git-commit-vandalism/t/t2006-checkout-index-basic.sh

105 lines
3.0 KiB
Bash
Raw Normal View History

#!/bin/sh
test_description='basic checkout-index tests
'
leak tests: mark passing SANITIZE=leak tests as leak-free Mark those remaining tests that pass when run under SANITIZE=leak with TEST_PASSES_SANITIZE_LEAK=true, these were either omitted in f346fcb62a0 (Merge branch 'ab/mark-leak-free-tests-even-more', 2021-12-15) and 5a4f8381b68 (Merge branch 'ab/mark-leak-free-tests', 2021-10-25), or have had their memory leaks fixed since then. With this change there's now a a one-to-one mapping between those tests that we have opted-in via "TEST_PASSES_SANITIZE_LEAK=true", and those that pass with the new "check" mode: GIT_TEST_PASSING_SANITIZE_LEAK=check \ GIT_TEST_SANITIZE_LEAK_LOG=true \ make test SANITIZE=leak Note that the "GIT_TEST_SANITIZE_LEAK_LOG=true" is needed due to the edge cases noted in a preceding commit, i.e. in some cases we'd pass the test itself, but still have outstanding leaks due to ignored exit codes. The "GIT_TEST_SANITIZE_LEAK_LOG=true" corrects for that, we're only marking those tests as passing that really don't have any leaks, whether that was reflected in their exit code or not. Note that the change here to "t9100-git-svn-basic.sh" is marking that test as passing under SANITIZE=leak, we're removing a "TEST_FAILS_SANITIZE_LEAK=true" line, not "TEST_PASSES_SANITIZE_LEAK=true". See 7a98d9ab00d (revisions API: have release_revisions() release "cmdline", 2022-04-13) for the introduction of that t/lib-git-svn.sh-specific variable. Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-07-28 01:13:41 +02:00
TEST_PASSES_SANITIZE_LEAK=true
. ./test-lib.sh
test_expect_success 'checkout-index --gobbledegook' '
test_expect_code 129 git checkout-index --gobbledegook 2>err &&
test_i18ngrep "[Uu]sage" err
'
test_expect_success 'checkout-index -h in broken repository' '
mkdir broken &&
(
cd broken &&
git init &&
>.git/index &&
test_expect_code 129 git checkout-index -h >usage 2>&1
) &&
test_i18ngrep "[Uu]sage" broken/usage
'
test_expect_success 'checkout-index reports errors (cmdline)' '
test_must_fail git checkout-index -- does-not-exist 2>stderr &&
test_i18ngrep not.in.the.cache stderr
'
test_expect_success 'checkout-index reports errors (stdin)' '
echo does-not-exist |
test_must_fail git checkout-index --stdin 2>stderr &&
test_i18ngrep not.in.the.cache stderr
'
checkout: fix bug that makes checkout follow symlinks in leading path Before checking out a file, we have to confirm that all of its leading components are real existing directories. And to reduce the number of lstat() calls in this process, we cache the last leading path known to contain only directories. However, when a path collision occurs (e.g. when checking out case-sensitive files in case-insensitive file systems), a cached path might have its file type changed on disk, leaving the cache on an invalid state. Normally, this doesn't bring any bad consequences as we usually check out files in index order, and therefore, by the time the cached path becomes outdated, we no longer need it anyway (because all files in that directory would have already been written). But, there are some users of the checkout machinery that do not always follow the index order. In particular: checkout-index writes the paths in the same order that they appear on the CLI (or stdin); and the delayed checkout feature -- used when a long-running filter process replies with "status=delayed" -- postpones the checkout of some entries, thus modifying the checkout order. When we have to check out an out-of-order entry and the lstat() cache is invalid (due to a previous path collision), checkout_entry() may end up using the invalid data and thrusting that the leading components are real directories when, in reality, they are not. In the best case scenario, where the directory was replaced by a regular