1
0
Fork 0
mirror of https://github.com/git/git.git synced 2024-10-31 06:17:56 +01:00
git/t/t1450-fsck.sh

762 lines
22 KiB
Bash
Raw Normal View History

#!/bin/sh
test_description='git fsck random collection of tests
* (HEAD) B
* (master) A
'
. ./test-lib.sh
test_expect_success setup '
git config gc.auto 0 &&
git config i18n.commitencoding ISO-8859-1 &&
test_commit A fileA one &&
git config --unset i18n.commitencoding &&
git checkout HEAD^0 &&
test_commit B fileB two &&
git tag -d A B &&
git reflog expire --expire=now --all &&
>empty
'
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 &&
git fsck --no-dangling >../actual 2>&1
) &&
test_cmp empty actual
'
test_expect_success 'HEAD is part of refs, valid objects appear valid' '
git fsck >actual 2>&1 &&
test_cmp empty actual
'
# Corruption tests follow. Make sure to remove all traces of the
# specific corruption you test afterwards, lest a later test trip over
# it.
test_expect_success 'setup: helpers for corruption tests' '
sha1_file() {
remainder=${1#??} &&
firsttwo=${1%$remainder} &&
echo ".git/objects/$firsttwo/$remainder"
} &&
remove_object() {
rm "$(sha1_file "$1")"
}
'
test_expect_success 'object with bad sha1' '
sha=$(echo blob | git hash-object -w --stdin) &&
old=$(echo $sha | sed "s+^..+&/+") &&
new=$(dirname $old)/ffffffffffffffffffffffffffffffffffffff &&
sha="$(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 $cmt &&
test_when_finished "git update-ref -d refs/heads/bogus" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "$sha.*corrupt" out
'
test_expect_success 'branch pointing to non-commit' '
git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
test_when_finished "git update-ref -d refs/heads/invalid" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "not a commit" out
'
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 &&
echo 0000000000000000000000000000000000000000 >.git/HEAD &&
# avoid corrupt/broken HEAD from interfering with repo discovery
test_must_fail env GIT_DIR=.git git fsck 2>out &&
cat out &&
grep "detached HEAD points" out
'
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 &&
cat out &&
grep "HEAD points to something strange" out
'
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) &&
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 $new" out
'
test_expect_success 'email with embedded > is not okay' '
git cat-file commit HEAD >basis &&
sed "s/@[a-z]/&>/" basis >bad-email &&
new=$(git hash-object -t commit -w --stdin <bad-email) &&
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 &&
cat out &&
grep "error in commit $new" out
'
test_expect_success 'missing < email delimiter is reported nicely' '
git cat-file commit HEAD >basis &&
sed "s/<//" basis >bad-email-2 &&
new=$(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" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "error in commit $new.* - bad name" out
'
test_expect_success 'missing email is reported nicely' '
git cat-file commit HEAD >basis &&
sed "s/[a-z]* <[^>]*>//" basis >bad-email-3 &&
new=$(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" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "error in commit $new.* - missing email" out
'
test_expect_success '> in name is reported' '
git cat-file commit HEAD >basis &&
sed "s/ </> </" basis >bad-email-4 &&
new=$(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" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "error in commit $new" out
'
# date is 2^64 + 1
test_expect_success 'integer overflow in timestamps is reported' '
git cat-file commit HEAD >basis &&
sed "s/^\\(author .*>\\) [0-9]*/\\1 18446744073709551617/" \
<basis >bad-timestamp &&
new=$(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" &&
test_must_fail git fsck 2>out &&
cat out &&
grep "error in commit $new.*integer overflow" out
'
test_expect_success 'commit with NUL in header' '
git cat-file commit HEAD >basis &&
sed "s/author ./author Q/" <basis | q_to_nul >commit-NUL-header &&
new=$(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 &&
cat out &&
grep "error in commit $new.*unterminated header: NUL at offset" out
'
test_expect_success 'tree object with duplicate entries' '
test_when_finished "for i in \$T; do remove_object \$i; done" &&
T=$(
GIT_INDEX_FILE=test-index &&
export GIT_INDEX_FILE &&
rm -f test-index &&
>x &&
git add x &&
git rev-parse :x &&
T=$(git write-tree) &&
echo $T &&
(
git cat-file tree $T &&
git cat-file tree $T
) |
git hash-object -w -t tree --stdin
) &&
test_must_fail git fsck 2>out &&
grep "error in tree .*contains duplicate file entries" out
'
test_expect_success 'unparseable tree object' '
test_when_finished "git update-ref -d refs/heads/wrong" &&
test_when_finished "remove_object \$tree_sha1" &&
test_when_finished "remove_object \$commit_sha1" &&
tree_sha1=$(printf "100644 \0twenty-bytes-of-junk" | git hash-object -t tree --stdin -w --literally) &&
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
'
hex2oct() {
perl -ne 'printf "\\%03o", hex for /../g'
}
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 $commit &&
test_must_fail git fsck >out 2>&1 &&
test_i18ngrep "is a blob, not a tree" out &&
test_i18ngrep ! "dangling blob" out
'
test_expect_success 'tag pointing to nonexistent' '
cat >invalid-tag <<-\EOF &&
object ffffffffffffffffffffffffffffffffffffffff
type commit
tag invalid
tagger T A Gger <tagger@example.com> 1234567890 -0000
This is an invalid tag.
