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         .git/archives2gitrc
  51                 This repository configuration file is preferred over the global one
  52                 if it exists.
  53 EXAMPLES
  54         Some possible options:
  55           $ ARCHIVES2GIT_ENV=/dev/null \\
  56             archives2git \$(ls ../oldproject-*.tgz | sort -V) \\
  57             # keep (only) the .gitignore file
  58             --keep-filter 'test x.gitignore = x"\$file"'
  59             --keep-filter 'echo "\$file" | grep -q -f \$HOME/filepats'
  60             # suppress the version info from the root directory in the archive
  61             --rename 'echo "\${file%-*}"'
  62             # allow to add ignored files
  63             --gitadd-arg -f
  64             # use the archive modification time as the author date (with GNU date)
  65             --date 'LC_ALL=C date -r "\$arch"'
  66             # do not include the path components in the commit title
  67             --title 'echo "\${arch##*/}"'
  68             # interactively edit each commit message
  69             --gitci-arg "--edit"
  70             # keep only the files of the project in the root of the tree
  71             --gitfilter $'--subdirectory-filter\noldproject'
  72             # move the project's files out of their subdir (while keeping other files)
  73             --gitfilter $'--tree-filter\nmv oldproject/.* .; mv oldproject/* .'
  74         See also the example configuration file.
  75 BUGS
  76         Trailing whitespaces in filenames are not supported.
  77 AUTHOR
  78         $PROGRAM_NAME was written by G.raud Meyer.
  79 SEE ALSO
  80         git-commit-tree(1), git_load_dirs(1), git-rev-parse(1)
  81 HelpMessage
  82 }
  83 OPTS_SPEC="\
  84 $PROGRAM_USAGE
  85 --
  86 unpack=      command to extract the files of the \$arch in \$TMPDIR
  87 strip-ext=   sh code to strip the filename extension from \$file
  88 repo=        work tree subdir (current dir by default)
  89 rename=      sh code to rename the current \$file of the root
  90 keep-filter= sh code to decide whether not to remove the \$file of the root
  91 gitadd-arg=  argument to pass to git-add
  92 date=        sh code printing the commit author date for the current \$arch
  93 author=      sh code printing the commit author name for the current \$arch
  94 title=       sh code printing the commit title (and an optional body start)
  95 body=        string appended to the title as a new paragraph
  96 gitci-arg=   argument to pass to git-commit
  97 gitfilter=   line separated arguments to pass to git-filter
  98 tag          tag the commits with the default tagname
  99 tagname=     tag the commits with the name printed by the given sh code
 100 command=     sh code to eval after having commited and tagged
 101 debug        enable the verbose printing of what is done
 102 keep-tempdir do not delete the temporary extraction dirs
 103 h,help       print a usage message and exit
 104 helpm        print the manual and exit
 105 man          view the manual in the pager and exit
 106 version      print the version information and exit
 107 "
 108
 109 # helper functions
 110 helpm_options () {
 111         printf %s "$OPTS_SPEC" | LC_ALL=C awk '/^--$/ { parse=1; next }
 112           function opt(spec) {
 113             if (sub(/^[a-zA-Z0-9][=*?!]{0,3}(,|$)/, "-&", spec)) sub(/^-.[=*?!]{0,3},/, "& --", spec)
 114             else sub(/^/, "--&", spec)
 115             return spec
 116           }
 117           parse { print " \t" opt($1); sub(/^[^ \t]*[ \t]*/,""); if ($0) print "\t\t" $0 }'
 118 }
 119 get_version () {
 120         _file=$1; shift
 121         _base=$( printf "%s\n" "$_file" |
 122           LC_ALL=C sed -r -e '$s;(([0-9]{1,}\.)|[-_])[0-9]{1,}([-_.]{0,1}([0-9]{1,}|[ab]|alpha|beta|gamma))*$;;' )
 123         _version=${_file#"$_base"}
