hh' specifiers anymore, only 'll' for quads and nothing for the rest. This is so that magic writing is simpler. Fix memory leak. Fix out of bounds read. Fix issue with long formats not matching fmtcheck.