Wdiff project

The Wdiff program, now maintained by Denver Gingerich, contains a word difference finder and a diff format converter.  This Web site you are visiting is being deprecated, please use the Savannah project page instead. The remaining bits of this site, including this page of course, are doomed to disappear.

Here are a few random notes from the previous maintainer.

• 1   Recovering old distributions
• 2   Remaining cleanup
• 3   wdiff option integration
• 4   diff notes

1   Recovering old distributions

To unpack any of the oldish versions 0.5, 0.5h or 0.5g of Wdiff, use a a recipe like this one:

       git clone git://github.com/pinard/wdiff.git
    cd wdiff
       git checkout ':/Release 0.5g'

2   Remaining cleanup

• Submit accumulated suggestions from local email folders

3   wdiff option integration

  -C => ..
  -a => -A
  -c => -i
  -h => ..

* -1, --no-deleted           inhibit output of deleted words
* -2, --no-inserted          inhibit output of inserted words
* -3, --no-common            inhibit output of common words
  -B, --ignore-blank-lines   ignore blank lines
  -C, --context=LINES        as -c, also select context size in lines
  -D, --ifdef=NAME           output #ifdef NAME format
  -F, --show-function-line=REGEXP   show previous context matching REGEXP
  -H, --speed-large-files    go faster, for when numerous small changes
  -I, --ignore-matching-lines=REGEXP   ignore lines matching REGEXP
  -L, --label=LABEL          use from/to LABEL instead of file name (twice)
  -N, --new-file             consider missing files to be empty
  -P, --unidirectional-new-file     consider missing old files to be empty
  -S, --starting-file=FILE   resume directory comparison with FILE
  -T, --initial-tab          produce TAB instead of initial space
  -U, --unified=LINES        as -u, also select context size in lines
  -W, --width=COLUMNS        use width of COLONS
  -X, --exclude-from=FILE    ignore files matching patterns from FILE
  -a, --text                 report line differences (text file default)
  -b, --ignore-space-change  ignore amount of white space
  -c, --context              output context diffs (default 3 context lines)
  -d, --minimal              try harder for a smaller set of changes
  -e, --ed                   output a valid ed script
  -f, --forward-ed           mix between -e and -n (not very useful)
  -h                         (ignored)
  -i, --ignore-case          ignore case differences
  -l, --paginate             paginate output through pr
* -l, --less-mode            variation of printer mode for less
  -n, --rcs                  output RCS format (internally used by RCS)
* -n, --avoid-wraps          do not extend fields through newlines
  -p, --show-c-function      show which C function for each change
* -p, --printer              overstrike as for printers
  -q, --brief                only says if files differ (binary default)
  -r, --recursive            recursively compare subdirectories
  -s, --report-identical-files      report when two files are the same
* -s, --statistics           say how many words deleted, inserted etc.
  -t, --expand-tabs          expand tabs to spaces in the output
* -t, --terminal             use termcap as for terminal displays
  -u, --unified              output unidiffs (default 3 context lines)
  -v, --version              output version information and exit
  -w, --ignore-all-space     ignore white space
* -w, --start-delete STRING  string to mark beginning of delete region
  -x, --exclude=PATTERN      ignore files (dirs) matching PATTERN
* -x, --end-delete STRING    string to mark end of delete region
  -y, --side-by-side         use side by side output format
* -y, --start-insert STRING  string to mark beginning of insert region

• -z, --end-insert STRING    string to mark end of insert region

4   diff notes

• -Wb should disregard white lines even if non empty
• --ifdef=NAME doc may be wrong in Invoking diff
• Notes on Handling Files that Do Not Fit in Memory
• $(diff_o) (so getopt, regex, etc.) has diff.h as dependencies