t/t6112-rev-list-filters-objects.sh

   1 #!/bin/sh
   2
   3 test_description='git rev-list using object filtering'
   4
   5 . ./test-lib.sh
   6
   7 # Test the blob:none filter.
   8
   9 test_expect_success 'setup r1' '
  10         echo "{print \$1}" >print_1.awk &&
  11         echo "{print \$2}" >print_2.awk &&
  12
  13         git init r1 &&
  14         for n in 1 2 3 4 5
  15         do
  16                 echo "This is file: $n" > r1/file.$n
  17                 git -C r1 add file.$n
  18                 git -C r1 commit -m "$n"
  19         done
  20 '
  21
  22 test_expect_success 'verify blob:none omits all 5 blobs' '
  23         git -C r1 ls-files -s file.1 file.2 file.3 file.4 file.5 \
  24                 >ls_files_result &&
  25         awk -f print_2.awk ls_files_result |
  26         sort >expected &&
  27
  28         git -C r1 rev-list --quiet --objects --filter-print-omitted \
  29                 --filter=blob:none HEAD >revs &&
  30         awk -f print_1.awk revs |
  31         sed "s/~//" |
  32         sort >observed &&
  33
  34         test_cmp expected observed
  35 '
  36
  37 test_expect_success 'specify blob explicitly prevents filtering' '
  38         file_3=$(git -C r1 ls-files -s file.3 |
  39                  awk -f print_2.awk) &&
  40
  41         file_4=$(git -C r1 ls-files -s file.4 |
  42                  awk -f print_2.awk) &&
  43
  44         git -C r1 rev-list --objects --filter=blob:none HEAD $file_3 >observed &&
  45         grep "$file_3" observed &&
  46         ! grep "$file_4" observed
  47 '
  48
  49 test_expect_success 'verify emitted+omitted == all' '
  50         git -C r1 rev-list --objects HEAD >revs &&
  51         awk -f print_1.awk revs |
  52         sort >expected &&
  53
  54         git -C r1 rev-list --objects --filter-print-omitted --filter=blob:none \
  55                 HEAD >revs &&
  56         awk -f print_1.awk revs |
  57         sed "s/~//" |
  58         sort >observed &&
  59
  60         test_cmp expected observed
  61 '
  62
  63
  64 # Test blob:limit=<n>[kmg] filter.
  65 # We boundary test around the size parameter.  The filter is strictly less than
  66 # the value, so size 500 and 1000 should have the same results, but 1001 should
  67 # filter more.
  68
  69 test_expect_success 'setup r2' '
  70         git init r2 &&
  71         for n in 1000 10000
  72         do
  73                 printf "%"$n"s" X > r2/large.$n
  74                 git -C r2 add large.$n
  75                 git -C r2 commit -m "$n"
  76         done
  77 '
  78
  79 test_expect_success 'verify blob:limit=500 omits all blobs' '
  80         git -C r2 ls-files -s large.1000 large.10000 >ls_files_result &&
  81         awk -f print_2.awk ls_files_result |
  82         sort >expected &&
  83
  84         git -C r2 rev-list --quiet --objects --filter-print-omitted \
  85                 --filter=blob:limit=500 HEAD >revs &&
  86         awk -f print_1.awk revs |
  87         sed "s/~//" |
  88         sort >observed &&
  89
  90         test_cmp expected observed
  91 '
  92
  93 test_expect_success 'verify emitted+omitted == all' '
  94         git -C r2 rev-list --objects HEAD >revs &&
  95         awk -f print_1.awk revs |
  96         sort >expected &&
  97
  98         git -C r2 rev-list --objects --filter-print-omitted \
  99                 --filter=blob:limit=500 HEAD >revs &&
 100         awk -f print_1.awk revs |
 101         sed "s/~//" |
 102         sort >observed &&
 103
 104         test_cmp expected observed
 105 '
 106
 107 test_expect_success 'verify blob:limit=1000' '
 108         git -C r2 ls-files -s large.1000 large.10000 >ls_files_result &&
 109         awk -f print_2.awk ls_files_result |
 110         sort >expected &&
 111
 112         git -C r2 rev-list --quiet --objects --filter-print-omitted \
 113                 --filter=blob:limit=1000 HEAD >revs &&
 114         awk -f print_1.awk revs |
 115         sed "s/~//" |
 116         sort >observed &&
 117
 118         test_cmp expected observed
 119 '
 120
 121 test_expect_success 'verify blob:limit=1001' '
 122         git -C r2 ls-files -s large.10000 >ls_files_result &&
 123         awk -f print_2.awk ls_files_result |
 124         sort >expected &&
 125
 126         git -C r2 rev-list --quiet --objects --filter-print-omitted \
 127                 --filter=blob:limit=1001 HEAD >revs &&
 128         awk -f print_1.awk revs |
 129         sed "s/~//" |
 130         sort >observed &&
 131
 132         test_cmp expected observed
 133 '
 134
 135 test_expect_success 'verify blob:limit=1k' '
 136         git -C r2 ls-files -s large.10000 >ls_files_result &&
 137         awk -f print_2.awk ls_files_result |
 138         sort >expected &&
 139
 140         git -C r2 rev-list --quiet --objects --filter-print-omitted \
 141                 --filter=blob:limit=1k HEAD >revs &&
 142         awk -f print_1.awk revs |
 143         sed "s/~//" |
 144         sort >observed &&
 145
 146         test_cmp expected observed
 147 '
 148
 149 test_expect_success 'verify blob:limit=1m' '
 150         git -C r2 rev-list --quiet --objects --filter-print-omitted \
 151                 --filter=blob:limit=1m HEAD >revs &&
 152         awk -f print_1.awk revs |
 153         sed "s/~//" |
 154         sort >observed &&
 155
 156         test_must_be_empty observed
 157 '
 158
 159 # Test sparse:path=<path> filter.
