File indexing completed on 2024-12-15 05:11:24

 #!/bin/sh
 
 # Apply kdelibs coding style to all c, cpp and header files in and below the current directory
 #
 # The coding style is defined in http://techbase.kde.org/Policies/Kdelibs_Coding_Style
 #
 # Requirements:
 # - installed astyle, with patches below
 # - $QT_NORMALIZE_TOOL pointing to qtrepotools/util/normalize/normalize (after compiling it)
 #
 # IMPORTANT: use astyle 2.05 or later. Patch astyle if you use older version.
 # Older versions of astyle misparse the Qt "keywords" like foreach and Q_FOREACH, which
 # inside the parenthesis, leading to things like:
 # Q_FOREACH(const QString & it, l) // note the space after the '&'.
 #
 # To fix this, patch astyle with http://www.davidfaure.fr/kde/astyle_qt.diff
 # Mentionned upstream in https://sourceforge.net/p/astyle/bugs/154/
 #
 # Another feature was implemented: removing spaces before ',' and ';'.
 # Apply http://www.davidfaure.fr/kde/astyle_comma.diff
 # This has been sent upstream in https://sourceforge.net/p/astyle/bugs/100/
 #
 # Instructions for OpenSuSE users:
 #   zypper si astyle
 #   cd ~/rpmbuild/SOURCES ; wget http://www.davidfaure.fr/kde/astyle_qt.diff
 #                           wget http://www.davidfaure.fr/kde/astyle_comma.diff
 #   cd ../SPECS ; wget http://www.davidfaure.fr/kde/astyle.spec.diff
 #   patch astyle.spec < astyle.spec.diff
 #   rpmbuild -ba astyle.spec
 #   sudo rpm -Uvh --force ~/rpmbuild/RPMS/x86_64/astyle-*.rpm
 
 files=`find -type f -name '*.c' -or -name '*.cpp' -or -name '*.cc' -or -name '*.h'`
 if [ -z "$files" ]; then
   # nothing to do
   exit 0
 fi
 
 if test -z "$QT_NORMALIZE_TOOL"; then
   echo "Please export QT_NORMALIZE_TOOL=<qt5>/qtrepotools/util/normalize/normalize"
   exit 1
 fi
 
 astyle -n -Q \
       --indent=spaces=4 --style=otbs \
       --indent-labels --pad-oper --unpad-paren --pad-header \
       --keep-one-line-statements \
       --convert-tabs \
       --indent-preprocessor \
       --align-pointer=name \
       $files
 
 # --max-code-length=100 is not used, on purpose. It's a hard limit, sometimes making things
 # less readable than if the code is kept on a single line. 100 is just a recommendation
 # in the coding style.
 
 # Watch out for things that lead to method implementations being parsed as if inside other methods,
 # e.g. due to '{' inside #ifdef and #else and '}' outside.
 grep '^\S* \S*::.*) {$' $files && echo "WARNING: check for wrong '{' placement in method definitions, in above grep results"
 
 # Remove old emacs mode-lines
 perl -pi -e '$_ = "" if /c-basic-offset: [1-8]/' $files
 
 # Remove old kate mode-lines
 perl -pi -e '$_ = "" if /kate: .*indent-width/ || /kate:.*tab-width/' $files
 
 # Remove old vim mode-lines
 perl -pi -e '$_ = "" if /\/\/.* vim:/' $files
 # They are often in a two-liner C comment, so we need a bit of perl magic to remove these
 perl - $files <<EOF
   foreach my \$file (@ARGV) {
    open F, "+<", \$file or do { print STDERR "open(\$file) failed : \"\$!\"\n"; next };
    my \$str = join '', <F>;
    if( \$str =~ m/vim:/ ) {
      #print STDERR "Removing multi-line vim modeline from \$file\n";
      \$str =~ s!/\*\**\s*\**\s*vim:[^\n]*\n\s*\*/!!smg;
      seek F, 0, 0;
      print F \$str;
      truncate F, tell(F);
    }
    close F;
   }
 EOF
 
 # Remove consecutive blank lines
 perl - $files <<EOF
   foreach my \$file (@ARGV) {
    open F, "+<", \$file or do { print STDERR "open(\$file) failed : \"\$!\"\n"; next };
    my \$str = join '', <F>;
    if (\$str =~ s/\s*\n\s*\n\s*\n\n*/\n\n/smg ) {
      seek F, 0, 0;
      print F \$str;
      truncate F, tell(F);
    }
    close F;
   }
 EOF
 
 # Normalize signals/slots
 $QT_NORMALIZE_TOOL --modify .
 
 # TODO: add command-line option to trigger this
 # It's not wanted when working on a fix
 #git commit -q -a -m "Code reformatted using kde-dev-scripts/astyle-kdelibs.
 #Use git blame -w `git rev-parse --short HEAD` to show authorship as it was before this commit."