2009-01-30 09:33:00 +01:00
|
|
|
#!/bin/sh
|
|
|
|
|
|
2010-09-07 03:47:07 +02:00
|
|
|
test_description='git fsck random collection of tests
|
|
|
|
|
|
|
|
|
|
* (HEAD) B
|
2020-11-19 00:44:21 +01:00
|
|
|
* (main) A
|
2010-09-07 03:47:07 +02:00
|
|
|
'
|
2009-01-30 09:33:00 +01:00
|
|
|
|
|
|
|
|
. ./test-lib.sh
|
|
|
|
|
|
|
|
|
|
test_expect_success setup '
|
2010-09-07 03:47:07 +02:00
|
|
|
git config gc.auto 0 &&
|
2010-05-26 23:50:34 +02:00
|
|
|
git config i18n.commitencoding ISO-8859-1 &&
|
2009-01-30 09:33:00 +01:00
|
|
|
test_commit A fileA one &&
|
2010-05-26 23:50:34 +02:00
|
|
|
git config --unset i18n.commitencoding &&
|
2009-01-30 09:33:00 +01:00
|
|
|
git checkout HEAD^0 &&
|
|
|
|
|
test_commit B fileB two &&
|
|
|
|
|
git tag -d A B &&
|
tests: use 'test_must_be_empty' instead of 'test_cmp <empty> <out>'
Using 'test_must_be_empty' is shorter and more idiomatic than
>empty &&
test_cmp empty out
as it saves the creation of an empty file. Furthermore, sometimes the
expected empty file doesn't have such a descriptive name like 'empty',
and its creation is far away from the place where it's finally used
for comparison (e.g. in 't7600-merge.sh', where two expected empty
files are created in the 'setup' test, but are used only about 500
lines later).
These cases were found by instrumenting 'test_cmp' to error out the
test script when it's used to compare empty files, and then converted
manually.
Note that even after this patch there still remain a lot of cases
where we use 'test_cmp' to check empty files:
- Sometimes the expected output is not hard-coded in the test, but
'test_cmp' is used to ensure that two similar git commands produce
the same output, and that output happens to be empty, e.g. the
test 'submodule update --merge - ignores --merge for new
submodules' in 't7406-submodule-update.sh'.
- Repetitive common tasks, including preparing the expected results
and running 'test_cmp', are often extracted into a helper
function, and some of this helper's callsites expect no output.
- For the same reason as above, the whole 'test_expect_success'
block is within a helper function, e.g. in 't3070-wildmatch.sh'.
- Or 'test_cmp' is invoked in a loop, e.g. the test 'cvs update
(-p)' in 't9400-git-cvsserver-server.sh'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-19 23:57:25 +02:00
|
|
|
git reflog expire --expire=now --all
|
2009-01-30 09:33:00 +01:00
|
|
|
'
|
|
|
|
|
|
2009-01-30 09:50:54 +01:00
|
|
|
test_expect_success 'loose objects borrowed from alternate are not missing' '
|
|
|
|
|
mkdir another &&
|
|
|
|
|
(
|
|
|
|
|
cd another &&
|
|
|
|
|
git init &&
|
|
|
|
|
echo ../../../.git/objects >.git/objects/info/alternates &&
|
|
|
|
|
test_commit C fileC one &&
|
2012-02-28 23:55:39 +01:00
|
|
|
git fsck --no-dangling >../actual 2>&1
|
2010-09-07 03:47:07 +02:00
|
|
|
) &&
|
tests: use 'test_must_be_empty' instead of 'test_cmp <empty> <out>'
Using 'test_must_be_empty' is shorter and more idiomatic than
>empty &&
test_cmp empty out
as it saves the creation of an empty file. Furthermore, sometimes the
expected empty file doesn't have such a descriptive name like 'empty',
and its creation is far away from the place where it's finally used
for comparison (e.g. in 't7600-merge.sh', where two expected empty
files are created in the 'setup' test, but are used only about 500
lines later).
These cases were found by instrumenting 'test_cmp' to error out the
test script when it's used to compare empty files, and then converted
manually.
Note that even after this patch there still remain a lot of cases
where we use 'test_cmp' to check empty files:
- Sometimes the expected output is not hard-coded in the test, but
'test_cmp' is used to ensure that two similar git commands produce
the same output, and that output happens to be empty, e.g. the
test 'submodule update --merge - ignores --merge for new
submodules' in 't7406-submodule-update.sh'.
- Repetitive common tasks, including preparing the expected results
and running 'test_cmp', are often extracted into a helper
function, and some of this helper's callsites expect no output.
- For the same reason as above, the whole 'test_expect_success'
block is within a helper function, e.g. in 't3070-wildmatch.sh'.
- Or 'test_cmp' is invoked in a loop, e.g. the test 'cvs update
(-p)' in 't9400-git-cvsserver-server.sh'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-19 23:57:25 +02:00
|
|
|
test_must_be_empty actual
|
2009-01-30 09:50:54 +01:00
|
|
|
'
|
|
|
|
|
|
2010-09-07 03:47:07 +02:00
|
|
|
test_expect_success 'HEAD is part of refs, valid objects appear valid' '
|
|
|
|
|
git fsck >actual 2>&1 &&
|
tests: use 'test_must_be_empty' instead of 'test_cmp <empty> <out>'
Using 'test_must_be_empty' is shorter and more idiomatic than
>empty &&
test_cmp empty out
as it saves the creation of an empty file. Furthermore, sometimes the
expected empty file doesn't have such a descriptive name like 'empty',
and its creation is far away from the place where it's finally used
for comparison (e.g. in 't7600-merge.sh', where two expected empty
files are created in the 'setup' test, but are used only about 500
lines later).
These cases were found by instrumenting 'test_cmp' to error out the
test script when it's used to compare empty files, and then converted
manually.
Note that even after this patch there still remain a lot of cases
where we use 'test_cmp' to check empty files:
- Sometimes the expected output is not hard-coded in the test, but
'test_cmp' is used to ensure that two similar git commands produce
the same output, and that output happens to be empty, e.g. the
test 'submodule update --merge - ignores --merge for new
submodules' in 't7406-submodule-update.sh'.
- Repetitive common tasks, including preparing the expected results
and running 'test_cmp', are often extracted into a helper
function, and some of this helper's callsites expect no output.
- For the same reason as above, the whole 'test_expect_success'
block is within a helper function, e.g. in 't3070-wildmatch.sh'.
- Or 'test_cmp' is invoked in a loop, e.g. the test 'cvs update
(-p)' in 't9400-git-cvsserver-server.sh'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-19 23:57:25 +02:00
|
|
|
test_must_be_empty actual
|
2010-05-26 23:50:34 +02:00
|
|
|
'
|
|
|
|
|
|
2009-02-19 12:13:39 +01:00
|
|
|
# Corruption tests follow. Make sure to remove all traces of the
|
|
|
|
|
# specific corruption you test afterwards, lest a later test trip over
|
|
|
|
|
# it.
|
|
|
|
|
|
2021-02-10 19:01:29 +01:00
|
|
|
sha1_file () {
|
|
|
|
|
git rev-parse --git-path objects/$(test_oid_to_path "$1")
|
|
|
|
|
}
|
2010-09-07 03:47:07 +02:00
|
|
|
|
2021-02-10 19:01:29 +01:00
|
|
|
remove_object () {
|
|
|
|
|
rm "$(sha1_file "$1")"
|
|
|
|
|
}
|
2010-09-07 03:47:07 +02:00
|
|
|
|
2021-10-01 11:16:38 +02:00
|
|
|
test_expect_success 'object with hash mismatch' '
|
|
|
|
|
git init --bare hash-mismatch &&
|
|
|
|
|
(
|
|
|
|
|
cd hash-mismatch &&
|
2010-09-07 03:47:07 +02:00
|
|
|
|
2021-10-01 11:16:38 +02:00
|
|
|
oid=$(echo blob | git hash-object -w --stdin) &&
|
|
|
|
|
old=$(test_oid_to_path "$oid") &&
|
|
|
|
|
new=$(dirname $old)/$(test_oid ff_2) &&
|
|
|
|
|
oid="$(dirname $new)$(basename $new)" &&
|
|
|
|
|
|
|
|
|
|
mv objects/$old objects/$new &&
|
|
|
|
|
git update-index --add --cacheinfo 100644 $oid foo &&
|
|
|
|
|
tree=$(git write-tree) &&
|
|
|
|
|
cmt=$(echo bogus | git commit-tree $tree) &&
|
|
|
|
|
git update-ref refs/heads/bogus $cmt &&
|
|
|
|
|
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
grep "$oid.*corrupt" out
|
|
|
|
|
)
|
2009-02-19 12:13:39 +01:00
|
|
|
'
|
|
|
|
|
|
2021-10-01 11:16:39 +02:00
|
|
|
test_expect_success 'object with hash and type mismatch' '
|
|
|
|
|
git init --bare hash-type-mismatch &&
|
|
|
|
|
(
|
|
|
|
|
cd hash-type-mismatch &&
|
|
|
|
|
|
|
|
|
|
oid=$(echo blob | git hash-object -w --stdin -t garbage --literally) &&
|
|
|
|
|
old=$(test_oid_to_path "$oid") &&
|
|
|
|
|
new=$(dirname $old)/$(test_oid ff_2) &&
|
|
|
|
|
oid="$(dirname $new)$(basename $new)" &&
|
|
|
|
|
|
|
|
|
|
mv objects/$old objects/$new &&
|
|
|
|
|
git update-index --add --cacheinfo 100644 $oid foo &&
|
|
|
|
|
tree=$(git write-tree) &&
|
|
|
|
|
cmt=$(echo bogus | git commit-tree $tree) &&
|
|
|
|
|
git update-ref refs/heads/bogus $cmt &&
|
|
|
|
|
|
|
|
|
|
cat >expect <<-\EOF &&
|
|
|
|
|
fatal: invalid object type
|
|
|
|
|
EOF
|
|
|
|
|
test_must_fail git fsck 2>actual &&
|
|
|
|
|
test_cmp expect actual
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2021-10-01 11:16:40 +02:00
|
|
|
test_expect_success POSIXPERM 'zlib corrupt loose object output ' '
|
|
|
|
|
git init --bare corrupt-loose-output &&
|
|
|
|
|
(
|
|
|
|
|
cd corrupt-loose-output &&
|
|
|
|
|
oid=$(git hash-object -w --stdin --literally </dev/null) &&
|
|
|
|
|
oidf=objects/$(test_oid_to_path "$oid") &&
|
|
|
|
|
chmod 755 $oidf &&
|
|
|
|
|
echo extra garbage >>$oidf &&
|
|
|
|
|
|
|
|
|
|
cat >expect.error <<-EOF &&
|
|
|
|
|
error: garbage at end of loose object '\''$oid'\''
|
|
|
|
|
error: unable to unpack contents of ./$oidf
|
|
|
|
|
error: $oid: object corrupt or missing: ./$oidf
|
|
|
|
|
EOF
|
|
|
|
|
test_must_fail git fsck 2>actual &&
|
|
|
|
|
grep ^error: actual >error &&
|
|
|
|
|
