Merging upstream version 1.21.

Signed-off-by: Daniel Baumann <daniel@debian.org>

NEWS

@@ -1,46 +1,52 @@
-Changes in version 1.20:
+Changes in version 1.21:
 
-Splitting was broken for files larger than 64 KiB because of a bug
-introduced in version 1.16.
+The options '--dump', '--remove' and '--strip' have been added, mainly as
+support for the tarlz archive format: http://www.nongnu.org/lzip/tarlz.html
+These options replace '--dump-tdata', '--remove-tdata' and '--strip-tdata',
+which are now aliases and will be removed in version 1.22.
 
-The options "--dump-tdata", "--remove-tdata", and "--strip-tdata" have
-been added to ease the management of metadata stored as trailing data:
+  '--dump=[<member_list>][:damaged][:tdata]' dumps the members listed, the
+  damaged members (if any), or the trailing data (if any) of one or more
+  regular multimember files to standard output.
 
-  "--dump-tdata" dumps the trailing data (if any) of one or more regular
-  files to standard output.
+  '--remove=[<member_list>][:damaged][:tdata]' removes the members listed,
+  the damaged members (if any), or the trailing data (if any) from regular
+  multimember files in place.
 
-  "--remove-tdata" removes the trailing data from regular files in place.
+  '--strip=[<member_list>][:damaged][:tdata]' copies one or more regular
+  multimember files to standard output, stripping the members listed, the
+  damaged members (if any), or the trailing data (if any) from each file.
 
-  "--strip-tdata" copies one or more regular files to standard output,
-  stripping the trailing data (if any) from each file.
+Detection of forbidden combinations of characters in trailing data has been
+improved.
 
-The option '--loose-trailing', has been added.
+'--split' can now detect trailing data and gaps between members, and save
+each gap in its own file. Trailing data (if any) are saved alone in the last
+file. (Gaps may contain garbage or may be members with corrupt headers or
+trailers).
 
-The test used by lziprecover to discriminate trailing data from a corrupt
-header in multimember or concatenated files has been improved to a
-Hamming distance (HD) of 3, and the 3 bit flips must happen in different
-magic bytes for the test to fail. As a consequence some kinds of files
-no longer can be appended to a lzip file as trailing data unless the
-'--loose-trailing' option is used when decompressing.
-Lziprecover can be used to remove conflicting trailing data from a file.
+'--ignore-errors' now makes '--list' show gaps between members, ignoring
+format errors.
 
-The contents of a corrupt or truncated header found in a multimember
-file is now shown, after the error message, in the same format as
-trailing data.
+'--ignore-errors' now makes '--range-decompress' ignore a truncated last
+member.
 
-The 'bits/byte' ratio has been replaced with the inverse compression
-ratio in the output.
+Errors are now also checked when closing the input file in decompression
+mode.
 
-The progress of decompression is now shown at verbosity level 2 (-vv) or
-higher.
+Some diagnostic messages have been improved.
 
-Progress of decompression is only shown if stderr is a terminal.
+'\n' is now printed instead of '\r' when showing progress of merge or repair
+if stdout is not a terminal.
 
-A final diagnostic is now shown at verbosity level 1 (-v) or higher if
-any file fails the test when testing multiple files.
+Lziprecover now compiles on DOS with DJGPP. (Patch from Robert Riebisch).
 
-In case of (de)compressed size mismatch, the stored size is now also
-shown in hexadecimal to ease visual comparison.
+The new chapter 'Tarlz', explaining the ways in which lziprecover can
+recover and process multimember tar.lz archives, has been added to the
+manual.
 
-The dictionary size is now shown at verbosity level 4 (-vvvv) when
-decompressing or testing.
+The configure script now accepts appending options to CXXFLAGS using the
+syntax 'CXXFLAGS+=OPTIONS'.
+
+It has been documented in INSTALL the use of
+CXXFLAGS+='-D __USE_MINGW_ANSI_STDIO' when compiling on MinGW.