 160 # !!!!
 161 # NOTE: sparse:path filter support has been dropped for security reasons,
 162 # so the tests have been changed to make sure that using it fails.
 163 # !!!!
 164 # Use a local file containing a sparse-checkout specification to filter
 165 # out blobs not required for the corresponding sparse-checkout.  We do not
 166 # require sparse-checkout to actually be enabled.
 167
 168 test_expect_success 'setup r3' '
 169         git init r3 &&
 170         mkdir r3/dir1 &&
 171         for n in sparse1 sparse2
 172         do
 173                 echo "This is file: $n" > r3/$n
 174                 git -C r3 add $n
 175                 echo "This is file: dir1/$n" > r3/dir1/$n
 176                 git -C r3 add dir1/$n
 177         done &&
 178         git -C r3 commit -m "sparse" &&
 179         echo dir1/ >pattern1 &&
 180         echo sparse1 >pattern2
 181 '
 182
 183 test_expect_success 'verify sparse:path=pattern1 fails' '
 184         test_must_fail git -C r3 rev-list --quiet --objects \
 185                 --filter-print-omitted --filter=sparse:path=../pattern1 HEAD
 186 '
 187
 188 test_expect_success 'verify sparse:path=pattern2 fails' '
 189         test_must_fail git -C r3 rev-list --quiet --objects \
 190                 --filter-print-omitted --filter=sparse:path=../pattern2 HEAD
 191 '
 192
 193 # Test sparse:oid=<oid-ish> filter.
 194 # Use a blob containing a sparse-checkout specification to filter
 195 # out blobs not required for the corresponding sparse-checkout.  We do not
 196 # require sparse-checkout to actually be enabled.
 197
 198 test_expect_success 'setup r3 part 2' '
 199         echo dir1/ >r3/pattern &&
 200         git -C r3 add pattern &&
 201         git -C r3 commit -m "pattern"
 202 '
 203
 204 test_expect_success 'verify sparse:oid=OID omits top-level files' '
 205         git -C r3 ls-files -s pattern sparse1 sparse2 >ls_files_result &&
 206         awk -f print_2.awk ls_files_result |
 207         sort >expected &&
 208
 209         oid=$(git -C r3 ls-files -s pattern | awk -f print_2.awk) &&
 210
 211         git -C r3 rev-list --quiet --objects --filter-print-omitted \
 212                 --filter=sparse:oid=$oid HEAD >revs &&
 213         awk -f print_1.awk revs |
 214         sed "s/~//" |
 215         sort >observed &&
 216
 217         test_cmp expected observed
 218 '
 219
 220 test_expect_success 'verify sparse:oid=oid-ish omits top-level files' '
 221         git -C r3 ls-files -s pattern sparse1 sparse2 >ls_files_result &&
 222         awk -f print_2.awk ls_files_result |
 223         sort >expected &&
 224
 225         git -C r3 rev-list --quiet --objects --filter-print-omitted \
 226                 --filter=sparse:oid=master:pattern HEAD >revs &&
 227         awk -f print_1.awk revs |
 228         sed "s/~//" |
 229         sort >observed &&
 230
 231         test_cmp expected observed
 232 '
 233
 234 test_expect_success 'rev-list W/ --missing=print and --missing=allow-any for trees' '
 235         TREE=$(git -C r3 rev-parse HEAD:dir1) &&
 236
 237         # Create a spare repo because we will be deleting objects from this one.