test_cmp expect.error error
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2009-02-19 12:13:39 +01:00
|
|
|
test_expect_success 'branch pointing to non-commit' '
|
2010-09-07 03:47:07 +02:00
|
|
|
git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/invalid" &&
|
2015-09-23 22:46:39 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "not a commit" out
|
2009-02-19 12:13:39 +01:00
|
|
|
'
|
|
|
|
|
|
2015-09-23 22:46:39 +02:00
|
|
|
test_expect_success 'HEAD link pointing at a funny object' '
|
|
|
|
|
test_when_finished "mv .git/SAVED_HEAD .git/HEAD" &&
|
|
|
|
|
mv .git/HEAD .git/SAVED_HEAD &&
|
2019-06-29 00:59:21 +02:00
|
|
|
echo $ZERO_OID >.git/HEAD &&
|
2015-09-23 22:46:39 +02:00
|
|
|
# avoid corrupt/broken HEAD from interfering with repo discovery
|
|
|
|
|
test_must_fail env GIT_DIR=.git git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "detached HEAD points" out
|
2015-09-23 22:46:39 +02:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'HEAD link pointing at a funny place' '
|
|
|
|
|
test_when_finished "mv .git/SAVED_HEAD .git/HEAD" &&
|
|
|
|
|
mv .git/HEAD .git/SAVED_HEAD &&
|
|
|
|
|
echo "ref: refs/funny/place" >.git/HEAD &&
|
|
|
|
|
# avoid corrupt/broken HEAD from interfering with repo discovery
|
|
|
|
|
test_must_fail env GIT_DIR=.git git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "HEAD points to something strange" out
|
2015-09-23 22:46:39 +02:00
|
|
|
'
|
|
|
|
|
|
2018-10-21 10:08:58 +02:00
|
|
|
test_expect_success 'HEAD link pointing at a funny object (from different wt)' '
|
|
|
|
|
test_when_finished "mv .git/SAVED_HEAD .git/HEAD" &&
|
|
|
|
|
test_when_finished "rm -rf .git/worktrees wt" &&
|
|
|
|
|
git worktree add wt &&
|
|
|
|
|
mv .git/HEAD .git/SAVED_HEAD &&
|
|
|
|
|
echo $ZERO_OID >.git/HEAD &&
|
|
|
|
|
# avoid corrupt/broken HEAD from interfering with repo discovery
|
|
|
|
|
test_must_fail git -C wt fsck 2>out &&
|
2019-01-04 22:33:31 +01:00
|
|
|
test_i18ngrep "main-worktree/HEAD: detached HEAD points" out
|
2018-10-21 10:08:58 +02:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'other worktree HEAD link pointing at a funny object' '
|
|
|
|
|
test_when_finished "rm -rf .git/worktrees other" &&
|
|
|
|
|
git worktree add other &&
|
|
|
|
|
echo $ZERO_OID >.git/worktrees/other/HEAD &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2019-01-04 22:33:31 +01:00
|
|
|
test_i18ngrep "worktrees/other/HEAD: detached HEAD points" out
|
2018-10-21 10:08:58 +02:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'other worktree HEAD link pointing at missing object' '
|
|
|
|
|
test_when_finished "rm -rf .git/worktrees other" &&
|
|
|
|
|
git worktree add other &&
|
|
|
|
|
echo "Contents missing from repo" | git hash-object --stdin >.git/worktrees/other/HEAD &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2019-01-04 22:33:31 +01:00
|
|
|
test_i18ngrep "worktrees/other/HEAD: invalid sha1 pointer" out
|
2018-10-21 10:08:58 +02:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'other worktree HEAD link pointing at a funny place' '
|
|
|
|
|
test_when_finished "rm -rf .git/worktrees other" &&
|
|
|
|
|
git worktree add other &&
|
|
|
|
|
echo "ref: refs/funny/place" >.git/worktrees/other/HEAD &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2019-01-04 22:33:31 +01:00
|
|
|
test_i18ngrep "worktrees/other/HEAD points to something strange" out
|
2018-10-21 10:08:58 +02:00
|
|
|
'
|
|
|
|
|
|
2020-02-22 21:17:42 +01:00
|
|
|
test_expect_success 'commit with multiple signatures is okay' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
cat >sigs <<-EOF &&
|
|
|
|
|
gpgsig -----BEGIN PGP SIGNATURE-----
|
|
|
|
|
VGhpcyBpcyBub3QgcmVhbGx5IGEgc2lnbmF0dXJlLg==
|
|
|
|
|
-----END PGP SIGNATURE-----
|
|
|
|
|
gpgsig-sha256 -----BEGIN PGP SIGNATURE-----
|
|
|
|
|
VGhpcyBpcyBub3QgcmVhbGx5IGEgc2lnbmF0dXJlLg==
|
|
|
|
|
-----END PGP SIGNATURE-----
|
|
|
|
|
EOF
|
|
|
|
|
sed -e "/^committer/q" basis >okay &&
|
|
|
|
|
cat sigs >>okay &&
|
|
|
|
|
echo >>okay &&
|
|
|
|
|
sed -e "1,/^$/d" basis >>okay &&
|
|
|
|
|
cat okay &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <okay) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
|
|
|
|
git fsck 2>out &&
|
|
|
|
|
cat out &&
|
|
|
|
|
! grep "commit $new" out
|
|
|
|
|
'
|
|
|
|
|
|
2010-04-24 18:06:08 +02:00
|
|
|
test_expect_success 'email without @ is okay' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/@/AT/" basis >okay &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <okay) &&
|
2010-09-07 03:47:07 +02:00
|
|
|
test_when_finished "remove_object $new" &&
|
2010-04-24 18:06:08 +02:00
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
2010-09-07 03:47:07 +02:00
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2010-04-24 18:06:08 +02:00
|
|
|
git fsck 2>out &&
|
2010-09-07 03:47:07 +02:00
|
|
|
! grep "commit $new" out
|
2010-04-24 18:06:08 +02:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'email with embedded > is not okay' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/@[a-z]/&>/" basis >bad-email &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <bad-email) &&
|
2010-09-07 03:47:07 +02:00
|
|
|
test_when_finished "remove_object $new" &&
|
2010-04-24 18:06:08 +02:00
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
2010-09-07 03:47:07 +02:00
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new" out
|
2010-04-24 18:06:08 +02:00
|
|
|
'
|
2009-02-19 12:13:39 +01:00
|
|
|
|
2011-08-11 12:21:10 +02:00
|
|
|
test_expect_success 'missing < email delimiter is reported nicely' '
|
2011-08-11 12:21:09 +02:00
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/<//" basis >bad-email-2 &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <bad-email-2) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new.* - bad name" out
|
2011-08-11 12:21:09 +02:00
|
|
|
'
|
|
|
|
|
|
2011-08-11 12:21:10 +02:00
|
|
|
test_expect_success 'missing email is reported nicely' '
|
2011-08-11 12:21:09 +02:00
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/[a-z]* <[^>]*>//" basis >bad-email-3 &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <bad-email-3) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new.* - missing email" out
|
2011-08-11 12:21:09 +02:00
|
|
|
'
|
|
|
|
|
|
2011-08-11 12:21:10 +02:00
|
|
|
test_expect_success '> in name is reported' '
|
2011-08-11 12:21:09 +02:00
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/ </> </" basis >bad-email-4 &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <bad-email-4) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new" out
|
2011-08-11 12:21:09 +02:00
|
|
|
'
|
|
|
|
|
|
fsck: report integer overflow in author timestamps
When we check commit objects, we complain if commit->date is
ULONG_MAX, which is an indication that we saw integer
overflow when parsing it. However, we do not do any check at
all for author lines, which also contain a timestamp.
Let's actually check the timestamps on each ident line
with strtoul. This catches both author and committer lines,
and we can get rid of the now-redundant commit->date check.
Note that like the existing check, we compare only against
ULONG_MAX. Now that we are calling strtoul at the site of
the check, we could be slightly more careful and also check
that errno is set to ERANGE. However, this will make further
refactoring in future patches a little harder, and it
doesn't really matter in practice.
For 32-bit systems, one would have to create a commit at the
exact wrong second in 2038. But by the time we get close to
that, all systems will hopefully have moved to 64-bit (and
if they haven't, they have a real problem one second later).
For 64-bit systems, by the time we get close to ULONG_MAX,
all systems will hopefully have been consumed in the fiery
wrath of our expanding Sun.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-02-24 08:39:04 +01:00
|
|
|
# date is 2^64 + 1
|
|
|
|
|
test_expect_success 'integer overflow in timestamps is reported' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/^\\(author .*>\\) [0-9]*/\\1 18446744073709551617/" \
|
|
|
|
|
<basis >bad-timestamp &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <bad-timestamp) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new.*integer overflow" out
|
fsck: report integer overflow in author timestamps
When we check commit objects, we complain if commit->date is
ULONG_MAX, which is an indication that we saw integer
overflow when parsing it. However, we do not do any check at
all for author lines, which also contain a timestamp.
Let's actually check the timestamps on each ident line
with strtoul. This catches both author and committer lines,
and we can get rid of the now-redundant commit->date check.
Note that like the existing check, we compare only against
ULONG_MAX. Now that we are calling strtoul at the site of
the check, we could be slightly more careful and also check
that errno is set to ERANGE. However, this will make further
refactoring in future patches a little harder, and it
doesn't really matter in practice.
For 32-bit systems, one would have to create a commit at the
exact wrong second in 2038. But by the time we get close to
that, all systems will hopefully have moved to 64-bit (and
if they haven't, they have a real problem one second later).