EOF
tag=$(git hash-object -t tag -w --stdin <invalid-tag) &&
test_when_finished "remove_object $tag" &&
echo $tag >.git/refs/tags/invalid &&
test_when_finished "git update-ref -d refs/tags/invalid" &&
test_must_fail git fsck --tags >out &&
cat out &&
grep "broken link" out
'
test_expect_success 'tag pointing to something else than its type' '
sha=$(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 $tag >.git/refs/tags/wrong &&
test_when_finished "git update-ref -d refs/tags/wrong" &&
test_must_fail git fsck --tags
'
test_expect_success 'tag with incorrect tag name & missing tagger' '
sha=$(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 &&
cat >expect <<-EOF &&
warning in tag $tag: badTagName: invalid '\''tag'\'' name: wrong name format
warning in tag $tag: missingTaggerEntry: invalid format - expected '\''tagger'\'' line
EOF
test_cmp expect out
'
test_expect_success 'tag with bad tagger' '
sha=$(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 $tag >.git/refs/tags/wrong &&
test_when_finished "git update-ref -d refs/tags/wrong" &&
test_must_fail git fsck --tags 2>out &&
grep "error in tag .*: invalid author/committer" out
'
test_expect_success 'tag with NUL in header' '
sha=$(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 &&
cat out &&
grep "error in tag $tag.*unterminated header: NUL at offset" out
'
test_expect_success 'cleaned up' '
git fsck >actual 2>&1 &&
test_cmp empty actual
'
test_expect_success 'rev-list --verify-objects' '
git rev-list --verify-objects --all >/dev/null 2>out &&
test_cmp empty out
'
test_expect_success 'rev-list --verify-objects with bad sha1' '
sha=$(echo blob | git hash-object -w --stdin) &&
old=$(echo $sha | sed "s+^..+&/+") &&
new=$(dirname $old)/ffffffffffffffffffffffffffffffffffffff &&
sha="$(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 $cmt &&
test_when_finished "git update-ref -d refs/heads/bogus" &&
test_might_fail git rev-list --verify-objects refs/heads/bogus >/dev/null 2>out &&
cat out &&
grep -q "error: sha1 mismatch 63ffffffffffffffffffffffffffffffffffffff" out
'
test_expect_success 'force fsck to ignore double author' '
git cat-file commit HEAD >basis &&
sed "s/^author .*/&,&/" <basis | tr , \\n >multiple-authors &&
new=$(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
'
_bz='\0'
_bz5="$_bz$_bz$_bz$_bz$_bz"
_bz20="$_bz5$_bz5$_bz5$_bz5"
test_expect_success 'fsck notices blob entry pointing to null sha1' '
(git init null-blob &&
cd null-blob &&
sha=$(printf "100644 file$_bz$_bz20" |
git hash-object -w --stdin -t tree) &&
git fsck 2>out &&
cat out &&
grep "warning.*null sha1" out
)
'
test_expect_success 'fsck notices submodule entry pointing to null sha1' '
(git init null-commit &&
cd null-commit &&
sha=$(printf "160000 submodule$_bz$_bz20" |
git hash-object -w --stdin -t tree) &&
git fsck 2>out &&
cat out &&
grep "warning.*null sha1" out
)
'
while read name path pretty; do
while read mode type; do
: ${pretty:=$path}
test_expect_success "fsck notices $pretty as $type" '
(
git init $name-$type &&
cd $name-$type &&
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" "$path" >bad &&
bad_tree=$(git mktree <bad) &&
git fsck 2>out &&
cat out &&
grep "warning.*tree $bad_tree" out
)'
done <<-\EOF
100644 blob
040000 tree
EOF
done <<-EOF
dot .
