hd2u 1.0.0 ---------- This stable release contains no functional changes since last development release (0.9.2). Only few typos in documentation has been fixed. Most notable changes since last stable release (0.8.2) are: * support for Macintosh text format * new --version (-V) option * documentation enhancements and fixes Peter Hanecak hd2u 0.9.2 ---------- In this release issues with some binary files being recognized as text files in Macintosh format should be resolved. Also documentation has been enhanced and some typos in it fixed. Peter Hanecak hd2u 0.9.1 ---------- New --version (-V) option has been added to display version information and build process has been enhanced. Peter Hanecak hd2u 0.9.0 ---------- This release starts development period. Most notable change is addition of support for Macintosh text format. Peter Hanecak hd2u 0.8.2 ---------- This release introduces read/write buffers which increases conversion performance about 9 times. Peter Hanecak hd2u 0.8.1 ---------- Temporary files are now created in directory defined by TMPDIR or TMP environment variable. If no such variable is found, '/tmp' directory is used. Also security has been enhanced by usage of 'mkstemp()' instead of 'tmpnam()'. Peter Hanecak hd2u 0.8.0 ---------- This stable release consists of documentation updates and fixes. Peter Hanecak hd2u 0.7.2 ---------- Compilation issues with gcc 2.96 were resolved and documentation has been fixed. Peter Hanecak hd2u 0.7.1 ---------- In this release, there are following changes: 1) convert() is reworked so it always "caches" input into tempfile (even when no conversion is done *). This makes dos2unix able to convert data from stdin again. Also scan/detect code has been separated from logic (conversion type selection) and verbose output. Also every conversion type (dos -> unix, unix -> dos, copy) has now its own optimised loop. 2) From above changes comes some performance issues (I do not know if anyone cares about them, but still I'm going to write about it): - conversion of big files should be faster (optimised loops for each conversion) - conversion of a lot of files should be slower (while there is full input scan and more logic) - even when just testing (--test and --verbose, --test alone makes just file reads without any output) temporary file is used to store input - this is because I did not wanted to slow scan with condition 'if (!testmode)' on each 'fputc()' thus slowing conversion; I think test will be used much less then conversions 3) As opposed to version 0.7.0, binary files are skipped only when --skipbin or -b option is used. I'm from Slovakia and our charset uses codes above 126 thus all our conversion which uses dos2unix 0.7.0 wont work on such files. So to maintain some backward compatibility from now binary files will be detected but not skipped by default. By the way, if you are using 'dos2unix *' you should be VERY CAREFULL!!! It's something like 'rm -rf .' if missused. So 'dos2unix *.txt *.c *.h doc/*.html' would be far better (even if not that comfortable). 4) Rouge '\r' characters (DOS format line breaks with missing '\n' after them) are now reported but affected only by 'dos -> unix' conversion - they are skipped. 5) If you give directory as input file to 'dos2unix' it will be skipped (I have better no idea what it was doing untill now). Peter Hanecak hd2u 0.7.0 ---------- This is development release based on changes from Rob Ginda which breaks conversion functionality when converting data from stdin. So users which requires proper functionality of stdin conversion should continue to use version 0.6.0 (or earlier) and wait for fixed version to be released (or make a fix for themselves - in such case I would be happy to receive that fix :). Peter Hanecak