For 64-bit systems, by the time we get close to ULONG_MAX,
all systems will hopefully have been consumed in the fiery
wrath of our expanding Sun.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-02-24 08:39:04 +01:00
|
|
|
'
|
|
|
|
|
|
2015-11-19 17:20:14 +01:00
|
|
|
test_expect_success 'commit with NUL in header' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/author ./author Q/" <basis | q_to_nul >commit-NUL-header &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <commit-NUL-header) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $new.*unterminated header: NUL at offset" out
|
2015-11-19 17:20:14 +01:00
|
|
|
'
|
|
|
|
|
|
2016-09-27 22:59:51 +02:00
|
|
|
test_expect_success 'tree object with duplicate entries' '
|
2017-01-16 22:24:03 +01:00
|
|
|
test_when_finished "for i in \$T; do remove_object \$i; done" &&
|
2014-08-29 22:31:46 +02:00
|
|
|
T=$(
|
|
|
|
|
GIT_INDEX_FILE=test-index &&
|
|
|
|
|
export GIT_INDEX_FILE &&
|
|
|
|
|
rm -f test-index &&
|
|
|
|
|
>x &&
|
|
|
|
|
git add x &&
|
2017-01-16 22:24:03 +01:00
|
|
|
git rev-parse :x &&
|
2014-08-29 22:31:46 +02:00
|
|
|
T=$(git write-tree) &&
|
2017-01-16 22:24:03 +01:00
|
|
|
echo $T &&
|
2014-08-29 22:31:46 +02:00
|
|
|
(
|
|
|
|
|
git cat-file tree $T &&
|
|
|
|
|
git cat-file tree $T
|
|
|
|
|
) |
|
|
|
|
|
git hash-object -w -t tree --stdin
|
|
|
|
|
) &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in tree .*contains duplicate file entries" out
|
2014-08-29 22:31:46 +02:00
|
|
|
'
|
|
|
|
|
|
2020-05-21 11:52:28 +02:00
|
|
|
check_duplicate_names () {
|
|
|
|
|
expect=$1 &&
|
|
|
|
|
shift &&
|
|
|
|
|
names=$@ &&
|
|
|
|
|
test_expect_$expect "tree object with duplicate names: $names" '
|
|
|
|
|
test_when_finished "remove_object \$blob" &&
|
|
|
|
|
test_when_finished "remove_object \$tree" &&
|
|
|
|
|
test_when_finished "remove_object \$badtree" &&
|
|
|
|
|
blob=$(echo blob | git hash-object -w --stdin) &&
|
|
|
|
|
printf "100644 blob %s\t%s\n" $blob x.2 >tree &&
|
|
|
|
|
tree=$(git mktree <tree) &&
|
|
|
|
|
for name in $names
|
|
|
|
|
do
|
|
|
|
|
case "$name" in
|
|
|
|
|
*/) printf "040000 tree %s\t%s\n" $tree "${name%/}" ;;
|
|
|
|
|
*) printf "100644 blob %s\t%s\n" $blob "$name" ;;
|
|
|
|
|
esac
|
|
|
|
|
done >badtree &&
|
|
|
|
|
badtree=$(git mktree <badtree) &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
test_i18ngrep "$badtree" out &&
|
|
|
|
|
test_i18ngrep "error in tree .*contains duplicate file entries" out
|
|
|
|
|
'
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
check_duplicate_names success x x.1 x/
|
|
|
|
|
check_duplicate_names success x x.1.2 x.1/ x/
|
2020-05-21 11:52:54 +02:00
|
|
|
check_duplicate_names success x x.1 x.1.2 x/
|
2020-05-10 18:12:16 +02:00
|
|
|
|
2016-09-27 22:59:51 +02:00
|
|
|
test_expect_success 'unparseable tree object' '
|
2019-06-29 00:59:21 +02:00
|
|
|
test_oid_cache <<-\EOF &&
|
|
|
|
|
junk sha1:twenty-bytes-of-junk
|
|
|
|
|
junk sha256:twenty-bytes-of-junk-twelve-more
|
|
|
|
|
EOF
|
|
|
|
|
|
2016-09-27 22:59:51 +02:00
|
|
|
test_when_finished "git update-ref -d refs/heads/wrong" &&
|
|
|
|
|
test_when_finished "remove_object \$tree_sha1" &&
|
|
|
|
|
test_when_finished "remove_object \$commit_sha1" &&
|
2019-06-29 00:59:21 +02:00
|
|
|
junk=$(test_oid junk) &&
|
|
|
|
|
tree_sha1=$(printf "100644 \0$junk" | git hash-object -t tree --stdin -w --literally) &&
|
2016-09-27 22:59:51 +02:00
|
|
|
commit_sha1=$(git commit-tree $tree_sha1) &&
|
|
|
|
|
git update-ref refs/heads/wrong $commit_sha1 &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
test_i18ngrep "error: empty filename in tree entry" out &&
|
|
|
|
|
test_i18ngrep "$tree_sha1" out &&
|
|
|
|
|
test_i18ngrep ! "fatal: empty filename in tree entry" out
|
|
|
|
|
'
|
|
|
|
|
|
2017-10-05 21:41:26 +02:00
|
|
|
test_expect_success 'tree entry with type mismatch' '
|
|
|
|
|
test_when_finished "remove_object \$blob" &&
|
|
|
|
|
test_when_finished "remove_object \$tree" &&
|
|
|
|
|
test_when_finished "remove_object \$commit" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/type_mismatch" &&
|
|
|
|
|
blob=$(echo blob | git hash-object -w --stdin) &&
|
|
|
|
|
blob_bin=$(echo $blob | hex2oct) &&
|
|
|
|
|
tree=$(
|
|
|
|
|
printf "40000 dir\0${blob_bin}100644 file\0${blob_bin}" |
|
|
|
|
|
git hash-object -t tree --stdin -w --literally
|
|
|
|
|
) &&
|
|
|
|
|
commit=$(git commit-tree $tree) &&
|
|
|
|
|
git update-ref refs/heads/type_mismatch $commit &&
|
|
|
|
|
test_must_fail git fsck >out 2>&1 &&
|
|
|
|
|
test_i18ngrep "is a blob, not a tree" out &&
|
|
|
|
|
test_i18ngrep ! "dangling blob" out
|
|
|
|
|
'
|
|
|
|
|
|
2010-02-20 01:18:44 +01:00
|
|
|
test_expect_success 'tag pointing to nonexistent' '
|
2019-06-29 00:59:21 +02:00
|
|
|
badoid=$(test_oid deadbeef) &&
|
|
|
|
|
cat >invalid-tag <<-EOF &&
|
|
|
|
|
object $badoid
|
2010-09-07 03:47:07 +02:00
|
|
|
type commit
|
|
|
|
|
tag invalid
|
|
|
|
|
tagger T A Gger <tagger@example.com> 1234567890 -0000
|
|
|
|
|
|
|
|
|
|
This is an invalid tag.
|
|
|
|
|
EOF
|
|
|
|
|
|
|
|
|
|
tag=$(git hash-object -t tag -w --stdin <invalid-tag) &&
|
|
|
|
|
test_when_finished "remove_object $tag" &&
|
|
|
|
|
echo $tag >.git/refs/tags/invalid &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/tags/invalid" &&
|
2010-02-20 01:18:44 +01:00
|
|
|
test_must_fail git fsck --tags >out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "broken link" out
|
2009-02-19 12:13:39 +01:00
|
|
|
'
|
|
|
|
|
|
2010-02-20 01:18:44 +01:00
|
|
|
test_expect_success 'tag pointing to something else than its type' '
|
2010-09-07 03:47:07 +02:00
|
|
|
sha=$(echo blob | git hash-object -w --stdin) &&
|
|
|
|
|
test_when_finished "remove_object $sha" &&
|
|
|
|
|
cat >wrong-tag <<-EOF &&
|
|
|
|
|
object $sha
|
|
|
|
|
type commit
|
|
|
|
|
tag wrong
|
|
|
|
|
tagger T A Gger <tagger@example.com> 1234567890 -0000
|
|
|
|
|
|
|
|
|
|
This is an invalid tag.
|
|
|
|
|
EOF
|
|
|
|
|
|
|
|
|
|
tag=$(git hash-object -t tag -w --stdin <wrong-tag) &&
|
|
|
|
|
test_when_finished "remove_object $tag" &&
|
|
|
|
|
echo $tag >.git/refs/tags/wrong &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/tags/wrong" &&
|
t1450: the order the objects are checked is undefined
When a tag T points at an object X that is of a type that is
different from what the tag records as, fsck should report it as an
error.
However, depending on the order X and T are checked individually,
the actual error message can be different. If X is checked first,
fsck remembers X's type and then when it checks T, it notices that T
records X as a wrong type (i.e. the complaint is about a broken tag
T). If T is checked first, on the other hand, fsck remembers that we
need to verify X is of the type tag records, and when it later
checks X, it notices that X is of a wrong type (i.e. the complaint
is about a broken object X).
The important thing is that fsck notices such an error and diagnoses
the issue on object X, but the test was expecting that we happen to
check objects in the order to make us detect issues with tag T, not
with object X. Remove this unwarranted assumption.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-03 00:08:16 +02:00
|
|
|
test_must_fail git fsck --tags
|
2009-02-19 12:13:39 +01:00
|
|
|
'
|
|
|
|
|
|
2014-09-11 16:26:41 +02:00
|
|
|
test_expect_success 'tag with incorrect tag name & missing tagger' '
|
|
|
|
|
sha=$(git rev-parse HEAD) &&
|
|
|
|
|
cat >wrong-tag <<-EOF &&
|
|
|
|
|
object $sha
|
|
|
|
|
type commit
|
|
|
|
|
tag wrong name format
|
|
|
|
|
|
|
|
|
|
This is an invalid tag.
|
|
|
|
|
EOF
|
|
|
|
|
|
|
|
|
|
tag=$(git hash-object -t tag -w --stdin <wrong-tag) &&
|
|
|
|
|
test_when_finished "remove_object $tag" &&
|
|
|
|
|
echo $tag >.git/refs/tags/wrong &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/tags/wrong" &&
|
|
|
|
|
git fsck --tags 2>out &&
|
2014-12-08 06:48:13 +01:00
|
|
|
|
|
|
|
|
cat >expect <<-EOF &&
|
2015-06-22 17:25:52 +02:00
|
|
|
warning in tag $tag: badTagName: invalid '\''tag'\'' name: wrong name format
|
|
|
|
|
warning in tag $tag: missingTaggerEntry: invalid format - expected '\''tagger'\'' line
|
2014-12-08 06:48:13 +01:00
|
|
|
EOF
|
2021-02-11 02:53:53 +01:00
|
|
|
test_cmp expect out
|
2014-09-11 16:26:41 +02:00
|
|
|
'
|
|
|
|
|
|
2014-09-11 23:16:36 +02:00
|
|
|
test_expect_success 'tag with bad tagger' '
|
|
|
|
|
sha=$(git rev-parse HEAD) &&
|
|
|
|
|
cat >wrong-tag <<-EOF &&
|
|
|
|
|
object $sha
|
|
|
|
|
type commit
|
|
|
|
|
tag not-quite-wrong
|
|
|
|
|
tagger Bad Tagger Name
|
|
|
|
|
|
|
|
|
|
This is an invalid tag.
