t/t7704-repack-cruft.sh

   1 #!/bin/sh
   2
   3 test_description='git repack works correctly'
   4
   5 . ./test-lib.sh
   6
   7 objdir=.git/objects
   8 packdir=$objdir/pack
   9
  10 test_expect_success '--expire-to stores pruned objects (now)' '
  11         git init expire-to-now &&
  12         (
  13                 cd expire-to-now &&
  14
  15                 git branch -M main &&
  16
  17                 test_commit base &&
  18
  19                 git checkout -b cruft &&
  20                 test_commit --no-tag cruft &&
  21
  22                 git rev-list --objects --no-object-names main..cruft >moved.raw &&
  23                 sort moved.raw >moved.want &&
  24
  25                 git rev-list --all --objects --no-object-names >expect.raw &&
  26                 sort expect.raw >expect &&
  27
  28                 git checkout main &&
  29                 git branch -D cruft &&
  30                 git reflog expire --all --expire=all &&
  31
  32                 git init --bare expired.git &&
  33                 git repack -d \
  34                         --cruft --cruft-expiration="now" \
  35                         --expire-to="expired.git/objects/pack/pack" &&
  36
  37                 expired="$(ls expired.git/objects/pack/pack-*.idx)" &&
  38                 test_path_is_file "${expired%.idx}.mtimes" &&
  39
  40                 # Since the `--cruft-expiration` is "now", the effective
  41                 # behavior is to move _all_ unreachable objects out to
  42                 # the location in `--expire-to`.
  43                 git show-index <$expired >expired.raw &&
  44                 cut -d" " -f2 expired.raw | sort >expired.objects &&
  45                 git rev-list --all --objects --no-object-names \
  46                         >remaining.objects &&
  47
  48                 # ...in other words, the combined contents of this
  49                 # repository and expired.git should be the same as the
  50                 # set of objects we started with.
  51                 sort expired.objects remaining.objects >actual &&
  52                 test_cmp expect actual &&
  53
  54                 # The "moved" objects (i.e., those in expired.git)
  55                 # should be the same as the cruft objects which were
  56                 # expired in the previous step.
  57                 test_cmp moved.want expired.objects
  58         )
  59 '
  60
  61 test_expect_success '--expire-to stores pruned objects (5.minutes.ago)' '
  62         git init expire-to-5.minutes.ago &&
  63         (
  64                 cd expire-to-5.minutes.ago &&
  65
  66                 git branch -M main &&
  67
  68                 test_commit base &&
  69
  70                 # Create two classes of unreachable objects, one which
  71                 # is older than 5 minutes (stale), and another which is
  72                 # newer (recent).
  73                 for kind in stale recent
  74                 do
  75                         git checkout -b $kind main &&
  76                         test_commit --no-tag $kind || return 1
  77                 done &&
  78
  79                 git rev-list --objects --no-object-names main..stale >in &&
  80                 stale="$(git pack-objects $objdir/pack/pack <in)" &&
  81                 mtime="$(test-tool chmtime --get =-600 $objdir/pack/pack-$stale.pack)" &&
  82
  83                 # expect holds the set of objects we expect to find in
  84                 # this repository after repacking
  85                 git rev-list --objects --no-object-names recent >expect.raw &&
  86                 sort expect.raw >expect &&
  87
  88                 # moved.want holds the set of objects we expect to find
  89                 # in expired.git
  90                 git rev-list --objects --no-object-names main..stale >out &&
  91                 sort out >moved.want &&
  92
  93                 git checkout main &&
  94                 git branch -D stale recent &&
  95                 git reflog expire --all --expire=all &&
  96                 git prune-packed &&
  97
  98                 git init --bare expired.git &&
  99                 git repack -d \
 100                         --cruft --cruft-expiration=5.minutes.ago \
 101                         --expire-to="expired.git/objects/pack/pack" &&
 102
 103                 # Some of the remaining objects in this repository are
 104                 # unreachable, so use `cat-file --batch-all-objects`
 105                 # instead of `rev-list` to get their names
 106                 git cat-file --batch-all-objects --batch-check="%(objectname)" \
 107                         >remaining.objects &&
 108                 sort remaining.objects >actual &&
 109                 test_cmp expect actual &&
 110
 111                 (
 112                         cd expired.git &&
 113
 114                         expired="$(ls objects/pack/pack-*.mtimes)" &&
 115                         test-tool pack-mtimes $(basename $expired) >out &&
 116                         cut -d" " -f1 out | sort >../moved.got &&
 117
 118                         # Ensure that there are as many objects with the
 119                         # expected mtime as were moved to expired.git.