 238         git clone r3 r3.b &&
 239
 240         rm r3.b/.git/objects/$(echo $TREE | sed "s|^..|&/|") &&
 241
 242         git -C r3.b rev-list --quiet --missing=print --objects HEAD \
 243                 >missing_objs 2>rev_list_err &&
 244         echo "?$TREE" >expected &&
 245         test_cmp expected missing_objs &&
 246
 247         # do not complain when a missing tree cannot be parsed
 248         test_must_be_empty rev_list_err &&
 249
 250         git -C r3.b rev-list --missing=allow-any --objects HEAD \
 251                 >objs 2>rev_list_err &&
 252         ! grep $TREE objs &&
 253         test_must_be_empty rev_list_err
 254 '
 255
 256 # Test tree:0 filter.
 257
 258 test_expect_success 'verify tree:0 includes trees in "filtered" output' '
 259         git -C r3 rev-list --quiet --objects --filter-print-omitted \
 260                 --filter=tree:0 HEAD >revs &&
 261
 262         awk -f print_1.awk revs |
 263         sed s/~// |
 264         xargs -n1 git -C r3 cat-file -t >unsorted_filtered_types &&
 265
 266         sort -u unsorted_filtered_types >filtered_types &&
 267         test_write_lines blob tree >expected &&
 268         test_cmp expected filtered_types
 269 '
 270
 271 # Make sure tree:0 does not iterate through any trees.
 272
 273 test_expect_success 'verify skipping tree iteration when not collecting omits' '
 274         GIT_TRACE=1 git -C r3 rev-list \
 275                 --objects --filter=tree:0 HEAD 2>filter_trace &&
 276         grep "Skipping contents of tree [.][.][.]" filter_trace >actual &&
 277         # One line for each commit traversed.
 278         test_line_count = 2 actual &&
 279
 280         # Make sure no other trees were considered besides the root.
 281         ! grep "Skipping contents of tree [^.]" filter_trace
 282 '
 283
 284 # Test tree:# filters.
 285
 286 expect_has () {
 287         commit=$1 &&
 288         name=$2 &&
 289
 290         hash=$(git -C r3 rev-parse $commit:$name) &&
 291         grep "^$hash $name$" actual
 292 }
 293
 294 test_expect_success 'verify tree:1 includes root trees' '
 295         git -C r3 rev-list --objects --filter=tree:1 HEAD >actual &&
 296
 297         # We should get two root directories and two commits.
 298         expect_has HEAD "" &&
 299         expect_has HEAD~1 ""  &&
 300         test_line_count = 4 actual
 301 '
 302
 303 test_expect_success 'verify tree:2 includes root trees and immediate children' '
 304         git -C r3 rev-list --objects --filter=tree:2 HEAD >actual &&
 305
 306         expect_has HEAD "" &&
 307         expect_has HEAD~1 "" &&
 308         expect_has HEAD dir1 &&
 309         expect_has HEAD pattern &&
 310         expect_has HEAD sparse1 &&
 311         expect_has HEAD sparse2 &&
 312
 313         # There are also 2 commit objects
 314         test_line_count = 8 actual
 315 '
 316
 317 test_expect_success 'verify tree:3 includes everything expected' '
 318         git -C r3 rev-list --objects --filter=tree:3 HEAD >actual &&
 319
 320         expect_has HEAD "" &&
 321         expect_has HEAD~1 "" &&
 322         expect_has HEAD dir1 &&
 323         expect_has HEAD dir1/sparse1 &&
 324         expect_has HEAD dir1/sparse2 &&
 325         expect_has HEAD pattern &&
 326         expect_has HEAD sparse1 &&
 327         expect_has HEAD sparse2 &&
 328
 329         # There are also 2 commit objects
 330         test_line_count = 10 actual
 331 '
 332
 333 # Test provisional omit collection logic with a repo that has objects appearing
 334 # at multiple depths - first deeper than the filter's threshold, then shallow.