|
|
|
|
|
EOF
|
|
|
|
|
|
|
|
|
|
tag=$(git hash-object --literally -t tag -w --stdin <wrong-tag) &&
|
|
|
|
|
test_when_finished "remove_object $tag" &&
|
|
|
|
|
echo $tag >.git/refs/tags/wrong &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/tags/wrong" &&
|
|
|
|
|
test_must_fail git fsck --tags 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in tag .*: invalid author/committer" out
|
2014-09-11 23:16:36 +02:00
|
|
|
'
|
|
|
|
|
|
2015-11-19 17:25:31 +01:00
|
|
|
test_expect_success 'tag with NUL in header' '
|
2015-11-19 17:20:14 +01:00
|
|
|
sha=$(git rev-parse HEAD) &&
|
|
|
|
|
q_to_nul >tag-NUL-header <<-EOF &&
|
|
|
|
|
object $sha
|
|
|
|
|
type commit
|
|
|
|
|
tag contains-Q-in-header
|
|
|
|
|
tagger T A Gger <tagger@example.com> 1234567890 -0000
|
|
|
|
|
|
|
|
|
|
This is an invalid tag.
|
|
|
|
|
EOF
|
|
|
|
|
|
|
|
|
|
tag=$(git hash-object --literally -t tag -w --stdin <tag-NUL-header) &&
|
|
|
|
|
test_when_finished "remove_object $tag" &&
|
|
|
|
|
echo $tag >.git/refs/tags/wrong &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/tags/wrong" &&
|
|
|
|
|
test_must_fail git fsck --tags 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in tag $tag.*unterminated header: NUL at offset" out
|
2015-11-19 17:20:14 +01:00
|
|
|
'
|
|
|
|
|
|
2010-09-07 03:47:07 +02:00
|
|
|
test_expect_success 'cleaned up' '
|
|
|
|
|
git fsck >actual 2>&1 &&
|
tests: use 'test_must_be_empty' instead of 'test_cmp <empty> <out>'
Using 'test_must_be_empty' is shorter and more idiomatic than
>empty &&
test_cmp empty out
as it saves the creation of an empty file. Furthermore, sometimes the
expected empty file doesn't have such a descriptive name like 'empty',
and its creation is far away from the place where it's finally used
for comparison (e.g. in 't7600-merge.sh', where two expected empty
files are created in the 'setup' test, but are used only about 500
lines later).
These cases were found by instrumenting 'test_cmp' to error out the
test script when it's used to compare empty files, and then converted
manually.
Note that even after this patch there still remain a lot of cases
where we use 'test_cmp' to check empty files:
- Sometimes the expected output is not hard-coded in the test, but
'test_cmp' is used to ensure that two similar git commands produce
the same output, and that output happens to be empty, e.g. the
test 'submodule update --merge - ignores --merge for new
submodules' in 't7406-submodule-update.sh'.
- Repetitive common tasks, including preparing the expected results
and running 'test_cmp', are often extracted into a helper
function, and some of this helper's callsites expect no output.
- For the same reason as above, the whole 'test_expect_success'
block is within a helper function, e.g. in 't3070-wildmatch.sh'.
- Or 'test_cmp' is invoked in a loop, e.g. the test 'cvs update
(-p)' in 't9400-git-cvsserver-server.sh'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-19 23:57:25 +02:00
|
|
|
test_must_be_empty actual
|
2010-09-07 03:47:07 +02:00
|
|
|
'
|
2009-02-19 12:13:39 +01:00
|
|
|
|
2012-02-13 21:17:11 +01:00
|
|
|
test_expect_success 'rev-list --verify-objects' '
|
|
|
|
|
git rev-list --verify-objects --all >/dev/null 2>out &&
|
tests: use 'test_must_be_empty' instead of 'test_cmp <empty> <out>'
Using 'test_must_be_empty' is shorter and more idiomatic than
>empty &&
test_cmp empty out
as it saves the creation of an empty file. Furthermore, sometimes the
expected empty file doesn't have such a descriptive name like 'empty',
and its creation is far away from the place where it's finally used
for comparison (e.g. in 't7600-merge.sh', where two expected empty
files are created in the 'setup' test, but are used only about 500
lines later).
These cases were found by instrumenting 'test_cmp' to error out the
test script when it's used to compare empty files, and then converted
manually.
Note that even after this patch there still remain a lot of cases
where we use 'test_cmp' to check empty files:
- Sometimes the expected output is not hard-coded in the test, but
'test_cmp' is used to ensure that two similar git commands produce
the same output, and that output happens to be empty, e.g. the
test 'submodule update --merge - ignores --merge for new
submodules' in 't7406-submodule-update.sh'.
- Repetitive common tasks, including preparing the expected results
and running 'test_cmp', are often extracted into a helper
function, and some of this helper's callsites expect no output.
- For the same reason as above, the whole 'test_expect_success'
block is within a helper function, e.g. in 't3070-wildmatch.sh'.
- Or 'test_cmp' is invoked in a loop, e.g. the test 'cvs update
(-p)' in 't9400-git-cvsserver-server.sh'.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-19 23:57:25 +02:00
|
|
|
test_must_be_empty out
|
2012-02-13 21:17:11 +01:00
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'rev-list --verify-objects with bad sha1' '
|
|
|
|
|
sha=$(echo blob | git hash-object -w --stdin) &&
|
2019-06-29 00:59:21 +02:00
|
|
|
old=$(test_oid_to_path $sha) &&
|
|
|
|
|
new=$(dirname $old)/$(test_oid ff_2) &&
|
2012-02-13 21:17:11 +01:00
|
|
|
sha="$(dirname $new)$(basename $new)" &&
|
|
|
|
|
mv .git/objects/$old .git/objects/$new &&
|
|
|
|
|
test_when_finished "remove_object $sha" &&
|
|
|
|
|
git update-index --add --cacheinfo 100644 $sha foo &&
|
|
|
|
|
test_when_finished "git read-tree -u --reset HEAD" &&
|
|
|
|
|
tree=$(git write-tree) &&
|
|
|
|
|
test_when_finished "remove_object $tree" &&
|
|
|
|
|
cmt=$(echo bogus | git commit-tree $tree) &&
|
|
|
|
|
test_when_finished "remove_object $cmt" &&
|
|
|
|
|
git update-ref refs/heads/bogus $cmt &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
|
|
|
|
|
|
|
|
|
test_might_fail git rev-list --verify-objects refs/heads/bogus >/dev/null 2>out &&
|
2019-06-29 00:59:21 +02:00
|
|
|
test_i18ngrep -q "error: hash mismatch $(dirname $new)$(test_oid ff_2)" out
|
2012-02-13 21:17:11 +01:00
|
|
|
'
|
|
|
|
|
|
2015-06-22 17:27:06 +02:00
|
|
|
test_expect_success 'force fsck to ignore double author' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/^author .*/&,&/" <basis | tr , \\n >multiple-authors &&
|
|
|
|
|
new=$(git hash-object -t commit -w --stdin <multiple-authors) &&
|
|
|
|
|
test_when_finished "remove_object $new" &&
|
|
|
|
|
git update-ref refs/heads/bogus "$new" &&
|
|
|
|
|
test_when_finished "git update-ref -d refs/heads/bogus" &&
|
|
|
|
|
test_must_fail git fsck &&
|
|
|
|
|
git -c fsck.multipleAuthors=ignore fsck
|
|
|
|
|
'
|
|
|
|
|
|
2012-07-28 17:06:29 +02:00
|
|
|
_bz='\0'
|
2019-06-29 00:59:21 +02:00
|
|
|
_bzoid=$(printf $ZERO_OID | sed -e 's/00/\\0/g')
|
2012-07-28 17:06:29 +02:00
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices blob entry pointing to null sha1' '
|
|
|
|
|
(git init null-blob &&
|
|
|
|
|
cd null-blob &&
|
2019-06-29 00:59:21 +02:00
|
|
|
sha=$(printf "100644 file$_bz$_bzoid" |
|
2012-07-28 17:06:29 +02:00
|
|
|
git hash-object -w --stdin -t tree) &&
|
|
|
|
|
git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "warning.*null sha1" out
|
2012-07-28 17:06:29 +02:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices submodule entry pointing to null sha1' '
|
|
|
|
|
(git init null-commit &&
|
|
|
|
|
cd null-commit &&
|
2019-06-29 00:59:21 +02:00
|
|
|
sha=$(printf "160000 submodule$_bz$_bzoid" |
|
2012-07-28 17:06:29 +02:00
|
|
|
git hash-object -w --stdin -t tree) &&
|
|
|
|
|
git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "warning.*null sha1" out
|
2012-07-28 17:06:29 +02:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
fsck: complain about HFS+ ".git" aliases in trees
Now that the index can block pathnames that case-fold to
".git" on HFS+, it would be helpful for fsck to notice such
problematic paths. This lets servers which use
receive.fsckObjects block them before the damage spreads.
Note that the fsck check is always on, even for systems
without core.protectHFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
HFS+.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on HFS+ themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git with invisible Unicode code-points mixed
in, meaning mischief is almost certainly what the tree
author had in mind).
Ideally these would be controlled by a separate
"fsck.protectHFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 00:21:57 +01:00
|
|
|
while read name path pretty; do
|
2014-11-24 19:40:11 +01:00
|
|
|
while read mode type; do
|
fsck: complain about HFS+ ".git" aliases in trees
Now that the index can block pathnames that case-fold to
".git" on HFS+, it would be helpful for fsck to notice such
problematic paths. This lets servers which use
receive.fsckObjects block them before the damage spreads.
Note that the fsck check is always on, even for systems
without core.protectHFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
HFS+.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on HFS+ themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git with invisible Unicode code-points mixed
in, meaning mischief is almost certainly what the tree
author had in mind).
Ideally these would be controlled by a separate
"fsck.protectHFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 00:21:57 +01:00
|
|
|
: ${pretty:=$path}
|
|
|
|
|
test_expect_success "fsck notices $pretty as $type" '
|
2014-11-24 19:40:11 +01:00
|
|
|
(
|
|
|
|
|
git init $name-$type &&
|
|
|
|
|
cd $name-$type &&
|
2019-09-12 14:54:05 +02:00
|
|
|
git config core.protectNTFS false &&
|
2014-11-24 19:40:11 +01:00
|
|
|
echo content >file &&
|
|
|
|
|
git add file &&
|
|
|
|
|
git commit -m base &&
|
|
|
|
|
blob=$(git rev-parse :file) &&
|
|
|
|
|
tree=$(git rev-parse HEAD^{tree}) &&
|
|
|
|
|
value=$(eval "echo \$$type") &&
|
|
|
|
|
printf "$mode $type %s\t%s" "$value" "$path" >bad &&
|
fsck: complain about NTFS ".git" aliases in trees
Now that the index can block pathnames that can be mistaken
to mean ".git" on NTFS and FAT32, it would be helpful for
fsck to notice such problematic paths. This lets servers
which use receive.fsckObjects block them before the damage
spreads.
