t/perf/repos/many-files.sh

   1 #!/bin/sh
   2 # Generate test data repository using the given parameters.
   3 # When omitted, we create "gen-many-files-d-w-f.git".
   4 #
   5 # Usage: [-r repo] [-d depth] [-w width] [-f files]
   6 #
   7 # -r repo: path to the new repo to be generated
   8 # -d depth: the depth of sub-directories
   9 # -w width: the number of sub-directories at each level
  10 # -f files: the number of files created in each directory
  11 #
  12 # Note that all files will have the same SHA-1 and each
  13 # directory at a level will have the same SHA-1, so we
  14 # will potentially have a large index, but not a large
  15 # ODB.
  16 #
  17 # Ballast will be created under "ballast/".
  18
  19 EMPTY_BLOB=e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
  20
  21 set -e
  22
  23 # (5, 10, 9) will create 999,999 ballast files.
  24 # (4, 10, 9) will create  99,999 ballast files.
  25 depth=5
  26 width=10
  27 files=9
  28
  29 while test "$#" -ne 0
  30 do
  31     case "$1" in
  32         -r)
  33             shift;
  34             test "$#" -ne 0 || { echo 'error: -r requires an argument' >&2; exit 1; }
  35             repo=$1;
  36             shift ;;
  37         -d)
  38             shift;
  39             test "$#" -ne 0 || { echo 'error: -d requires an argument' >&2; exit 1; }
  40             depth=$1;
  41             shift ;;
  42         -w)
  43             shift;
  44             test "$#" -ne 0 || { echo 'error: -w requires an argument' >&2; exit 1; }
  45             width=$1;
  46             shift ;;
  47         -f)
  48             shift;
  49             test "$#" -ne 0 || { echo 'error: -f requires an argument' >&2; exit 1; }
  50             files=$1;
  51             shift ;;
  52         *)
  53             echo "error: unknown option '$1'" >&2; exit 1 ;;
  54         esac
  55 done
  56
  57 # Inflate the index with thousands of empty files.
  58 # usage: dir depth width files
  59 fill_index() {
  60         awk -v arg_dir=$1 -v arg_depth=$2 -v arg_width=$3 -v arg_files=$4 '
  61                 function make_paths(dir, depth, width, files, f, w) {
  62                         for (f = 1; f <= files; f++) {
  63                                 print dir "/file" f
  64                         }
  65                         if (depth > 0) {
  66                                 for (w = 1; w <= width; w++) {
  67                                         make_paths(dir "/dir" w, depth - 1, width, files)
  68                                 }
  69                         }
  70                 }
  71                 END { make_paths(arg_dir, arg_depth, arg_width, arg_files) }
  72                 ' </dev/null |
  73         sed "s/^/100644 $EMPTY_BLOB     /" |
  74         git update-index --index-info
  75         return 0
  76 }
  77
  78 [ -z "$repo" ] && repo=gen-many-files-$depth.$width.$files.git
  79
  80 mkdir $repo
  81 cd $repo
  82 git init .
  83
  84 # Create an initial commit just to define master.
  85 touch many-files.empty
  86 echo "$depth $width $files" >many-files.params
  87 git add many-files.*
  88 git commit -q -m params
  89
  90 # Create ballast for p0006 based upon the given params and
  91 # inflate the index with thousands of empty files and commit.
  92 git checkout -b p0006-ballast
  93 fill_index "ballast" $depth $width $files
  94 git commit -q -m "ballast"
  95
  96 nr_files=$(git ls-files | wc -l)
  97
  98 # Modify 1 file and commit.
  99 echo "$depth $width $files" >>many-files.params
 100 git add many-files.params
 101 git commit -q -m "ballast plus 1"
 102
 103 # Checkout master to put repo in canonical state (because
 104 # the perf test may need to clone and enable sparse-checkout
 105 # before attempting to checkout a commit with the ballast
 106 # (because it may contain 100K directories and 1M files)).
 107 git checkout master
 108
 109 echo "Repository "$repo" ($depth, $width, $files) created.  Ballast $nr_files."
 110 exit 0