archives2git

   1 #!/bin/sh
   2 # Copyright © 2013,2014,2017 Géraud Meyer <graud@gmx.com>
   3 #
   4 # This program is free software; you can redistribute it and/or modify
   5 # it under the terms of the GNU General Public License version 2 as
   6 # published by the Free Software Foundation.
   7 #
   8 # This program is distributed in the hope that it will be useful, but
   9 # WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  10 # or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  11 # for more details.
  12 #
  13 # You should have received a copy of the GNU General Public License along
  14 # with this program.  If not, see <http://www.gnu.org/licenses/>.
  15
  16 PROGRAM_NAME="archives2git"
  17 PROGRAM_VERSION="0.2+"
  18 PROGRAM_USAGE="$PROGRAM_NAME [<options>] [--] <archives_and_directories>"
  19 help_message () {
  20         cat <<HelpMessage
  21 NAME
  22         $PROGRAM_NAME - generate a Git history from a series of release tarballs
  23 USAGE
  24         $PROGRAM_USAGE
  25 DESCRIPTION
  26         $PROGRAM_NAME replaces the contents of a Git repository subdir with that of
  27         an <archive> file (or of another <directory>) and commits the changes.
  28 OPTIONS
  29         The options are parsed by git-rev-parse(1) with the option --parseopt.
  30 $( helpm_options )
  31 SHELL VARIABLES
  32         The following shell variables are usable if appropriate in the shell code
  33         snippets passed as command line options: \$arch, \$archdir, \$archname,
  34         \$file, \$date, \$author, \$title, \$tagname.
  35 SHELL FUNCTIONS
  36         The following shell functions are defined in the script and are usable in
  37         the shell code snippets: get_version(filename).
  38 ENVIRONMENT
  39         ARCHIVES2GIT_ENV
  40                 After parameter and command substitutions and arithmetic expansion, it
  41                 gives the path of the config file to source.  If unset the default
  42                 configuration file is sourced instead.
  43         GIT_AUTHOR_NAME
  44         GIT_AUTHOR_EMAIL
  45         GIT_COMMITTER_NAME
  46         GIT_COMMITTER_EMAIL
  47 FILES
  48         ~/.archives2gitrc
  49                 This file is sourced by the script.  An example file is distributed.
  50 EXAMPLES
  51         Some possible options:
  52           $ archives2git \$(ls ../oldproject-*.tgz | sort -V) \\
  53             # keep (only) the .gitignore file
  54             --keep-filter 'test x.gitignore = x"\$file"'
  55             --keep-filter 'echo "\$file" | grep -q -f \$HOME/filepats'
  56             # suppress the version info from the root directory in the archive
  57             --rename 'echo "\${file%-*}"'
  58             # allow to add ignored files
  59             --gitadd-arg -f
  60             # use the archive modification time as the author date (with GNU date)
  61             --date 'LC_ALL=C date -r "\$arch"'
  62             # do not include the path components in the commit title
  63             --title 'echo "\${arch##*/}"'
  64             # interactively edit each commit message
  65             --gitci-arg "--edit"
  66             # keep only the files of the project in the root of the tree
  67             --gitfilter $'--subdirectory-filter\noldproject'
  68             # move the project's files out of their subdir (while keeping other files)
  69             --gitfilter $'--tree-filter\nmv oldproject/.* .; mv oldproject/* .'
  70         See also the example configuration file.
  71 NOTES
  72         The code snippets passed to eval are not guaranteed to run in a subshell,
  73         so make sure they do not change the shell environment.
  74 BUGS
  75         Trailing whitespaces in filenames are not supported.
  76 AUTHOR
  77         $PROGRAM_NAME was written by G.raud Meyer.