Note that the fsck check is always on, even for systems
without core.protectNTFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
NTFS.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on NTFS themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git or git~1, meaning mischief is almost
certainly what the tree author had in mind).
Ideally these would be controlled by a separate
"fsck.protectNTFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-10 22:28:27 +01:00
|
|
|
bad_tree=$(git mktree <bad) &&
|
2014-11-24 19:40:11 +01:00
|
|
|
git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "warning.*tree $bad_tree" out
|
2014-11-24 19:40:11 +01:00
|
|
|
)'
|
|
|
|
|
done <<-\EOF
|
|
|
|
|
100644 blob
|
|
|
|
|
040000 tree
|
|
|
|
|
EOF
|
fsck: complain about HFS+ ".git" aliases in trees
Now that the index can block pathnames that case-fold to
".git" on HFS+, it would be helpful for fsck to notice such
problematic paths. This lets servers which use
receive.fsckObjects block them before the damage spreads.
Note that the fsck check is always on, even for systems
without core.protectHFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
HFS+.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on HFS+ themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git with invisible Unicode code-points mixed
in, meaning mischief is almost certainly what the tree
author had in mind).
Ideally these would be controlled by a separate
"fsck.protectHFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 00:21:57 +01:00
|
|
|
done <<-EOF
|
2014-11-24 19:40:11 +01:00
|
|
|
dot .
|
|
|
|
|
dotdot ..
|
|
|
|
|
dotgit .git
|
2014-11-24 19:40:44 +01:00
|
|
|
dotgit-case .GIT
|
fsck: complain about HFS+ ".git" aliases in trees
Now that the index can block pathnames that case-fold to
".git" on HFS+, it would be helpful for fsck to notice such
problematic paths. This lets servers which use
receive.fsckObjects block them before the damage spreads.
Note that the fsck check is always on, even for systems
without core.protectHFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
HFS+.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on HFS+ themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git with invisible Unicode code-points mixed
in, meaning mischief is almost certainly what the tree
author had in mind).
Ideally these would be controlled by a separate
"fsck.protectHFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 00:21:57 +01:00
|
|
|
dotgit-unicode .gI${u200c}T .gI{u200c}T
|
fsck: complain about NTFS ".git" aliases in trees
Now that the index can block pathnames that can be mistaken
to mean ".git" on NTFS and FAT32, it would be helpful for
fsck to notice such problematic paths. This lets servers
which use receive.fsckObjects block them before the damage
spreads.
Note that the fsck check is always on, even for systems
without core.protectNTFS set. This is technically more
restrictive than we need to be, as a set of users on ext4
could happily use these odd filenames without caring about
NTFS.
However, on balance, it's helpful for all servers to block
these (because the paths can be used for mischief, and
servers which bother to fsck would want to stop the spread
whether they are on NTFS themselves or not), and hardly
anybody will be affected (because the blocked names are
variants of .git or git~1, meaning mischief is almost
certainly what the tree author had in mind).
Ideally these would be controlled by a separate
"fsck.protectNTFS" flag. However, it would be much nicer to
be able to enable/disable _any_ fsck flag individually, and
any scheme we choose should match such a system. Given the
likelihood of anybody using such a path in practice, it is
not unreasonable to wait until such a system materializes.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-10 22:28:27 +01:00
|
|
|
dotgit-case2 .Git
|
|
|
|
|
git-tilde1 git~1
|
|
|
|
|
dotgitdot .git.
|
|
|
|
|
dot-backslash-case .\\\\.GIT\\\\foobar
|
|
|
|
|
dotgit-case-backslash .git\\\\foobar
|
2014-11-24 19:40:11 +01:00
|
|
|
EOF
|
2012-11-28 22:35:29 +01:00
|
|
|
|
2014-12-23 09:45:36 +01:00
|
|
|
test_expect_success 'fsck allows .Ňit' '
|
|
|
|
|
(
|
|
|
|
|
git init not-dotgit &&
|
|
|
|
|
cd not-dotgit &&
|
|
|
|
|
echo content >file &&
|
|
|
|
|
git add file &&
|
|
|
|
|
git commit -m base &&
|
|
|
|
|
blob=$(git rev-parse :file) &&
|
|
|
|
|
printf "100644 blob $blob\t.\\305\\207it" >tree &&
|
|
|
|
|
tree=$(git mktree <tree) &&
|
|
|
|
|
git fsck 2>err &&
|
|
|
|
|
test_line_count = 0 err
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2016-04-14 19:58:22 +02:00
|
|
|
test_expect_success 'NUL in commit' '
|
|
|
|
|
rm -fr nul-in-commit &&
|
|
|
|
|
git init nul-in-commit &&
|
|
|
|
|
(
|
|
|
|
|
cd nul-in-commit &&
|
|
|
|
|
git commit --allow-empty -m "initial commitQNUL after message" &&
|
|
|
|
|
git cat-file commit HEAD >original &&
|
|
|
|
|
q_to_nul <original >munged &&
|
|
|
|
|
git hash-object -w -t commit --stdin <munged >name &&
|
|
|
|
|
git branch bad $(cat name) &&
|
|
|
|
|
|
|
|
|
|
test_must_fail git -c fsck.nulInCommit=error fsck 2>warn.1 &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep nulInCommit warn.1 &&
|
2016-04-14 19:58:22 +02:00
|
|
|
git fsck 2>warn.2 &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep nulInCommit warn.2
|
2016-04-14 19:58:22 +02:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2014-09-12 05:38:30 +02:00
|
|
|
# create a static test repo which is broken by omitting
|
|
|
|
|
# one particular object ($1, which is looked up via rev-parse
|
|
|
|
|
# in the new repository).
|
|
|
|
|
create_repo_missing () {
|
|
|
|
|
rm -rf missing &&
|
|
|
|
|
git init missing &&
|
|
|
|
|
(
|
|
|
|
|
cd missing &&
|
|
|
|
|
git commit -m one --allow-empty &&
|
|
|
|
|
mkdir subdir &&
|
|
|
|
|
echo content >subdir/file &&
|
|
|
|
|
git add subdir/file &&
|
|
|
|
|
git commit -m two &&
|
|
|
|
|
unrelated=$(echo unrelated | git hash-object --stdin -w) &&
|
|
|
|
|
git tag -m foo tag $unrelated &&
|
|
|
|
|
sha1=$(git rev-parse --verify "$1") &&
|
|
|
|
|
path=$(echo $sha1 | sed 's|..|&/|') &&
|
|
|
|
|
rm .git/objects/$path
|
|
|
|
|
)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices missing blob' '
|
|
|
|
|
create_repo_missing HEAD:subdir/file &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices missing subtree' '
|
|
|
|
|
create_repo_missing HEAD:subdir &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices missing root tree' '
|
|
|
|
|
create_repo_missing HEAD^{tree} &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices missing parent' '
|
|
|
|
|
create_repo_missing HEAD^ &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices missing tagged object' '
|
|
|
|
|
create_repo_missing tag^{blob} &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices ref pointing to missing commit' '
|
|
|
|
|
create_repo_missing HEAD &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
|
|
|
|
test_expect_success 'fsck notices ref pointing to missing tag' '
|
|
|
|
|
create_repo_missing tag &&
|
|
|
|
|
test_must_fail git -C missing fsck
|
|
|
|
|
'
|
|
|
|
|
|
2015-06-22 17:27:12 +02:00
|
|
|
test_expect_success 'fsck --connectivity-only' '
|
|
|
|
|
rm -rf connectivity-only &&
|
|
|
|
|
git init connectivity-only &&
|
|
|
|
|
(
|
|
|
|
|
cd connectivity-only &&
|
|
|
|
|
touch empty &&
|
|
|
|
|
git add empty &&
|
|
|
|
|
test_commit empty &&
|
fsck: prepare dummy objects for --connectivity-check
Normally fsck makes a pass over all objects to check their
integrity, and then follows up with a reachability check to
make sure we have all of the referenced objects (and to know
which ones are dangling). The latter checks for the HAS_OBJ
flag in obj->flags to see if we found the object in the
first pass.
Commit 02976bf85 (fsck: introduce `git fsck --connectivity-only`,
2015-06-22) taught fsck to skip the initial pass, and to
fallback to has_sha1_file() instead of the HAS_OBJ check.
However, it converted only one HAS_OBJ check to use
has_sha1_file(). But there are many other places in
builtin/fsck.c that assume that the flag is set (or that
lookup_object() will return an object at all). This leads to
several bugs with --connectivity-only:
1. mark_object() will not queue objects for examination,
so recursively following links from commits to trees,
etc, did nothing. I.e., we were checking the
reachability of hardly anything at all.
2. When a set of heads is given on the command-line, we
use lookup_object() to see if they exist. But without
the initial pass, we assume nothing exists.
3. When loading reflog entries, we do a similar
lookup_object() check, and complain that the reflog is
broken if the object doesn't exist in our hash.
So in short, --connectivity-only is broken pretty badly, and
will claim that your repository is fine when it's not.
Presumably nobody noticed for a few reasons.
One is that the embedded test does not actually test the
recursive nature of the reachability check. All of the
missing objects are still in the index, and we directly
check items from the index. This patch modifies the test to
delete the index, which shows off breakage (1).
Another is that --connectivity-only just skips the initial
pass for loose objects. So on a real repository, the packed
objects were still checked correctly. But on the flipside,
it means that "git fsck --connectivity-only" still checks
the sha1 of all of the packed objects, nullifying its
original purpose of being a faster git-fsck.
And of course the final problem is that the bug only shows
up when there _is_ corruption, which is rare. So anybody
running "git fsck --connectivity-only" proactively would
assume it was being thorough, when it was not.
One possibility for fixing this is to find all of the spots
that rely on HAS_OBJ and tweak them for the connectivity-only
case. But besides the risk that we might miss a spot (and I
found three already, corresponding to the three bugs above),
there are other parts of fsck that _can't_ work without a
full list of objects. E.g., the list of dangling objects.
Instead, let's make the connectivity-only case look more
like the normal case. Rather than skip the initial pass
completely, we'll do an abbreviated one that sets up the
HAS_OBJ flag for each object, without actually loading the
object data.
That's simple and fast, and we don't have to care about the
connectivity_only flag in the rest of the code at all.