file, the user will get an error: "fatal: unable to create file 'foo/bar': Not a directory". But if the directory was replaced by a symlink, checkout could actually end up following the symlink and writing the file at a wrong place, even outside the repository. Since delayed checkout is affected by this bug, it could be used by an attacker to write arbitrary files during the clone of a maliciously crafted repository. Some candidate solutions considered were to disable the lstat() cache during unordered checkouts or sort the entries before passing them to the checkout machinery. But both ideas include some performance penalty and they don't future-proof the code against new unordered use cases. Instead, we now manually reset the lstat cache whenever we successfully remove a directory. Note: We are not even checking whether the directory was the same as the lstat cache points to because we might face a scenario where the paths refer to the same location but differ due to case folding, precomposed UTF-8 issues, or the presence of `..` components in the path. Two regression tests, with case-collisions and utf8-collisions, are also added for both checkout-index and delayed checkout. Note: to make the previously mentioned clone attack unfeasible, it would be sufficient to reset the lstat cache only after the remove_subtree() call inside checkout_entry(). This is the place where we would remove a directory whose path collides with the path of another entry that we are currently trying to check out (possibly a symlink). However, in the interest of a thorough fix that does not leave Git open to similar-but-not-identical attack vectors, we decided to intercept all `rmdir()` calls in one fell swoop. This addresses CVE-2021-21300. Co-authored-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
2020-12-10 14:27:55 +01:00
for mode in 'case' 'utf-8'
do
case "$mode" in
case) dir='A' symlink='a' mode_prereq='CASE_INSENSITIVE_FS' ;;
utf-8)
dir=$(printf "\141\314\210") symlink=$(printf "\303\244")
mode_prereq='UTF8_NFD_TO_NFC' ;;
esac
test_expect_success SYMLINKS,$mode_prereq \
"checkout-index with $mode-collision don't write to the wrong place" '
git init $mode-collision &&
(
cd $mode-collision &&
mkdir target-dir &&
empty_obj_hex=$(git hash-object -w --stdin </dev/null) &&
symlink_hex=$(printf "%s" "$PWD/target-dir" | git hash-object -w --stdin) &&
cat >objs <<-EOF &&
100644 blob ${empty_obj_hex} ${dir}/x
100644 blob ${empty_obj_hex} ${dir}/y
100644 blob ${empty_obj_hex} ${dir}/z
120000 blob ${symlink_hex} ${symlink}
EOF
git update-index --index-info <objs &&
# Note: the order is important here to exercise the
# case where the file at ${dir} has its type changed by
# the time Git tries to check out ${dir}/z.
#
# Also, we use core.precomposeUnicode=false because we
# want Git to treat the UTF-8 paths transparently on
# Mac OS, matching what is in the index.
#
git -c core.precomposeUnicode=false checkout-index -f \
${dir}/x ${dir}/y ${symlink} ${dir}/z &&
# Should not create ${dir}/z at ${symlink}/z
test_path_is_missing target-dir/z
)
'
done
test_expect_success 'checkout-index --temp correctly reports error on missing blobs' '
test_when_finished git reset --hard &&
missing_blob=$(echo "no such blob here" | git hash-object --stdin) &&
cat >objs <<-EOF &&
100644 $missing_blob file
120000 $missing_blob symlink
EOF
git update-index --index-info <objs &&
test_must_fail git checkout-index --temp symlink file 2>stderr &&
test_i18ngrep "unable to read sha1 file of file ($missing_blob)" stderr &&
test_i18ngrep "unable to read sha1 file of symlink ($missing_blob)" stderr
'
test_expect_success 'checkout-index --temp correctly reports error for submodules' '
git init sub &&
test_commit -C sub file &&
git submodule add ./sub &&
git commit -m sub &&
test_must_fail git checkout-index --temp sub 2>stderr &&
test_i18ngrep "cannot create temporary submodule sub" stderr
'
test_done