 120                         #
 121                         # In other words, ensure that the recorded
 122                         # mtimes of any moved objects was written
 123                         # correctly.
 124                         grep " $mtime$" out >matching &&
 125                         test_line_count = $(wc -l <../moved.want) matching
 126                 ) &&
 127                 test_cmp moved.want moved.got
 128         )
 129 '
 130
 131 generate_random_blob() {
 132         test-tool genrandom "$@" >blob &&
 133         git hash-object -w -t blob blob &&
 134         rm blob
 135 }
 136
 137 pack_random_blob () {
 138         generate_random_blob "$@" &&
 139         git repack -d -q >/dev/null
 140 }
 141
 142 generate_cruft_pack () {
 143         pack_random_blob "$@" >/dev/null &&
 144
 145         ls $packdir/pack-*.pack | xargs -n 1 basename >in &&
 146         pack="$(git pack-objects --cruft $packdir/pack <in)" &&
 147         git prune-packed &&
 148
 149         echo "$packdir/pack-$pack.mtimes"
 150 }
 151
 152 test_expect_success '--max-cruft-size creates new packs when above threshold' '
 153         git init max-cruft-size-large &&
 154         (
 155                 cd max-cruft-size-large &&
 156                 test_commit base &&
 157
 158                 foo="$(pack_random_blob foo $((1*1024*1024)))" &&
 159                 git repack --cruft -d &&
 160                 cruft_foo="$(ls $packdir/pack-*.mtimes)" &&
 161
 162                 bar="$(pack_random_blob bar $((1*1024*1024)))" &&
 163                 git repack --cruft -d --max-cruft-size=1M &&
 164                 cruft_bar="$(ls $packdir/pack-*.mtimes | grep -v $cruft_foo)" &&
 165
 166                 test-tool pack-mtimes $(basename "$cruft_foo") >foo.objects &&
 167                 test-tool pack-mtimes $(basename "$cruft_bar") >bar.objects &&
 168
 169                 grep "^$foo" foo.objects &&
 170                 test_line_count = 1 foo.objects &&
 171                 grep "^$bar" bar.objects &&
 172                 test_line_count = 1 bar.objects
 173         )
 174 '
 175
 176 test_expect_success '--max-cruft-size combines existing packs when below threshold' '
 177         git init max-cruft-size-small &&
 178         (
 179                 cd max-cruft-size-small &&
 180                 test_commit base &&
 181
 182                 foo="$(pack_random_blob foo $((1*1024*1024)))" &&
 183                 git repack --cruft -d &&
 184
 185                 bar="$(pack_random_blob bar $((1*1024*1024)))" &&
 186                 git repack --cruft -d --max-cruft-size=10M &&
 187
 188                 cruft=$(ls $packdir/pack-*.mtimes) &&
 189                 test-tool pack-mtimes $(basename "$cruft") >cruft.objects &&
 190
 191                 grep "^$foo" cruft.objects &&
 192                 grep "^$bar" cruft.objects &&
 193                 test_line_count = 2 cruft.objects
 194         )
 195 '
 196
 197 test_expect_success '--max-cruft-size combines smaller packs first' '
 198         git init max-cruft-size-consume-small &&
 199         (
 200                 cd max-cruft-size-consume-small &&
 201
 202                 test_commit base &&
 203                 git repack -ad &&
 204
 205                 cruft_foo="$(generate_cruft_pack foo 524288)" &&    # 0.5 MiB
 206                 cruft_bar="$(generate_cruft_pack bar 524288)" &&    # 0.5 MiB
 207                 cruft_baz="$(generate_cruft_pack baz 1048576)" &&   # 1.0 MiB