While we're at it, let's make sure we treat loose and packed
objects the same (i.e., setting up dummy objects for both
and skipping the actual sha1 check). That makes the
connectivity-only check actually fast on a real repo (40
seconds versus 180 seconds on my copy of linux.git).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-17 22:32:57 +01:00
|
|
|
|
|
|
|
|
# Drop the index now; we want to be sure that we
|
|
|
|
|
# recursively notice the broken objects
|
|
|
|
|
# because they are reachable from refs, not because
|
|
|
|
|
# they are in the index.
|
|
|
|
|
rm -f .git/index &&
|
|
|
|
|
|
|
|
|
|
# corrupt the blob, but in a way that we can still identify
|
|
|
|
|
# its type. That lets us see that --connectivity-only is
|
|
|
|
|
# not actually looking at the contents, but leaves it
|
|
|
|
|
# free to examine the type if it chooses.
|
2019-06-29 00:59:21 +02:00
|
|
|
empty=.git/objects/$(test_oid_to_path $EMPTY_BLOB) &&
|
fsck: prepare dummy objects for --connectivity-check
Normally fsck makes a pass over all objects to check their
integrity, and then follows up with a reachability check to
make sure we have all of the referenced objects (and to know
which ones are dangling). The latter checks for the HAS_OBJ
flag in obj->flags to see if we found the object in the
first pass.
Commit 02976bf85 (fsck: introduce `git fsck --connectivity-only`,
2015-06-22) taught fsck to skip the initial pass, and to
fallback to has_sha1_file() instead of the HAS_OBJ check.
However, it converted only one HAS_OBJ check to use
has_sha1_file(). But there are many other places in
builtin/fsck.c that assume that the flag is set (or that
lookup_object() will return an object at all). This leads to
several bugs with --connectivity-only:
1. mark_object() will not queue objects for examination,
so recursively following links from commits to trees,
etc, did nothing. I.e., we were checking the
reachability of hardly anything at all.
2. When a set of heads is given on the command-line, we
use lookup_object() to see if they exist. But without
the initial pass, we assume nothing exists.
3. When loading reflog entries, we do a similar
lookup_object() check, and complain that the reflog is
broken if the object doesn't exist in our hash.
So in short, --connectivity-only is broken pretty badly, and
will claim that your repository is fine when it's not.
Presumably nobody noticed for a few reasons.
One is that the embedded test does not actually test the
recursive nature of the reachability check. All of the
missing objects are still in the index, and we directly
check items from the index. This patch modifies the test to
delete the index, which shows off breakage (1).
Another is that --connectivity-only just skips the initial
pass for loose objects. So on a real repository, the packed
objects were still checked correctly. But on the flipside,
it means that "git fsck --connectivity-only" still checks
the sha1 of all of the packed objects, nullifying its
original purpose of being a faster git-fsck.
And of course the final problem is that the bug only shows
up when there _is_ corruption, which is rare. So anybody
running "git fsck --connectivity-only" proactively would
assume it was being thorough, when it was not.
One possibility for fixing this is to find all of the spots
that rely on HAS_OBJ and tweak them for the connectivity-only
case. But besides the risk that we might miss a spot (and I
found three already, corresponding to the three bugs above),
there are other parts of fsck that _can't_ work without a
full list of objects. E.g., the list of dangling objects.
Instead, let's make the connectivity-only case look more
like the normal case. Rather than skip the initial pass
completely, we'll do an abbreviated one that sets up the
HAS_OBJ flag for each object, without actually loading the
object data.
That's simple and fast, and we don't have to care about the
connectivity_only flag in the rest of the code at all.
While we're at it, let's make sure we treat loose and packed
objects the same (i.e., setting up dummy objects for both
and skipping the actual sha1 check). That makes the
connectivity-only check actually fast on a real repo (40
seconds versus 180 seconds on my copy of linux.git).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-17 22:32:57 +01:00
|
|
|
blob=$(echo unrelated | git hash-object -w --stdin) &&
|
2017-01-24 14:27:49 +01:00
|
|
|
mv -f $(sha1_file $blob) $empty &&
|
fsck: prepare dummy objects for --connectivity-check
Normally fsck makes a pass over all objects to check their
integrity, and then follows up with a reachability check to
make sure we have all of the referenced objects (and to know
which ones are dangling). The latter checks for the HAS_OBJ
flag in obj->flags to see if we found the object in the
first pass.
Commit 02976bf85 (fsck: introduce `git fsck --connectivity-only`,
2015-06-22) taught fsck to skip the initial pass, and to
fallback to has_sha1_file() instead of the HAS_OBJ check.
However, it converted only one HAS_OBJ check to use
has_sha1_file(). But there are many other places in
builtin/fsck.c that assume that the flag is set (or that
lookup_object() will return an object at all). This leads to
several bugs with --connectivity-only:
1. mark_object() will not queue objects for examination,
so recursively following links from commits to trees,
etc, did nothing. I.e., we were checking the
reachability of hardly anything at all.
2. When a set of heads is given on the command-line, we
use lookup_object() to see if they exist. But without
the initial pass, we assume nothing exists.
3. When loading reflog entries, we do a similar
lookup_object() check, and complain that the reflog is
broken if the object doesn't exist in our hash.
So in short, --connectivity-only is broken pretty badly, and
will claim that your repository is fine when it's not.
Presumably nobody noticed for a few reasons.
One is that the embedded test does not actually test the
recursive nature of the reachability check. All of the
missing objects are still in the index, and we directly
check items from the index. This patch modifies the test to
delete the index, which shows off breakage (1).
Another is that --connectivity-only just skips the initial
pass for loose objects. So on a real repository, the packed
objects were still checked correctly. But on the flipside,
it means that "git fsck --connectivity-only" still checks
the sha1 of all of the packed objects, nullifying its
original purpose of being a faster git-fsck.
And of course the final problem is that the bug only shows
up when there _is_ corruption, which is rare. So anybody
running "git fsck --connectivity-only" proactively would
assume it was being thorough, when it was not.
One possibility for fixing this is to find all of the spots
that rely on HAS_OBJ and tweak them for the connectivity-only
case. But besides the risk that we might miss a spot (and I
found three already, corresponding to the three bugs above),
there are other parts of fsck that _can't_ work without a
full list of objects. E.g., the list of dangling objects.
Instead, let's make the connectivity-only case look more
like the normal case. Rather than skip the initial pass
completely, we'll do an abbreviated one that sets up the
HAS_OBJ flag for each object, without actually loading the
object data.
That's simple and fast, and we don't have to care about the
connectivity_only flag in the rest of the code at all.
While we're at it, let's make sure we treat loose and packed
objects the same (i.e., setting up dummy objects for both
and skipping the actual sha1 check). That makes the
connectivity-only check actually fast on a real repo (40
seconds versus 180 seconds on my copy of linux.git).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-17 22:32:57 +01:00
|
|
|
|
2015-06-22 17:27:12 +02:00
|
|
|
test_must_fail git fsck --strict &&
|
|
|
|
|
git fsck --strict --connectivity-only &&
|
|
|
|
|
tree=$(git rev-parse HEAD:) &&
|
|
|
|
|
suffix=${tree#??} &&
|
|
|
|
|
tree=.git/objects/${tree%$suffix}/$suffix &&
|
|
|
|
|
rm -f $tree &&
|
|
|
|
|
echo invalid >$tree &&
|
|
|
|
|
test_must_fail git fsck --strict --connectivity-only
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
fsck: prepare dummy objects for --connectivity-check
Normally fsck makes a pass over all objects to check their
integrity, and then follows up with a reachability check to
make sure we have all of the referenced objects (and to know
which ones are dangling). The latter checks for the HAS_OBJ
flag in obj->flags to see if we found the object in the
first pass.
Commit 02976bf85 (fsck: introduce `git fsck --connectivity-only`,
2015-06-22) taught fsck to skip the initial pass, and to
fallback to has_sha1_file() instead of the HAS_OBJ check.
However, it converted only one HAS_OBJ check to use
has_sha1_file(). But there are many other places in
builtin/fsck.c that assume that the flag is set (or that
lookup_object() will return an object at all). This leads to
several bugs with --connectivity-only:
1. mark_object() will not queue objects for examination,
so recursively following links from commits to trees,
etc, did nothing. I.e., we were checking the
reachability of hardly anything at all.
2. When a set of heads is given on the command-line, we
use lookup_object() to see if they exist. But without
the initial pass, we assume nothing exists.
3. When loading reflog entries, we do a similar
lookup_object() check, and complain that the reflog is
broken if the object doesn't exist in our hash.
So in short, --connectivity-only is broken pretty badly, and
will claim that your repository is fine when it's not.
Presumably nobody noticed for a few reasons.
One is that the embedded test does not actually test the
recursive nature of the reachability check. All of the
missing objects are still in the index, and we directly
check items from the index. This patch modifies the test to
delete the index, which shows off breakage (1).
Another is that --connectivity-only just skips the initial
pass for loose objects. So on a real repository, the packed
objects were still checked correctly. But on the flipside,
it means that "git fsck --connectivity-only" still checks
the sha1 of all of the packed objects, nullifying its
original purpose of being a faster git-fsck.
And of course the final problem is that the bug only shows
up when there _is_ corruption, which is rare. So anybody
running "git fsck --connectivity-only" proactively would
assume it was being thorough, when it was not.
One possibility for fixing this is to find all of the spots
that rely on HAS_OBJ and tweak them for the connectivity-only
case. But besides the risk that we might miss a spot (and I
found three already, corresponding to the three bugs above),
there are other parts of fsck that _can't_ work without a
full list of objects. E.g., the list of dangling objects.
Instead, let's make the connectivity-only case look more
like the normal case. Rather than skip the initial pass
completely, we'll do an abbreviated one that sets up the
HAS_OBJ flag for each object, without actually loading the
object data.
That's simple and fast, and we don't have to care about the
connectivity_only flag in the rest of the code at all.