  78 SEE ALSO
  79         git-commit-tree(1), git_load_dirs(1), git-rev-parse(1)
  80 HelpMessage
  81 }
  82 OPTS_SPEC="\
  83 $PROGRAM_USAGE
  84 --
  85 unpack=      command to extract the files of the \$arch in \$TMPDIR
  86 strip-ext=   sh code to strip the filename extension from \$file
  87 repo=        work tree subdir (current dir by default)
  88 rename=      sh code to rename the current \$file of the root
  89 keep-filter= sh code to decide whether not to remove the \$file of the root
  90 gitadd-arg=  argument to pass to git-add
  91 date=        sh code printing the commit author date for the current \$arch
  92 author=      sh code printing the commit author name for the current \$arch
  93 title=       sh code printing the commit title (and an optional body start)
  94 body=        string appended to the title as a new paragraph
  95 gitci-arg=   argument to pass to git-commit
  96 gitfilter=   line separated arguments to pass to git-filter
  97 tag          tag the commits with the default tagname
  98 tagname=     tag the commits with the name printed by the given sh code
  99 command=     sh code to eval after having commited and tagged
 100 debug        enable the verbose printing of what is done
 101 h,help       print a usage message and exit
 102 helpm        print the manual and exit
 103 man          view the manual in the pager and exit
 104 version      print the version information and exit
 105 "
 106
 107 # helper functions
 108 helpm_options () {
 109         printf %s "$OPTS_SPEC" | LC_ALL=C awk '/^--$/ { parse=1; next }
 110           function opt(spec) {
 111             if (sub(/^[a-zA-Z0-9][=*?!]{0,3}(,|$)/, "-&", spec)) sub(/^-.[=*?!]{0,3},/, "& --", spec)
 112             else sub(/^/, "--&", spec)
 113             return spec
 114           }
 115           parse { print " \t" opt($1); sub(/^[^ \t]*[ \t]*/,""); if ($0) print "\t\t" $0 }'
 116 }
 117 get_version () {
 118         _file=$1; shift
 119         _base=$( printf "%s\n" "$_file" |
 120           LC_ALL=C sed -r -e '$s;(([0-9]{1,}\.)|[-_])[0-9]{1,}([-_.]{0,1}([0-9]{1,}|[ab]|alpha|beta|gamma))*$;;' )
 121         _version=${_file#"$_base"}
 122         _version=${_version#"-"}; _version=${_version#"."}; _version=${_version#"_"}
 123         printf "%s\n" "${_version}"
 124 }
 125 strip_extension () {
 126         printf "%s\n" "$1" | LC_ALL=C sed -r \
 127           -e '$s;^([^/]*/)*([^/]{1,})/{0,1}$;\2;' \
 128           -e '$s;\.([Zz]|lzo|bz|gz|bz2|xz|lzma|7z|lz|rz|lrz)$;;' \
 129           -e '$s;\.(tar|t[Zz]|tzo|tbz|tgz|tbz2|txz|t7z|tlz|lha|lzh|7z|alz|arj|zip|rar|jar|war|ace|cab|a|cpio|shar|rpm|deb)$;;' \
 130           -e '$s;\.(orig)$;;'
 131 }
 132
 133 # default parameters
 134 UNPACK=
 135 STRIP_EXT='strip_extension "$arch"'
 136 RENAME='' # $file $arch $archname
 137 FILTER='false' # $file $arch $archname
 138 DATE='' # $arch $archname
 139 TITLE='echo "$arch"' # $arch $archname
 140 BODY="" # message taken as is
 141 ADDARGS=
 142 CIARGS=
 143 FILTERHEAD= # line separated arguments
 144 TAG=
 145 TAGDEF='printf version/; get_version "$archname"'
 146 COMMAND=
 147 # config (file, environment, command line)
 148 if [ -n "${ARCHIVES2GIT_ENV+set}" ]
 149 then
 150         dquotedexpr=$( printf "%s" "$ARCHIVES2GIT_ENV" | LC_ALL=C sed 's/"/\\"/g' )
 151           # double quoting still allows backslash escapes (as well as substitutions
 152           # and expansions)
 153         configfile=
 154         (eval : '${configfile:="'"$dquotedexpr"'"}') &&
 155         eval : '${configfile:="'"$dquotedexpr"'"}'
 156           # use := instead of = as the substitution for more portability (which is
 157           # not that sensible since the + substitution is not implemented either
 158           # when = is not)
 159 else
 160         configfile=$HOME/.archives2gitrc
 161 fi
 162 [ -r "$configfile" ] && . "$configfile"
 163 GIT_WORK_TREE=${GIT_WORK_TREE-.}
 164   # may be a subdir of the top level
 165 parsing=$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@"); rc=$?
