examples/grok-pull.post_update_hook.sh

   1 #!/bin/sh
   2
   3 # use flock(1) from util-linux to avoid seek contention on slow HDDs
   4 # when using multiple `pull_threads' with grok-pull:
   5 # [ "${FLOCKER}" != "$0" ] && exec env FLOCKER="$0" flock "$0" "$0" "$@" || :
   6
   7 # post_update_hook for repos.conf as used by grok-pull, takes a full
   8 # git repo path as it's first and only arg.
   9 full_git_dir="$1"
  10
  11 url_base=http://127.0.0.1:8080/
  12
  13 # same default as other public-inbox-* tools
  14 PI_CONFIG=${PI_CONFIG-~/.public-inbox/config}
  15
  16 # FreeBSD expr(1) only supports BRE, so no '+'
  17 EPOCH2MAIN='\(..*\)/git/[0-9][0-9]*\.git'
  18
  19 # see if it's v2 or v1 based on tree contents, since somebody could
  20 # theoretically name a v1 inbox with a path that looks like a v2 epoch
  21 if git --git-dir="$full_git_dir" ls-tree --name-only HEAD | \
  22         grep -E '^(m|d)$' >/dev/null
  23 then
  24         inbox_fmt=2
  25         inbox_dir=$(expr "$full_git_dir" : "$EPOCH2MAIN")
  26         inbox_name=$(basename "$inbox_dir")
  27         msgmap="$inbox_dir"/msgmap.sqlite3
  28         inbox_lock="$inbox_dir"/inbox.lock
  29 else
  30         inbox_fmt=1
  31         inbox_dir="$full_git_dir"
  32         inbox_name=$(basename "$inbox_dir" .git)
  33         msgmap="$inbox_dir"/public-inbox/msgmap.sqlite3
  34         inbox_lock="$inbox_dir"/ssoma.lock
  35 fi
  36
  37 # run public-inbox-init iff unconfigured
  38 cfg_dir=$(git config -f "$PI_CONFIG" publicinbox."$inbox_name".inboxdir)
  39
  40 # check legacy name for "inboxdir"
  41 case $cfg_dir in
  42 '') cfg_dir=$(git config -f "$PI_CONFIG" publicinbox."$inbox_name".mainrepo) ;;
  43 esac
  44
  45 case $cfg_dir in
  46 '')
  47         remote_git_url=$(git --git-dir="$full_git_dir" config remote.origin.url)
  48         case $remote_git_url in
  49         '')
  50                 echo >&2 "remote.origin.url unset in $full_git_dir/config"
  51                 exit 1
  52                 ;;
  53         esac
  54
  55         case $inbox_fmt in
  56         1)
  57                 remote_inbox_url="$remote_git_url"
  58                 ;;
  59         2)
  60                 remote_inbox_url=$(expr "$remote_git_url" : "$EPOCH2MAIN")
  61                 ;;
  62         esac
  63
  64         config_url="$remote_inbox_url"/_/text/config/raw
  65         remote_config="$inbox_dir"/remote.config.$$
  66         infourls=
  67         trap 'rm -f "$remote_config"' EXIT
  68         if curl --compressed -sSf -v "$config_url" >"$remote_config"
  69         then
  70                 # n.b. inbox_name on the remote may not match our local
  71                 # inbox_name, so we match all addresses in the remote config
  72                 addresses=$(git config -f "$remote_config" -l | \
  73                         sed -ne 's/^publicinbox\..\+\.address=//p')
  74                 case $addresses in
  75                 '')
  76                         echo >&2 'unable to extract address(es) from ' \
  77                                 "$remote_config"
  78                         exit 1
  79                         ;;
  80                 esac
  81                 newsgroups=$(git config -f "$remote_config" -l | \
  82                         sed -ne 's/^publicinbox\..\+\.newsgroup=//p')
  83                 infourls=$(git config -f "$remote_config" -l | \
  84                         sed -ne 's/^publicinbox\..\+.infourl=//p')
  85         else
  86                 newsgroups=
  87                 addresses="$inbox_name@$$.$(hostname).example.com"
  88                 echo >&2 "E: curl $config_url failed"
  89                 echo >&2 "E: using bogus <$addresses> for $inbox_dir"
  90         fi
  91         local_url="$url_base$inbox_name"
  92         public-inbox-init -V$inbox_fmt "$inbox_name" \
  93                 "$inbox_dir" "$local_url" $addresses
  94
  95         if test $? -ne 0
  96         then
  97                 echo >&2 "E: public-inbox-init failed on $inbox_dir"
  98                 exit 1
  99         fi
 100
 101         for ng in $newsgroups
 102         do
 103                 git config -f "$PI_CONFIG" \
 104                         "publicinbox.$inbox_name.newsgroup" "$ng"
 105                 # only one newsgroup per inbox
 106                 break
 107         done
 108         for url in $infourls
 109         do
 110                 git config -f "$PI_CONFIG" \
 111                         "publicinbox.$inbox_name.infourl" "$url"
 112         done
 113         curl -sSfv "$remote_inbox_url"/description >"$inbox_dir"/description
 114         echo "# $inbox_name at $inbox_dir ($addresses) $local_url"
 115         ;;
 116 esac
 117
 118 # only run public-inbox-index if an index exists and has messages,
 119 # since epochs may be cloned out-of-order by grokmirror and we also
 120 # don't know what indexlevel a user wants
 121 if test -f "$msgmap"
 122 then
 123         # We need to use flock(1) (from util-linux) to avoid timeouts
 124         # and SQLite locking problems.
 125         # FreeBSD has a similar lockf(1) utility, but it unlinks by
 126         # default so we use `-k' to keep the lock on the FS.
 127         FLOCK=flock
 128         case $(uname -s) in
 129         FreeBSD) FLOCK='lockf -k' ;;
 130         # ... other OSes here
 131         esac
 132
 133         n=$(echo 'SELECT COUNT(*) FROM msgmap' | \
 134                 $FLOCK $inbox_lock sqlite3 -readonly "$msgmap")
 135         case $n in
 136         0|'')
 137                 : v2 inboxes may be init-ed with an empty msgmap
 138                 ;;
 139         *)
 140                 # if on HDD and limited RAM, add `--sequential-shard'
 141                 # and possibly a large `--batch-size' if you have much
 142                 # memory in public-inbox 1.6.0+
 143                 $EATMYDATA public-inbox-index -v "$inbox_dir"
 144                 ;;
 145         esac
 146 fi