 208                 cruft_quux="$(generate_cruft_pack quux 1572864)" && # 1.5 MiB
 209
 210                 test-tool pack-mtimes "$(basename $cruft_foo)" >expect.raw &&
 211                 test-tool pack-mtimes "$(basename $cruft_bar)" >>expect.raw &&
 212                 sort expect.raw >expect.objects &&
 213
 214                 # repacking with `--max-cruft-size=2M` should combine
 215                 # both 0.5 MiB packs together, instead of, say, one of
 216                 # the 0.5 MiB packs with the 1.0 MiB pack
 217                 ls $packdir/pack-*.mtimes | sort >cruft.before &&
 218                 git repack -d --cruft --max-cruft-size=2M &&
 219                 ls $packdir/pack-*.mtimes | sort >cruft.after &&
 220
 221                 comm -13 cruft.before cruft.after >cruft.new &&
 222                 comm -23 cruft.before cruft.after >cruft.removed &&
 223
 224                 test_line_count = 1 cruft.new &&
 225                 test_line_count = 2 cruft.removed &&
 226
 227                 # the two smaller packs should be rolled up first
 228                 printf "%s\n" $cruft_foo $cruft_bar | sort >expect.removed &&
 229                 test_cmp expect.removed cruft.removed &&
 230
 231                 # ...and contain the set of objects rolled up
 232                 test-tool pack-mtimes "$(basename $(cat cruft.new))" >actual.raw &&
 233                 sort actual.raw >actual.objects &&
 234
 235                 test_cmp expect.objects actual.objects
 236         )
 237 '
 238
 239 test_expect_success 'setup --max-cruft-size with freshened objects' '
 240         git init max-cruft-size-freshen &&
 241         (
 242                 cd max-cruft-size-freshen &&
 243
 244                 test_commit base &&
 245                 git repack -ad &&
 246
 247                 foo="$(generate_random_blob foo 64)" &&
 248                 test-tool chmtime --get -10000 \
 249                         "$objdir/$(test_oid_to_path "$foo")" >foo.mtime &&
 250
 251                 git repack --cruft -d &&
 252
 253                 cruft="$(ls $packdir/pack-*.mtimes)" &&
 254                 test-tool pack-mtimes "$(basename $cruft)" >actual &&
 255                 echo "$foo $(cat foo.mtime)" >expect &&
 256                 test_cmp expect actual
 257         )
 258 '
 259
 260 test_expect_success '--max-cruft-size with freshened objects (loose)' '
 261         (
 262                 cd max-cruft-size-freshen &&
 263
 264                 # regenerate the object, setting its mtime to be more recent
 265                 foo="$(generate_random_blob foo 64)" &&
 266                 test-tool chmtime --get -100 \
 267                         "$objdir/$(test_oid_to_path "$foo")" >foo.mtime &&
 268
 269                 git repack --cruft -d &&
 270
 271                 cruft="$(ls $packdir/pack-*.mtimes)" &&
 272                 test-tool pack-mtimes "$(basename $cruft)" >actual &&
 273                 echo "$foo $(cat foo.mtime)" >expect &&
 274                 test_cmp expect actual
 275         )
 276 '
 277
 278 test_expect_success '--max-cruft-size with freshened objects (packed)' '
 279         (
 280                 cd max-cruft-size-freshen &&
 281
 282                 # regenerate the object and store it in a packfile,
 283                 # setting its mtime to be more recent
 284                 #
 285                 # store it alongside another cruft object so that we
 286                 # do not create an identical copy of the existing
 287                 # cruft pack (which contains $foo).
