archives2git

   1 #!/bin/sh
   2 # Copyright © 2013,2014,2017,2018 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="0.3"
  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
  27         that of an <archive> file (or of another <directory>) and commits the
  28         changes.
  29 OPTIONS
  30         The options are parsed by git-rev-parse(1) with the option --parseopt.
  31 $( helpm_options )
  32 SHELL VARIABLES
  33         The following shell variables are usable if appropriate in the shell
  34         code snippets passed as command line options: \$arch, \$archdir,
  35         \$archname, \$file, \$date, \$author, \$title, \$tagname.
  36 SHELL FUNCTIONS
  37         The following shell functions are defined in the script and are usable
  38         in the shell code snippets: get_version(filename).
  39 ENVIRONMENT
  40         ARCHIVES2GIT_ENV
  41                 After parameter and command substitutions and arithmetic
  42                 expansion, it gives the path of the config file to source.  If
  43                 unset the default configuration file is sourced instead.
  44         GIT_AUTHOR_NAME
  45         GIT_AUTHOR_EMAIL
  46         GIT_COMMITTER_NAME
  47         GIT_COMMITTER_EMAIL
  48 FILES
  49         ~/.archives2gitrc
  50                 This file is sourced by the script.  An example file is
  51                 distributed.
  52         .git/archives2gitrc
  53                 This repository configuration file is preferred over the global
  54                 one if it exists.
  55 EXAMPLES
  56         Some possible options:
  57           $ ARCHIVES2GIT_ENV=/dev/null \\
  58             archives2git \$(ls ../oldproject-*.tgz | sort -V) \\
  59             # keep (only) the .gitignore file
  60             --keep-filter 'test x.gitignore = x"\$file"'
  61             --keep-filter 'echo "\$file" | grep -q -f \$HOME/filepats'
  62             # suppress the version info from the root directory in the archive
  63             --rename 'echo "\${file%-*}"'
  64             # allow to add ignored files
  65             --gitadd-arg -f
  66             # use the archive modification time as the author date (with GNU date)
  67             --date 'LC_ALL=C date -r "\$arch"'
  68             # do not include the path components in the commit title
  69             --title 'echo "\${arch##*/}"'
  70             # interactively edit each commit message
  71             --gitci-arg "--edit"
  72             # keep only the files of the project in the root of the tree
  73             --gitfilter $'--subdirectory-filter\noldproject'
  74             # move the project's files out of their subdir (while keeping other files)
  75             --gitfilter $'--tree-filter\nmv oldproject/.* .; mv oldproject/* .'
  76         See also the example configuration file.
  77 BUGS
  78         Trailing whitespaces in filenames are not supported.
  79 AUTHOR
  80         $PROGRAM_NAME was written by G.raud Meyer.
  81 SEE ALSO
  82         git-commit-tree(1), git_load_dirs(1), git-rev-parse(1)
  83 HelpMessage
  84 }
  85 OPTS_SPEC="\
  86 $PROGRAM_USAGE
  87 --
  88 unpack=      command to extract the files of the \$arch in \$TMPDIR
  89 strip-ext=   sh code to strip the filename extension from \$file
  90 repo=        work tree subdir (current dir by default)
  91 rename=      sh code to rename the current \$file of the root
  92 keep-filter= sh code to decide whether not to remove the \$file of the root
  93 gitadd-arg=  argument to pass to git-add
  94 date=        sh code printing the commit author date for the current \$arch
  95 author=      sh code printing the commit author name for the current \$arch
  96 title=       sh code printing the commit title (and an optional body start)
  97 body=        string appended to the title as a new paragraph
  98 gitci-arg=   argument to pass to git-commit
  99 gitfilter=   line separated arguments to pass to git-filter
 100 tag          tag the commits with the default tagname
 101 tagname=     tag the commits with the name printed by the given sh code
 102 command=     sh code to eval after having commited and tagged
 103 debug        enable the verbose printing of what is done
 104 keep-tempdir do not delete the temporary extraction dirs
 105 h,help       print a usage message and exit
 106 helpm        print the manual and exit
 107 man          view the manual in the pager and exit