While we're at it, let's make sure we treat loose and packed
objects the same (i.e., setting up dummy objects for both
and skipping the actual sha1 check). That makes the
connectivity-only check actually fast on a real repo (40
seconds versus 180 seconds on my copy of linux.git).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-17 22:32:57 +01:00
|
|
|
test_expect_success 'fsck --connectivity-only with explicit head' '
|
|
|
|
|
rm -rf connectivity-only &&
|
|
|
|
|
git init connectivity-only &&
|
|
|
|
|
(
|
|
|
|
|
cd connectivity-only &&
|
|
|
|
|
test_commit foo &&
|
|
|
|
|
rm -f .git/index &&
|
|
|
|
|
tree=$(git rev-parse HEAD^{tree}) &&
|
|
|
|
|
remove_object $(git rev-parse HEAD:foo.t) &&
|
|
|
|
|
test_must_fail git fsck --connectivity-only $tree
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2016-07-17 13:00:02 +02:00
|
|
|
test_expect_success 'fsck --name-objects' '
|
|
|
|
|
rm -rf name-objects &&
|
|
|
|
|
git init name-objects &&
|
|
|
|
|
(
|
|
|
|
|
cd name-objects &&
|
2021-02-10 19:01:30 +01:00
|
|
|
git config core.logAllRefUpdates false &&
|
2016-07-17 13:00:02 +02:00
|
|
|
test_commit julius caesar.t &&
|
2021-02-10 19:01:30 +01:00
|
|
|
test_commit augustus44 &&
|
|
|
|
|
test_commit caesar &&
|
t1450: refactor loose-object removal
Commit 90cf590f5 (fsck: optionally show more helpful info
for broken links, 2016-07-17) added a remove_loose_object()
helper, but we already had a remove_object() helper that did
the same thing. Let's combine these into one.
The implementations had a few subtle differences, so I've
tried to take the best of both:
- the original used "sed", but the newer version avoids
spawning an extra process
- the original processed "$*", which was nonsense, as it
assumed only a single sha1. Use "$1" to make that more
clear.
- the newer version ran an extra rev-parse, but it was not
necessary; it's sole caller already converted the
argument into a raw sha1
- the original used "rm -f", whereas the new one uses
"rm". The latter is better because it may notice a bug
or other unexpected failure in the test. (The original
does check that the object exists before we remove it,
which is good, but that's a subset of the possible
unexpected conditions).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-13 18:54:10 +01:00
|
|
|
remove_object $(git rev-parse julius:caesar.t) &&
|
2016-07-17 13:00:02 +02:00
|
|
|
tree=$(git rev-parse --verify julius:) &&
|
2021-02-10 19:01:30 +01:00
|
|
|
git tag -d julius &&
|
|
|
|
|
test_must_fail git fsck --name-objects >out &&
|
|
|
|
|
test_i18ngrep "$tree (refs/tags/augustus44\\^:" out
|
2016-07-17 13:00:02 +02:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2017-01-13 18:54:39 +01:00
|
|
|
test_expect_success 'alternate objects are correctly blamed' '
|
|
|
|
|
test_when_finished "rm -rf alt.git .git/objects/info/alternates" &&
|
2019-06-29 00:59:21 +02:00
|
|
|
name=$(test_oid numeric) &&
|
|
|
|
|
path=$(test_oid_to_path "$name") &&
|
2017-01-13 18:54:39 +01:00
|
|
|
git init --bare alt.git &&
|
|
|
|
|
echo "../../alt.git/objects" >.git/objects/info/alternates &&
|
2019-06-29 00:59:21 +02:00
|
|
|
mkdir alt.git/objects/$(dirname $path) &&
|
|
|
|
|
>alt.git/objects/$(dirname $path)/$(basename $path) &&
|
2017-01-13 18:54:39 +01:00
|
|
|
test_must_fail git fsck >out 2>&1 &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep alt.git out
|
2017-01-13 18:54:39 +01:00
|
|
|
'
|
|
|
|
|
|
2017-01-13 18:55:55 +01:00
|
|
|
test_expect_success 'fsck errors in packed objects' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
sed "s/</one/" basis >one &&
|
|
|
|
|
sed "s/</foo/" basis >two &&
|
|
|
|
|
one=$(git hash-object -t commit -w one) &&
|
|
|
|
|
two=$(git hash-object -t commit -w two) &&
|
|
|
|
|
pack=$(
|
|
|
|
|
{
|
|
|
|
|
echo $one &&
|
|
|
|
|
echo $two
|
|
|
|
|
} | git pack-objects .git/objects/pack/pack
|
|
|
|
|
) &&
|
|
|
|
|
test_when_finished "rm -f .git/objects/pack/pack-$pack.*" &&
|
|
|
|
|
remove_object $one &&
|
|
|
|
|
remove_object $two &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
2018-11-10 06:16:15 +01:00
|
|
|
test_i18ngrep "error in commit $one.* - bad name" out &&
|
|
|
|
|
test_i18ngrep "error in commit $two.* - bad name" out &&
|
2017-01-13 18:55:55 +01:00
|
|
|
! grep corrupt out
|
|
|
|
|
'
|
|
|
|
|
|
2017-07-28 22:08:02 +02:00
|
|
|
test_expect_success 'fsck fails on corrupt packfile' '
|
|
|
|
|
hsh=$(git commit-tree -m mycommit HEAD^{tree}) &&
|
|
|
|
|
pack=$(echo $hsh | git pack-objects .git/objects/pack/pack) &&
|
|
|
|
|
|
|
|
|
|
# Corrupt the first byte of the first object. (It contains 3 type bits,
|
|
|
|
|
# at least one of which is not zero, so setting the first byte to 0 is
|
|
|
|
|
# sufficient.)
|
|
|
|
|
chmod a+w .git/objects/pack/pack-$pack.pack &&
|
2020-08-02 00:06:11 +02:00
|
|
|
printf "\0" | dd of=.git/objects/pack/pack-$pack.pack bs=1 conv=notrunc seek=12 &&
|
2017-07-28 22:08:02 +02:00
|
|
|
|
|
|
|
|
test_when_finished "rm -f .git/objects/pack/pack-$pack.*" &&
|
|
|
|
|
remove_object $hsh &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
test_i18ngrep "checksum mismatch" out
|
|
|
|
|
'
|
|
|
|
|
|
fsck: parse loose object paths directly
When we iterate over the list of loose objects to check, we
get the actual path of each object. But we then throw it
away and pass just the sha1 to fsck_sha1(), which will do a
fresh lookup. Usually it would find the same object, but it
may not if an object exists both as a loose and a packed
object. We may end up checking the packed object twice, and
never look at the loose one.
In practice this isn't too terrible, because if fsck doesn't
complain, it means you have at least one good copy. But
since the point of fsck is to look for corruption, we should
be thorough.
The new read_loose_object() interface can help us get the
data from disk, and then we replace parse_object() with
parse_object_buffer(). As a bonus, our error messages now
mention the path to a corrupted object, which should make it
easier to track down errors when they do happen.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-13 18:59:44 +01:00
|
|
|
test_expect_success 'fsck finds problems in duplicate loose objects' '
|
|
|
|
|
rm -rf broken-duplicate &&
|
|
|
|
|
git init broken-duplicate &&
|
|
|
|
|
(
|
|
|
|
|
cd broken-duplicate &&
|
|
|
|
|
test_commit duplicate &&
|
|
|
|
|
# no "-d" here, so we end up with duplicates
|
|
|
|
|
git repack &&
|
|
|
|
|
# now corrupt the loose copy
|
|
|
|
|
file=$(sha1_file "$(git rev-parse HEAD)") &&
|
|
|
|
|
rm "$file" &&
|
|
|
|
|
echo broken >"$file" &&
|
|
|
|
|
test_must_fail git fsck
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2017-01-13 19:00:25 +01:00
|
|
|
test_expect_success 'fsck detects trailing loose garbage (commit)' '
|
|
|
|
|
git cat-file commit HEAD >basis &&
|
|
|
|
|
echo bump-commit-sha1 >>basis &&
|
|
|
|
|
commit=$(git hash-object -w -t commit basis) &&
|
|
|
|
|
file=$(sha1_file $commit) &&
|
|
|
|
|
test_when_finished "remove_object $commit" &&
|
|
|
|
|
chmod +w "$file" &&
|
|
|
|
|
echo garbage >>"$file" &&
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
test_i18ngrep "garbage.*$commit" out
|
|
|
|
|
'
|
|
|
|
|
|
2018-10-31 00:18:51 +01:00
|
|
|
test_expect_success 'fsck detects trailing loose garbage (large blob)' '
|
2017-01-13 19:00:25 +01:00
|
|
|
blob=$(echo trailing | git hash-object -w --stdin) &&
|
|
|
|
|
file=$(sha1_file $blob) &&
|
|
|
|
|
test_when_finished "remove_object $blob" &&
|
|
|
|
|
chmod +w "$file" &&
|
|
|
|
|
echo garbage >>"$file" &&
|
2018-10-31 00:18:51 +01:00
|
|
|
test_must_fail git -c core.bigfilethreshold=5 fsck 2>out &&
|
2017-01-13 19:00:25 +01:00
|
|
|
test_i18ngrep "garbage.*$blob" out
|
|
|
|
|
'
|
|
|
|
|
|
check_stream_sha1(): handle input underflow
This commit fixes an infinite loop when fscking large
truncated loose objects.
The check_stream_sha1() function takes an mmap'd loose
object buffer and streams 4k of output at a time, checking
its sha1. The loop quits when we've output enough bytes (we
know the size from the object header), or when zlib tells us
anything except Z_OK or Z_BUF_ERROR.
The latter is expected because zlib may run out of room in
our 4k buffer, and that is how it tells us to process the
output and loop again.
But Z_BUF_ERROR also covers another case: one in which zlib
cannot make forward progress because it needs more _input_.
This should never happen in this loop, because though we're
streaming the output, we have the entire deflated input
available in the mmap'd buffer. But since we don't check
this case, we'll just loop infinitely if we do see a
truncated object, thinking that zlib is asking for more
output space.
It's tempting to fix this by checking stream->avail_in as
part of the loop condition (and quitting if all of our bytes
have been consumed). But that assumes that once zlib has
consumed the input, there is nothing left to do. That's not
necessarily the case: it may have read our input into its
internal state, but still have bytes to output.
Instead, let's continue on Z_BUF_ERROR only when we see the
case we're expecting: the previous round filled our output
buffer completely. If it didn't (and we still saw
Z_BUF_ERROR), we know something is wrong and should break
out of the loop.
The bug comes from commit f6371f9210 (sha1_file: add
read_loose_object() function, 2017-01-13), which
reimplemented some of the existing loose object functions.
So it's worth checking if this bug was inherited from any of
those. The answers seems to be no. The two obvious
candidates are both OK:
1. unpack_sha1_rest(); this doesn't need to loop on
Z_BUF_ERROR at all, since it allocates the expected
output buffer in advance (which we can't do since we're
explicitly streaming here)
2. check_object_signature(); the streaming path relies on
the istream interface, which uses read_istream_loose()
for this case. That function uses a similar "is our
output buffer full" check with Z_BUF_ERROR (which is
where I stole it from for this patch!)
Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-10-31 00:23:12 +01:00
|
|
|
test_expect_success 'fsck detects truncated loose object' '
|
|
|
|
|
# make it big enough that we know we will truncate in the data
|
|
|
|
|
# portion, not the header
|
2018-10-31 05:12:12 +01:00
|
|
|
test-tool genrandom truncate 4096 >file &&
|
check_stream_sha1(): handle input underflow
This commit fixes an infinite loop when fscking large
truncated loose objects.