 288                 foo="$(generate_random_blob foo 64)" &&
 289                 bar="$(generate_random_blob bar 64)" &&
 290                 foo_pack="$(printf "%s\n" $foo $bar | git pack-objects $packdir/pack)" &&
 291                 git prune-packed &&
 292
 293                 test-tool chmtime --get -10 \
 294                         "$packdir/pack-$foo_pack.pack" >foo.mtime &&
 295
 296                 git repack --cruft -d &&
 297
 298                 cruft="$(ls $packdir/pack-*.mtimes)" &&
 299                 test-tool pack-mtimes "$(basename $cruft)" >actual &&
 300                 echo "$foo $(cat foo.mtime)" >expect.raw &&
 301                 echo "$bar $(cat foo.mtime)" >>expect.raw &&
 302                 sort expect.raw >expect &&
 303                 test_cmp expect actual
 304         )
 305 '
 306
 307 test_expect_success '--max-cruft-size with pruning' '
 308         git init max-cruft-size-prune &&
 309         (
 310                 cd max-cruft-size-prune &&
 311
 312                 test_commit base &&
 313                 foo="$(generate_random_blob foo $((1024*1024)))" &&
 314                 bar="$(generate_random_blob bar $((1024*1024)))" &&
 315                 baz="$(generate_random_blob baz $((1024*1024)))" &&
 316
 317                 test-tool chmtime -10000 "$objdir/$(test_oid_to_path "$foo")" &&
 318
 319                 git repack -d --cruft --max-cruft-size=1M &&
 320
 321                 # backdate the mtimes of all cruft packs to validate
 322                 # that they were rewritten as a result of pruning
 323                 ls $packdir/pack-*.mtimes | sort >cruft.before &&
 324                 for cruft in $(cat cruft.before)
 325                 do
 326                         mtime="$(test-tool chmtime --get -10000 "$cruft")" &&
 327                         echo $cruft $mtime >>mtimes || return 1
 328                 done &&
 329
 330                 # repack (and prune) with a --max-cruft-size to ensure
 331                 # that we appropriately split the resulting set of packs
 332                 git repack -d --cruft --max-cruft-size=1M \
 333                         --cruft-expiration=10.seconds.ago &&
 334                 ls $packdir/pack-*.mtimes | sort >cruft.after &&
 335
 336                 for cruft in $(cat cruft.after)
 337                 do
 338                         old_mtime="$(grep $cruft mtimes | cut -d" " -f2)" &&
 339                         new_mtime="$(test-tool chmtime --get $cruft)" &&
 340                         test $old_mtime -lt $new_mtime || return 1
 341                 done &&
 342
 343                 test_line_count = 3 cruft.before &&
 344                 test_line_count = 2 cruft.after &&
 345                 test_must_fail git cat-file -e $foo &&
 346                 git cat-file -e $bar &&
 347                 git cat-file -e $baz
 348         )
 349 '
 350
 351 test_expect_success '--max-cruft-size ignores non-local packs' '
 352         repo="max-cruft-size-non-local" &&
 353         git init $repo &&
 354         (
 355                 cd $repo &&
 356                 test_commit base &&
 357                 generate_random_blob foo 64 &&
 358                 git repack --cruft -d
 359         ) &&
 360
 361         git clone --reference=$repo $repo $repo-alt &&
 362         (
 363                 cd $repo-alt &&
 364
 365                 test_commit other &&
 366                 generate_random_blob bar 64 &&
 367
 368                 # ensure that we do not attempt to pick up packs from
 369                 # the non-alternated repository, which would result in a
 370                 # crash
 371                 git repack --cruft --max-cruft-size=1M -d
 372         )
 373 '
 374
 375 test_expect_success 'reachable packs are preferred over cruft ones' '
 376         repo="cruft-preferred-packs" &&
 377         git init "$repo" &&
 378         (
 379                 cd "$repo" &&
 380
 381                 # This test needs to exercise careful control over when a MIDX
 382                 # is and is not written. Unset the corresponding TEST variable
 383                 # accordingly.
 384                 sane_unset GIT_TEST_MULTI_PACK_INDEX &&
 385
 386                 test_commit base &&
 387                 test_commit --no-tag cruft &&
 388
 389                 non_cruft="$(echo base | git pack-objects --revs $packdir/pack)" &&
 390                 # Write a cruft pack which both (a) sorts ahead of the non-cruft
 391                 # pack in lexical order, and (b) has an older mtime to appease
 392                 # the MIDX preferred pack selection routine.
 393                 cruft="$(echo pack-$non_cruft.pack | git pack-objects --cruft $packdir/pack-A)" &&
 394                 test-tool chmtime -1000 $packdir/pack-A-$cruft.pack &&
 395
 396                 test_commit other &&
 397                 git repack -d &&
 398
 399                 git repack --geometric 2 -d --write-midx --write-bitmap-index &&
 400
 401                 # After repacking, there are two packs left: one reachable one
 402                 # (which is the result of combining both of the existing two
 403                 # non-cruft packs), and one cruft pack.
 404                 find .git/objects/pack -type f -name "*.pack" >packs &&
 405                 test_line_count = 2 packs &&
 406
 407                 # Make sure that the pack we just wrote is marked as preferred,
 408                 # not the cruft one.
 409                 pack="$(test-tool read-midx --preferred-pack $objdir)" &&
 410                 test_path_is_missing "$packdir/$(basename "$pack" ".idx").mtimes"
 411         )
 412 '
 413
 414 test_done