 124         _version=${_version#"-"}; _version=${_version#"."}; _version=${_version#"_"}
 125         printf "%s\n" "${_version}"
 126 }
 127 strip_extension () {
 128         printf "%s\n" "$1" | LC_ALL=C sed -r \
 129           -e '$s;^([^/]*/)*([^/]{1,})/{0,1}$;\2;' \
 130           -e '$s;\.([Zz]|lzo|bz|gz|bz2|xz|lzma|7z|lz|rz|lrz)$;;' \
 131           -e '$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)$;;' \
 132           -e '$s;\.(orig)$;;'
 133 }
 134 NL='
 135 '
 136
 137 # default parameters
 138 UNPACK=
 139 STRIP_EXT='strip_extension "$arch"'
 140 RENAME='' # $file $arch $archname
 141 FILTER='false' # $file $arch $archname
 142 DATE='' # $arch $archname
 143 TITLE='echo "$arch"' # $arch $archname
 144 BODY="" # message taken as is
 145 ADDARGS=
 146 CIARGS=
 147 FILTERHEAD= # line separated arguments
 148 TAG=
 149 TAGDEF='printf version/; get_version "$archname"'
 150 COMMAND=
 151 KEEP_TEMPDIR=
 152 # config (file, environment, command line)
 153 if [ -n "${ARCHIVES2GIT_ENV+set}" ]
 154 then
 155         dquotedexpr=$( printf "%s" "$ARCHIVES2GIT_ENV" | LC_ALL=C sed 's/"/\\"/g' )
 156           # double quoting still allows backslash escapes (as well as substitutions
 157           # and expansions)
 158         configfile=
 159         (eval : '${configfile:="'"$dquotedexpr"'"}') &&
 160         eval : '${configfile:="'"$dquotedexpr"'"}'
 161           # use := instead of = as the substitution for more portability (which is
 162           # not that sensible since the + substitution is not implemented either
 163           # when = is not)
 164 elif
 165         gitdir=$(git rev-parse --git-dir) &&
 166         [ -f "$gitdir"/archives2gitrc ]
 167 then
 168         configfile=$gitdir/archives2gitrc
 169 else
 170         configfile=$HOME/.archives2gitrc
 171 fi
 172 [ -r "$configfile" ] && . "$configfile"
 173 GIT_WORK_TREE=${GIT_WORK_TREE-.}
 174   # may be a subdir of the top level
 175 parsing=$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@"); rc=$?
 176 eval "$parsing"
 177 if [ $rc -ne 0 ]
 178 then printf %s "$parsing" | grep '^EOF' >/dev/null && exit 0 || exit $rc
 179 fi
 180 while [ $# -gt 0 ]
 181 do
 182         case $1 in
 183         --unpack)
 184                 shift; UNPACK=$1 ;;
 185         --strip-ext)
 186                 shift; STRIP_EXT=$1 ;;
 187         --repo)
 188                 shift; GIT_WORK_TREE=$1 ;;
 189         --rename)
 190                 shift; RENAME=$1 ;;
 191         --keep-filter)
 192                 shift; FILTER=$1 ;;
 193         --gitadd-arg)
 194                 shift; ADDARGS="$ADDARGS$NL$1" ;;
 195         --date)
 196                 shift; DATE=$1 ;;
 197         --author)
 198                 shift; AUTHOR=$1 ;;
 199         --title)
 200                 shift; TITLE=$1 ;;
 201         --body)
 202                 shift; BODY=$1 ;;
 203         --gitci-arg)
 204                 shift; CIARGS="$CIARGS$NL$1" ;;
 205         --gitfilter)
 206                 shift; FILTERHEAD=$1 ;;
 207         --tag)
 208                 TAG=$TAGDEF ;;
 209         --tagname)
 210                 shift; TAG=$1 ;;
 211         --command)
 212                 shift; COMMAND=$1 ;;
 213         --debug)
 214                 set -x ;;
 215         --keep-tempdir)
 216                 KEEP_TEMPDIR=yes ;;
 217         --helpm)
 218                 help_message
 219                 exit ;;
 220         --man)
 221                 help_message | helpm2pod --man - ARCHIVES2GIT
 222                 exit ;;
 223         --version)
 224                 echo "$PROGRAM_NAME version $PROGRAM_VERSION"
 225                 exit ;;
 226         --?*)
 227                 echo >&2 "$PROGRAM_NAME: internal error: invalid option $1"
 228                 exit 255 ;;
 229         --)
 230                 shift; break ;;
 231         *)