dotdot ..
dotgit .git
dotgit-case .GIT
dotgit-unicode .gI${u200c}T .gI{u200c}T
dotgit-case2 .Git
git-tilde1 git~1
dotgitdot .git.
dot-backslash-case .\\\\.GIT\\\\foobar
dotgit-case-backslash .git\\\\foobar
EOF
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
)
'
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 &&
grep nulInCommit warn.1 &&
git fsck 2>warn.2 &&
grep nulInCommit warn.2
)
'
# 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
'
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.
empty=.git/objects/e6/9de29bb2d1d6434b8b29ae775ad8c2e48c5391 &&
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) &&
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
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
)
'
test_expect_success 'fsck --name-objects' '
rm -rf name-objects &&
git init name-objects &&
(
cd name-objects &&
test_commit julius caesar.t &&
test_commit augustus &&
test_commit caesar &&
remove_object $(git rev-parse julius:caesar.t) &&
test_must_fail git fsck --name-objects >out &&
tree=$(git rev-parse --verify julius:) &&
egrep "$tree \((refs/heads/master|HEAD)@\{[0-9]*\}:" out
)
'
test_expect_success 'alternate objects are correctly blamed' '
test_when_finished "rm -rf alt.git .git/objects/info/alternates" &&
git init --bare alt.git &&
echo "../../alt.git/objects" >.git/objects/info/alternates &&
mkdir alt.git/objects/12 &&
>alt.git/objects/12/34567890123456789012345678901234567890 &&
test_must_fail git fsck >out 2>&1 &&
grep alt.git out
'
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 &&
grep "error in commit $one.* - bad name" out &&
grep "error in commit $two.* - bad name" out &&
! grep corrupt out
'
test_expect_success 'fsck fails on corrupt packfile' '
hsh=$(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 &&
printf '\0' | dd of=.git/objects/pack/pack-$pack.pack bs=1 conv=notrunc seek=12 &&
test_when_finished "rm -f .git/objects/pack/pack-$pack.*" &&
remove_object $hsh &&
test_must_fail git fsck 2>out &&
test_i18ngrep "checksum mismatch" out
'
test_expect_success 'fsck finds problems in duplicate loose objects' '
rm -rf broken-duplicate &&
git init broken-duplicate &&
(
cd broken-duplicate &&
test_commit duplicate &&
# no "-d" here, so we end up with duplicates
git repack &&
# now corrupt the loose copy
file=$(sha1_file "$(git rev-parse HEAD)") &&
rm "$file" &&
echo broken >"$file" &&
test_must_fail git fsck
)
'
test_expect_success 'fsck detects trailing loose garbage (commit)' '
git cat-file commit HEAD >basis &&
echo bump-commit-sha1 >>basis &&
commit=$(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
'
test_expect_success 'fsck detects trailing loose garbage (blob)' '
blob=$(echo trailing | git hash-object -w --stdin) &&
file=$(sha1_file $blob) &&
test_when_finished "remove_object $blob" &&
chmod +w "$file" &&
echo garbage >>"$file" &&
test_must_fail git fsck 2>out &&
test_i18ngrep "garbage.*$blob" out
'
# 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.
test_expect_success 'fsck notices dangling objects' '
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) &&
cat >expect <<-EOF &&
dangling blob $dblob
dangling commit $dcommit
dangling tree $dtree
EOF
git fsck >actual &&
# the output order is non-deterministic, as it comes from a hash
sort <actual >actual.sorted &&
test_cmp expect actual.sorted
)
'
test_expect_success 'fsck $name notices bogus $name' '
test_must_fail git fsck bogus &&
test_must_fail git fsck $_z40
'
test_expect_success 'bogus head does not fallback to all heads' '
# set up a case that will cause a reachability complaint
echo to-be-deleted >foo &&
git add foo &&
blob=$(git rev-parse :foo) &&
test_when_finished "git rm --cached foo" &&
remove_object $blob &&
test_must_fail git fsck $_z40 >out 2>&1 &&
! grep $blob out
'
# 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.
test_expect_success 'detect corrupt index file in fsck' '
cp .git/index .git/index.backup &&
test_when_finished "mv .git/index.backup .git/index" &&
corrupt_index_checksum &&
test_must_fail git fsck --cache 2>errors &&
grep "bad index file" errors
'
test_done