The check_stream_sha1() function takes an mmap'd loose
object buffer and streams 4k of output at a time, checking
its sha1. The loop quits when we've output enough bytes (we
know the size from the object header), or when zlib tells us
anything except Z_OK or Z_BUF_ERROR.
The latter is expected because zlib may run out of room in
our 4k buffer, and that is how it tells us to process the
output and loop again.
But Z_BUF_ERROR also covers another case: one in which zlib
cannot make forward progress because it needs more _input_.
This should never happen in this loop, because though we're
streaming the output, we have the entire deflated input
available in the mmap'd buffer. But since we don't check
this case, we'll just loop infinitely if we do see a
truncated object, thinking that zlib is asking for more
output space.
It's tempting to fix this by checking stream->avail_in as
part of the loop condition (and quitting if all of our bytes
have been consumed). But that assumes that once zlib has
consumed the input, there is nothing left to do. That's not
necessarily the case: it may have read our input into its
internal state, but still have bytes to output.
Instead, let's continue on Z_BUF_ERROR only when we see the
case we're expecting: the previous round filled our output
buffer completely. If it didn't (and we still saw
Z_BUF_ERROR), we know something is wrong and should break
out of the loop.
The bug comes from commit f6371f9210 (sha1_file: add
read_loose_object() function, 2017-01-13), which
reimplemented some of the existing loose object functions.
So it's worth checking if this bug was inherited from any of
those. The answers seems to be no. The two obvious
candidates are both OK:
1. unpack_sha1_rest(); this doesn't need to loop on
Z_BUF_ERROR at all, since it allocates the expected
output buffer in advance (which we can't do since we're
explicitly streaming here)
2. check_object_signature(); the streaming path relies on
the istream interface, which uses read_istream_loose()
for this case. That function uses a similar "is our
output buffer full" check with Z_BUF_ERROR (which is
where I stole it from for this patch!)
Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-10-31 00:23:12 +01:00
|
|
|
blob=$(git hash-object -w file) &&
|
|
|
|
|
file=$(sha1_file $blob) &&
|
|
|
|
|
test_when_finished "remove_object $blob" &&
|
|
|
|
|
test_copy_bytes 1024 <"$file" >tmp &&
|
|
|
|
|
rm "$file" &&
|
|
|
|
|
mv -f tmp "$file" &&
|
|
|
|
|
|
|
|
|
|
# check both regular and streaming code paths
|
|
|
|
|
test_must_fail git fsck 2>out &&
|
|
|
|
|
test_i18ngrep corrupt.*$blob out &&
|
|
|
|
|
|
|
|
|
|
test_must_fail git -c core.bigfilethreshold=128 fsck 2>out &&
|
|
|
|
|
test_i18ngrep corrupt.*$blob out
|
|
|
|
|
'
|
|
|
|
|
|
2017-01-16 22:25:35 +01:00
|
|
|
# for each of type, we have one version which is referenced by another object
|
|
|
|
|
# (and so while unreachable, not dangling), and another variant which really is
|
|
|
|
|
# dangling.
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
test_expect_success 'create dangling-object repository' '
|
2017-01-16 22:25:35 +01:00
|
|
|
git init dangling &&
|
|
|
|
|
(
|
|
|
|
|
cd dangling &&
|
|
|
|
|
blob=$(echo not-dangling | git hash-object -w --stdin) &&
|
|
|
|
|
dblob=$(echo dangling | git hash-object -w --stdin) &&
|
|
|
|
|
tree=$(printf "100644 blob %s\t%s\n" $blob one | git mktree) &&
|
|
|
|
|
dtree=$(printf "100644 blob %s\t%s\n" $blob two | git mktree) &&
|
|
|
|
|
commit=$(git commit-tree $tree) &&
|
|
|
|
|
dcommit=$(git commit-tree -p $commit $tree) &&
|
|
|
|
|
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
cat >expect <<-EOF
|
2017-01-16 22:25:35 +01:00
|
|
|
dangling blob $dblob
|
|
|
|
|
dangling commit $dcommit
|
|
|
|
|
dangling tree $dtree
|
|
|
|
|
EOF
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
)
|
|
|
|
|
'
|
2017-01-16 22:25:35 +01:00
|
|
|
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
test_expect_success 'fsck notices dangling objects' '
|
|
|
|
|
(
|
|
|
|
|
cd dangling &&
|
2017-01-16 22:25:35 +01:00
|
|
|
git fsck >actual &&
|
|
|
|
|
# the output order is non-deterministic, as it comes from a hash
|
|
|
|
|
sort <actual >actual.sorted &&
|
2021-02-11 02:53:53 +01:00
|
|
|
test_cmp expect actual.sorted
|
2017-01-16 22:25:35 +01:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
test_expect_success 'fsck --connectivity-only notices dangling objects' '
|
|
|
|
|
(
|
|
|
|
|
cd dangling &&
|
|
|
|
|
git fsck --connectivity-only >actual &&
|
|
|
|
|
# the output order is non-deterministic, as it comes from a hash
|
|
|
|
|
sort <actual >actual.sorted &&
|
2021-02-11 02:53:53 +01:00
|
|
|
test_cmp expect actual.sorted
|
fsck: always compute USED flags for unreachable objects
The --connectivity-only option avoids opening every object, and instead
just marks reachable objects with a flag and compares this to the set
of all objects. This strategy is discussed in more detail in 3e3f8bd608
(fsck: prepare dummy objects for --connectivity-check, 2017-01-17).
This means that we report _every_ unreachable object as dangling.
Whereas in a full fsck, we'd have actually opened and parsed each of
those unreachable objects, marking their child objects with the USED
flag, to mean "this was mentioned by another object". And thus we can
report only the tip of an unreachable segment of the object graph as
dangling.
You can see this difference with a trivial example:
tree=$(git hash-object -t tree -w /dev/null)
one=$(echo one | git commit-tree $tree)
two=$(echo two | git commit-tree -p $one $tree)
Running `git fsck` will report only $two as dangling, but with
--connectivity-only, both commits (and the tree) are reported. Likewise,
using --lost-found would write all three objects.
We can make --connectivity-only work like the normal case by taking a
separate pass over the unreachable objects, parsing them and marking
objects they refer to as USED. That still avoids parsing any blobs,
though we do pay the cost to access any unreachable commits and trees
(which may or may not be noticeable, depending on how many you have).
If neither --dangling nor --lost-found is in effect, then we can skip
this step entirely, just like we do now. That makes "--connectivity-only
--no-dangling" just as fast as the current "--connectivity-only". I.e.,
we do the correct thing always, but you can still tweak the options to
make it faster if you don't care about dangling objects.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-03-05 05:47:39 +01:00
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2017-01-16 22:33:29 +01:00
|
|
|
test_expect_success 'fsck $name notices bogus $name' '
|
|
|
|
|
test_must_fail git fsck bogus &&
|
2018-05-13 04:24:13 +02:00
|
|
|
test_must_fail git fsck $ZERO_OID
|
2017-01-16 22:33:29 +01:00
|
|
|
'
|
|
|
|
|
|
2017-01-16 22:34:21 +01:00
|
|
|
test_expect_success 'bogus head does not fallback to all heads' '
|
|
|
|
|
# set up a case that will cause a reachability complaint
|
|
|
|
|
echo to-be-deleted >foo &&
|
|
|
|
|
git add foo &&
|
|
|
|
|
blob=$(git rev-parse :foo) &&
|
|
|
|
|
test_when_finished "git rm --cached foo" &&
|
|
|
|
|
remove_object $blob &&
|
2018-05-13 04:24:13 +02:00
|
|
|
test_must_fail git fsck $ZERO_OID >out 2>&1 &&
|
2017-01-16 22:34:21 +01:00
|
|
|
! grep $blob out
|
|
|
|
|
'
|
|
|
|
|
|
2017-04-25 20:41:09 +02:00
|
|
|
# Corrupt the checksum on the index.
|
|
|
|
|
# Add 1 to the last byte in the SHA.
|
|
|
|
|
corrupt_index_checksum () {
|
|
|
|
|
perl -w -e '
|
|
|
|
|
use Fcntl ":seek";
|
|
|
|
|
open my $fh, "+<", ".git/index" or die "open: $!";
|
|
|
|
|
binmode $fh;
|
|
|
|
|
seek $fh, -1, SEEK_END or die "seek: $!";
|
|
|
|
|
read $fh, my $in_byte, 1 or die "read: $!";
|
|
|
|
|
|
|
|
|
|
$in_value = unpack("C", $in_byte);
|
|
|
|
|
$out_value = ($in_value + 1) & 255;
|
|
|
|
|
|
|
|
|
|
$out_byte = pack("C", $out_value);
|
|
|
|
|
|
|
|
|
|
seek $fh, -1, SEEK_END or die "seek: $!";
|
|
|
|
|
print $fh $out_byte;
|
|
|
|
|
close $fh or die "close: $!";
|
|
|
|
|
'
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
# Corrupt the checksum on the index and then
|
|
|
|
|
# verify that only fsck notices.
|
2017-04-14 22:32:21 +02:00
|
|
|
test_expect_success 'detect corrupt index file in fsck' '
|
|
|
|
|
cp .git/index .git/index.backup &&
|
|
|
|
|
test_when_finished "mv .git/index.backup .git/index" &&
|
2017-04-25 20:41:09 +02:00
|
|
|
corrupt_index_checksum &&
|
|
|
|
|
test_must_fail git fsck --cache 2>errors &&
|
2018-11-10 06:16:05 +01:00
|
|
|
test_i18ngrep "bad index file" errors
|
2017-04-14 22:32:21 +02:00
|
|
|
'
|
|
|
|
|
|
2021-10-01 11:16:37 +02:00
|
|
|
test_expect_success 'fsck hard errors on an invalid object type' '
|
|
|
|
|
git init --bare garbage-type &&
|
|
|
|
|
(
|
|
|
|
|
cd garbage-type &&
|
|
|
|
|
|
|
|
|
|
git hash-object --stdin -w -t garbage --literally </dev/null &&
|
|
|
|
|
|
|
|
|
|
cat >err.expect <<-\EOF &&
|
|
|
|
|
fatal: invalid object type
|
|
|
|
|
EOF
|
|
|
|
|
test_must_fail git fsck >out 2>err &&
|
|
|
|
|
test_cmp err.expect err &&
|
|
|
|
|
test_must_be_empty out
|
|
|
|
|
)
|
|
|
|
|
'
|
|
|
|
|
|
2009-01-30 09:33:00 +01:00
|
|
|
test_done
|