 108 version      print the version information and exit
 109 "
 110
 111 # helper functions
 112 helpm_options () {
 113         printf %s "$OPTS_SPEC" | LC_ALL=C awk '/^--$/ { parse=1; next }
 114           function opt(spec) {
 115             if (sub(/^[a-zA-Z0-9][=*?!]{0,3}(,|$)/, "-&", spec)) sub(/^-.[=*?!]{0,3},/, "& --", spec)
 116             else sub(/^/, "--&", spec)
 117             return spec
 118           }
 119           parse { print " \t" opt($1); sub(/^[^ \t]*[ \t]*/,""); if ($0) print "\t\t" $0 }'
 120 }
 121 get_version () {
 122         _file=$1; shift
 123         _base=$( printf "%s\n" "$_file" |
 124           LC_ALL=C sed -r -e '$s;(([0-9]{1,}\.)|[-_])[0-9]{1,}([-_.]{0,1}([0-9]{1,}|[ab]|alpha|beta|gamma))*$;;' )
 125         _version=${_file#"$_base"}
 126         _version=${_version#"-"}; _version=${_version#"."}; _version=${_version#"_"}
 127         printf "%s\n" "${_version}"
 128 }
 129 warn () { echo >&2 "$@"; }
 130 fail () { _rc=$1; shift; warn "$PROGRAM_NAME:" "$@"; exit $_rc; }
 131 strip_extension () {
 132         printf "%s\n" "$1" | LC_ALL=C sed -r \
 133           -e '$s;^([^/]*/)*([^/]{1,})/{0,1}$;\2;' \
 134           -e '$s;\.([Zz]|lzo|bz|gz|bz2|xz|lzma|7z|lz|rz|lrz)$;;' \
 135           -e '$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)$;;' \
 136           -e '$s;\.(orig)$;;'
 137 }
 138 NL='
 139 '
 140
 141 # default parameters
 142 UNPACK=
 143 STRIP_EXT='strip_extension "$arch"'
 144 RENAME='' # $file $arch $archname
 145 FILTER='false' # $file $arch $archname
 146 DATE='' # $arch $archname
 147 TITLE='echo "$arch"' # $arch $archname
 148 BODY="" # message taken as is
 149 ADDARGS=
 150 CIARGS=
 151 FILTERHEAD= # line separated arguments
 152 TAG=
 153 TAGDEF='printf version/; get_version "$archname"'
 154 COMMAND=
 155 KEEP_TEMPDIR=
 156 # config (file, environment, command line)
 157 if [ -n "${ARCHIVES2GIT_ENV+set}" ]
 158 then
 159         dquotedexpr=$( printf "%s" "$ARCHIVES2GIT_ENV" | LC_ALL=C sed 's/"/\\"/g' )
 160           # double quoting still allows backslash escapes (as well as
 161           # substitutions and expansions)
 162         configfile=
 163         (eval : '${configfile:="'"$dquotedexpr"'"}') &&
 164         eval : '${configfile:="'"$dquotedexpr"'"}'
 165           # use := instead of = as the substitution for more portability (which
 166           # is not that sensible since the + substitution is not implemented
 167           # either when = is not)
 168 elif
 169         gitdir=$(git rev-parse --git-dir) &&
 170         [ -f "$gitdir"/archives2gitrc ]
 171 then
 172         configfile=$gitdir/archives2gitrc
 173 else
 174         configfile=$HOME/.archives2gitrc
 175 fi
 176 [ -r "$configfile" ] && . "$configfile"
 177 GIT_WORK_TREE=${GIT_WORK_TREE-.}
 178   # may be a subdir of the top level
 179 parsing=$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@"); rc=$?
 180 eval "$parsing"
 181 if [ $rc -ne 0 ]
 182 then printf %s "$parsing" | grep '^EOF' >/dev/null && exit 0 || exit $rc
 183 fi
 184 while [ $# -gt 0 ]
 185 do
 186         case $1 in
 187         --unpack)
 188                 shift; UNPACK=$1 ;;
 189         --strip-ext)
 190                 shift; STRIP_EXT=$1 ;;
 191         --repo)
 192                 shift; GIT_WORK_TREE=$1 ;;
 193         --rename)
 194                 shift; RENAME=$1 ;;
 195         --keep-filter)
 196                 shift; FILTER=$1 ;;
 197         --gitadd-arg)
 198                 shift; ADDARGS="$ADDARGS$NL$1" ;;
 199         --date)
 200                 shift; DATE=$1 ;;
 201         --author)
 202                 shift; AUTHOR=$1 ;;
 203         --title)
 204                 shift; TITLE=$1 ;;
 205         --body)
 206                 shift; BODY=$1 ;;
 207         --gitci-arg)
 208                 shift; CIARGS="$CIARGS$NL$1" ;;
 209         --gitfilter)
 210                 shift; FILTERHEAD=$1 ;;
 211         --tag)
 212                 TAG=$TAGDEF ;;
 213         --tagname)
 214                 shift; TAG=$1 ;;
 215         --command)
 216                 shift; COMMAND=$1 ;;
 217         --debug)
 218                 set -x ;;
 219         --keep-tempdir)
 220                 KEEP_TEMPDIR=yes ;;
 221         --helpm)
 222                 help_message
 223                 exit ;;
 224         --man)
 225                 help_message | helpm2pod --man - ARCHIVES2GIT