 335
 336 test_expect_success 'setup r4' '
 337         git init r4 &&
 338
 339         echo foo > r4/foo &&
 340         mkdir r4/subdir &&
 341         echo bar > r4/subdir/bar &&
 342
 343         mkdir r4/filt &&
 344         cp -r r4/foo r4/subdir r4/filt &&
 345
 346         git -C r4 add foo subdir filt &&
 347         git -C r4 commit -m "commit msg"
 348 '
 349
 350 expect_has_with_different_name () {
 351         repo=$1 &&
 352         name=$2 &&
 353
 354         hash=$(git -C $repo rev-parse HEAD:$name) &&
 355         ! grep "^$hash $name$" actual &&
 356         grep "^$hash " actual &&
 357         ! grep "~$hash" actual
 358 }
 359
 360 test_expect_success 'test tree:# filter provisional omit for blob and tree' '
 361         git -C r4 rev-list --objects --filter-print-omitted --filter=tree:2 \
 362                 HEAD >actual &&
 363         expect_has_with_different_name r4 filt/foo &&
 364         expect_has_with_different_name r4 filt/subdir
 365 '
 366
 367 test_expect_success 'verify skipping tree iteration when collecting omits' '
 368         GIT_TRACE=1 git -C r4 rev-list --filter-print-omitted \
 369                 --objects --filter=tree:0 HEAD 2>filter_trace &&
 370         grep "^Skipping contents of tree " filter_trace >actual &&
 371
 372         echo "Skipping contents of tree subdir/..." >expect &&
 373         test_cmp expect actual
 374 '
 375
 376 # Test tree:<depth> where a tree is iterated to twice - once where a subentry is
 377 # too deep to be included, and again where the blob inside it is shallow enough
 378 # to be included. This makes sure we don't use LOFR_MARK_SEEN incorrectly (we
 379 # can't use it because a tree can be iterated over again at a lower depth).
 380
 381 test_expect_success 'tree:<depth> where we iterate over tree at two levels' '
 382         git init r5 &&
 383
 384         mkdir -p r5/a/subdir/b &&
 385         echo foo > r5/a/subdir/b/foo &&
 386
 387         mkdir -p r5/subdir/b &&
 388         echo foo > r5/subdir/b/foo &&
 389
 390         git -C r5 add a subdir &&
 391         git -C r5 commit -m "commit msg" &&
 392
 393         git -C r5 rev-list --objects --filter=tree:4 HEAD >actual &&
 394         expect_has_with_different_name r5 a/subdir/b/foo
 395 '
 396
 397 test_expect_success 'tree:<depth> which filters out blob but given as arg' '
 398         blob_hash=$(git -C r4 rev-parse HEAD:subdir/bar) &&
 399
 400         git -C r4 rev-list --objects --filter=tree:1 HEAD $blob_hash >actual &&
 401         grep ^$blob_hash actual
 402 '
 403
 404 # Delete some loose objects and use rev-list, but WITHOUT any filtering.
 405 # This models previously omitted objects that we did not receive.
 406
 407 test_expect_success 'rev-list W/ --missing=print' '
 408         git -C r1 ls-files -s file.1 file.2 file.3 file.4 file.5 \
 409                 >ls_files_result &&
 410         awk -f print_2.awk ls_files_result |
 411         sort >expected &&
 412
 413         for id in `cat expected | sed "s|..|&/|"`
 414         do
 415                 rm r1/.git/objects/$id
 416         done &&
 417
 418         git -C r1 rev-list --quiet --missing=print --objects HEAD >revs &&
 419         awk -f print_1.awk revs |
 420         sed "s/?//" |
 421         sort >observed &&
 422
 423         test_cmp expected observed
 424 '
 425
 426 test_expect_success 'rev-list W/O --missing fails' '
 427         test_must_fail git -C r1 rev-list --quiet --objects HEAD
 428 '
 429
 430 test_expect_success 'rev-list W/ missing=allow-any' '
 431         git -C r1 rev-list --quiet --missing=allow-any --objects HEAD
 432 '
 433
 434 # Test expansion of filter specs.
 435
 436 test_expect_success 'expand blob limit in protocol' '
 437         git -C r2 config --local uploadpack.allowfilter 1 &&
 438         GIT_TRACE_PACKET="$(pwd)/trace" git -c protocol.version=2 clone \
 439                 --filter=blob:limit=1k "file://$(pwd)/r2" limit &&
 440         ! grep "blob:limit=1k" trace &&
 441         grep "blob:limit=1024" trace
 442 '
 443
 444 test_expect_success 'expand tree depth limit in protocol' '
 445         GIT_TRACE_PACKET="$(pwd)/tree_trace" git -c protocol.version=2 clone \
 446                 --filter=tree:0k "file://$(pwd)/r2" tree &&
 447         ! grep "tree:0k" tree_trace &&
 448         grep "tree:0" tree_trace
 449 '
 450
 451 test_done