 232                 break ;;
 233         esac
 234         shift
 235 done
 236 # setup
 237 [ -z "$KEEP_TEMPDIR" ] && trap 'rm -rf "$TMPDIR"' EXIT INT QUIT TERM
 238 TMPDIR=$(mktemp -d) ||
 239 {
 240         echo >&2 "$PROGRAM_NAME: error: cannot create a temporary directory"
 241         exit 253
 242 }
 243 WRKDIR=$(pwd)
 244 # git repository check
 245 cd "$GIT_WORK_TREE" &&
 246 { test -d .git || git rev-parse --git-dir >/dev/null; } ||
 247 {
 248         echo >&2 "$PROGRAM_NAME: error: not in a git repository ($(pwd))"
 249         exit 255
 250 }
 251 test -z "$(git status --porcelain)" &&
 252 git update-index -q --refresh ||
 253 {
 254         echo >&2 "$PROGRAM_NAME: error: unstaged files or dirty index"
 255         exit 254
 256 }
 257
 258 # main
 259 set -e
 260 OLDIFS=$IFS
 261 for arch
 262 do
 263         # put the files in the archive dir
 264         cd "$WRKDIR"
 265         if [ -d "$arch" ]
 266         then
 267                 cp -R "$arch" "$TMPDIR"
 268         elif [ -n "$UNPACK" ]
 269         then
 270                 (eval "$UNPACK")
 271         else
 272                 aunpack -X "$TMPDIR" "$arch"
 273         fi
 274         archdir=$TMPDIR
 275         archname=$(eval "$STRIP_EXT")
 276         # remove (almost) everything from the repository dir
 277         cd "$GIT_WORK_TREE"
 278         for file in * .*
 279         do
 280                 [ -e "$file" ] || [ -h "$file" ] || continue
 281                 [ x"$file" = x"." ] || [ x"$file" = x".." ] || [ x"$file" = x".git" ] && continue
 282                 (eval "$FILTER") || { git rm -r "$file" || rm -R "$file"; }
 283         done
 284         # move the content of the temp dir to the repository and stage it
 285         for file in "$archdir"/* "$archdir"/.*
 286         do
 287                 [ -e "$file" ] || [ -h "$file" ] || continue
 288                 file=${file#$archdir/}
 289                 [ x"$file" = x"." ] || [ x"$file" = x".." ] && continue
 290                 [ -n "$RENAME" ] && name=$(eval "$RENAME") || name=$file
 291                 if [ -n "$name" ]
 292                 then
 293                         if [ -e ./"$name" ]
 294                         then
 295                           # remove conflicting file (previously kept or renamed to the same name)
 296                                 rm -R ./"$name"
 297                                 echo >&2 "WARNING: conflicting file removed: \`$name'"
 298                         fi
 299                         mv "$archdir"/"$file" ./"$name"
 300                         IFS=$NL; set -f
 301                         git add $ADDARGS ./"$name"
 302                         set +f; IFS=$OLDIFS
 303                 else
 304                         rm -R "$archdir"/"$file"
 305                 fi
 306                 unset name
 307         done
 308         unset file
 309         # commit, filter and clean up
 310         date=$(eval "$DATE")
 311         author=$(eval "$AUTHOR")
 312         title=$(eval "$TITLE")
 313         IFS=$NL; set -f
 314         git commit ${AUTHOR:+--author$NL"$author"} -m "$title"${BODY:+"$NL$NL$BODY"} ${DATE:+--date$NL"$date"} $CIARGS
 315         if [ -n "$FILTERHEAD" ]
 316         then
 317                 git filter-branch $FILTERHEAD -- HEAD^..HEAD
 318                 rm -R "$(git rev-parse --git-dir)"/refs/original/
 319         fi
 320         set +f; IFS=$OLDIFS
 321         if [ -n "$TAG" ]
 322         then
 323                 tagmessage=$(eval "$TAG" | LC_ALL=C tail -n +2)
 324                 tagname=$(eval "$TAG" | LC_ALL=C head -1)
 325                 git tag ${tagmessage:+-m "$tagmessage"} "$tagname"
 326         fi
 327         if [ -n "$COMMAND" ]
 328         then
 329                 (eval "$COMMAND")
 330         fi
 331         unset title date author tagname tagmessage
 332 done