 226                 exit ;;
 227         --version)
 228                 echo "$PROGRAM_NAME version $PROGRAM_VERSION"
 229                 exit ;;
 230         --?*)
 231                 fail 255 "internal error: invalid option $1" ;;
 232         --)
 233                 shift; break ;;
 234         *)
 235                 break ;;
 236         esac
 237         shift
 238 done
 239 # setup
 240 [ -z "$KEEP_TEMPDIR" ] && trap 'rm -rf "$TMPDIR"' EXIT INT QUIT TERM
 241 TMPDIR=$(mktemp -d) ||
 242 fail 253 "error: cannot create a temporary directory"
 243 WRKDIR=$(pwd)
 244 # git repository check
 245 cd "$GIT_WORK_TREE" &&
 246 { test -d .git || git rev-parse --git-dir >/dev/null; } ||
 247 fail 255 "error: not in a git repository ($(pwd))"
 248 test -z "$(git status --porcelain)" &&
 249 git update-index -q --refresh ||
 250 fail 254 "error: unstaged files or dirty index"
 251
 252 # main
 253 set -e
 254 (set -o pipefail) 2>/dev/null && set -o pipefail
 255 OLDIFS=$IFS
 256 for arch
 257 do
 258         # put the files in the archive dir
 259         cd "$WRKDIR"
 260         if [ -d "$arch" ]
 261         then
 262                 cp -R "$arch" "$TMPDIR"
 263         elif [ -n "$UNPACK" ]
 264         then
 265                 (eval "$UNPACK")
 266         else
 267                 aunpack -X "$TMPDIR" "$arch"
 268         fi
 269         archdir=$TMPDIR
 270         archname=$(eval "$STRIP_EXT") || fail 32 "\$STRIP_EXT failed"
 271         # remove (almost) everything from the repository dir
 272         cd "$GIT_WORK_TREE"
 273         for file in * .*
 274         do
 275                 [ -e "$file" ] || [ -h "$file" ] || continue
 276                 [ x"$file" = x"." ] || [ x"$file" = x".." ] ||
 277                 [ x"$file" = x".git" ] && continue
 278                 (eval "$FILTER") || { git rm -r "$file" || rm -R "$file"; }
 279         done
 280         # move the content of the temp dir to the repository and stage it
 281         for file in "$archdir"/* "$archdir"/.*
 282         do
 283                 [ -e "$file" ] || [ -h "$file" ] || continue
 284                 file=${file#$archdir/}
 285                 [ x"$file" = x"." ] || [ x"$file" = x".." ] && continue
 286                 [ -n "$RENAME" ] && name=$(eval "$RENAME") || name=$file
 287                 if [ -n "$name" ]
 288                 then
 289                         if [ -e ./"$name" ]
 290                         then
 291                           # remove conflicting file (previously kept or renamed
 292                           # to the same name)
 293                                 rm -R ./"$name"
 294                                 warn "Warning: conflicting file removed: \`$name'"
 295                         fi
 296                         mv "$archdir"/"$file" ./"$name"
 297                         IFS=$NL; set -f
 298                         git add $ADDARGS ./"$name"
 299                         set +f; IFS=$OLDIFS
 300                 else
 301                         rm -R "$archdir"/"$file"
 302                 fi
 303                 unset name
 304         done
 305         unset file
 306         # commit, filter and clean up
 307         date=$(eval "$DATE") || fail 32 "\$DATE failed with $?"
 308         author=$(eval "$AUTHOR") || fail 32 "\$AUTHOR failed with $?"
 309         title=$(eval "$TITLE") || fail 32 "\$TITLE failed with $?"
 310         IFS=$NL; set -f
 311         git commit ${AUTHOR:+--author$NL"$author"} ${DATE:+--date$NL"$date"} \
 312           -m "$title"${BODY:+"$NL$NL$BODY"} $CIARGS
 313         if [ -n "$FILTERHEAD" ]
 314         then
 315                 git filter-branch $FILTERHEAD -- HEAD^..HEAD
 316                 rm -R "$(git rev-parse --git-dir)"/refs/original/
 317         fi
 318         set +f; IFS=$OLDIFS
 319         if [ -n "$TAG" ]
 320         then
 321                 tag=$(eval "$TAG") || fail 32 "\$TAG failed with $?"
 322                 tagmessage=$(printf %s "$tag" | LC_ALL=C tail -n +2) || fail 32 "tail failed with $?"
 323                 tagname=$(printf %s "$tag" | LC_ALL=C head -1) || fail 32 "head failed with $?"
 324                 git tag ${tagmessage:+-m "$tagmessage"} "$tagname"
 325         fi
 326         if [ -n "$COMMAND" ]
 327         then
 328                 (eval "$COMMAND")
 329         fi
 330         unset title date author tag tagname tagmessage
 331 done