 166 eval "$parsing"
 167 if [ $rc -ne 0 ]
 168 then printf %s "$parsing" | grep '^EOF' >/dev/null && exit 0 || exit $rc
 169 fi
 170 while [ $# -gt 0 ]
 171 do
 172         case $1 in
 173         --unpack)
 174                 shift; UNPACK=$1 ;;
 175         --strip-ext)
 176                 shift; STRIP_EXT=$1 ;;
 177         --repo)
 178                 shift; GIT_WORK_TREE=$1 ;;
 179         --rename)
 180                 shift; RENAME=$1 ;;
 181         --keep-filter)
 182                 shift; FILTER=$1 ;;
 183         --gitadd-arg)
 184                 shift; ADDARGS="$ADDARGS $1" ;;
 185         --date)
 186                 shift; DATE=$1 ;;
 187         --author)
 188                 shift; AUTHOR=$1 ;;
 189         --title)
 190                 shift; TITLE=$1 ;;
 191         --body)
 192                 shift; BODY=$1 ;;
 193         --gitci-arg)
 194                 shift; CIARGS="$CIARGS $1" ;;
 195         --gitfilter)
 196                 shift; FILTERHEAD=$1 ;;
 197         --tag)
 198                 TAG=$TAGDEF ;;
 199         --tagname)
 200                 shift; TAG=$1 ;;
 201         --command)
 202                 shift; COMMAND=$1 ;;
 203         --debug)
 204                 set -x ;;
 205         --helpm)
 206                 help_message
 207                 exit ;;
 208         --man)
 209                 help_message | helpm2pod --man - ARCHIVES2GIT
 210                 exit ;;
 211         --version)
 212                 echo "$PROGRAM_NAME version $PROGRAM_VERSION"
 213                 exit ;;
 214         --?*)
 215                 echo >&2 "$PROGRAM_NAME: internal error: invalid option $1"
 216                 exit 255 ;;
 217         --)
 218                 shift; break ;;
 219         *)
 220                 break ;;
 221         esac
 222         shift
 223 done
 224 # setup
 225 NL='
 226 '
 227 trap 'rm -rf "$TMPDIR"' EXIT INT QUIT TERM
 228 TMPDIR=$(mktemp -d) ||
 229 {
 230         echo >&2 "$PROGRAM_NAME: error: cannot create a temporary directory"
 231         exit 253
 232 }
 233 WRKDIR=$(pwd)
 234 # git repository check
 235 cd "$GIT_WORK_TREE" &&
 236 { test -d .git || git rev-parse --git-dir >/dev/null; } ||
 237 {
 238         echo >&2 "$PROGRAM_NAME: error: not in a git repository ($(pwd))"
 239         exit 255
 240 }
 241 test -z "$(git status --porcelain)" &&
 242 git update-index -q --refresh ||
 243 {
 244         echo >&2 "$PROGRAM_NAME: error: unstaged files or dirty index"
 245         exit 254
 246 }
 247
 248 # main
 249 set -e
 250 OLDIFS=$IFS
 251 for arch
 252 do
 253         # put the files in the archive dir
 254         cd "$WRKDIR"
 255         if [ -d "$arch" ]
 256         then
 257                 cp -R "$arch" "$TMPDIR"
 258         elif [ -n "$UNPACK" ]
 259         then
 260                 eval "$UNPACK"
 261         else
 262                 aunpack -X "$TMPDIR" "$arch"
 263         fi
 264         archdir=$TMPDIR
 265         archname=$(eval "$STRIP_EXT")
 266         # remove (almost) everything from the repository dir
 267         cd "$GIT_WORK_TREE"
 268         for file in * .*
 269         do
 270                 [ -e "$file" ] || [ -h "$file" ] || continue
 271                 [ x"$file" = x"." ] || [ x"$file" = x".." ] || [ x"$file" = x".git" ] && continue
 272                 eval "$FILTER" || { git rm -r "$file" || rm -R "$file"; }
 273         done
 274         # move the content of the temp dir to the repository and stage it
 275         for file in "$archdir"/* "$archdir"/.*
 276         do
 277                 [ -e "$file" ] || [ -h "$file" ] || continue
 278                 file=${file#$archdir/}
 279                 [ x"$file" = x"." ] || [ x"$file" = x".." ] && continue
 280                 [ -n "$RENAME" ] && name=$(eval "$RENAME") || name=$file
 281                 if [ -n "$name" ]
 282                 then
 283                         if [ -e ./"$name" ]
 284                         then
 285                           # remove conflicting file (previously kept or renamed to the same name)
 286                                 rm -R ./"$name"
 287                                 echo >&2 "WARNING: conflicting file removed: \`$name'"
 288                         fi
 289                         mv "$archdir"/"$file" ./"$name"
 290                         git add $ADDARGS ./"$name"
 291                 else
 292                         rm -R "$archdir"/"$file"
 293                 fi
 294                 unset name
 295         done
 296         unset file
 297         # commit, filter and clean up
 298         date=$(eval "$DATE")
 299         author=$(eval "$AUTHOR")
 300         title=$(eval "$TITLE")
 301         git commit ${AUTHOR:+--author "$author"} -m "$title"${BODY:+"$NL$NL$BODY"} ${DATE:+--date "$date"} $CIARGS
 302         if [ -n "$FILTERHEAD" ]
 303         then
 304                 IFS=$NL
 305                 git filter-branch $FILTERHEAD -- HEAD^..HEAD
 306                 rm -R "$(git rev-parse --git-dir)"/refs/original/
 307                 IFS=$OLDIFS
 308         fi
 309         if [ -n "$TAG" ]
 310         then
 311                 tagmessage=$(eval "$TAG" | LC_ALL=C tail -n +2)
 312                 tagname=$(eval "$TAG" | LC_ALL=C head -1)
 313                 git tag ${tagmessage:+-m "$tagmessage"} "$tagname"
 314         fi
 315         if [ -n "$COMMAND" ]
 316         then
 317                 eval "$COMMAND"
 318         fi
 319         unset title date author tagname tagmessage
 320 done