[PATCH 05/11] remote-curl: fix parsing of detached SHA256 heads
- From
Patrick Steinhardt <ps@pks.im>
- Date
- Apr 19, 2024, 09:51 UTC
- Message-ID
- <32a429fb60e4a7e80afdbcf9f874d2b2d921badd.1713519789.git.ps@pks.im>
- In-Reply-To
- <cover.1713519789.git.ps@pks.im>
The dumb HTTP transport tries to read the remote HEAD reference by downloading the "HEAD" file and then parsing it via `http_fetch_ref()`. This function will either parse the file as an object ID in case it is exactly `the_hash_algo->hexsz` long, or otherwise it will check whether the reference starts with "ref :" and parse it as a symbolic ref.
This is broken when parsing detached HEADs of a remote SHA256 repository because we never update `the_hash_algo` to the discovered remote object hash. Consequently, `the_hash_algo` will always be the fallback SHA1 hash algorithm, which will cause us to fail parsing HEAD altogteher when it contains a SHA256 object ID.
Fix this issue by setting up `the_hash_algo` via `repo_set_hash_algo()`. While at it, let's make the expected SHA1 fallback explicit in our code, which also addresses an upcoming issue where we are going to remove the SHA1 fallback for `the_hash_algo`.
Signed-off-by: Patrick Steinhardt <ps@pks.im> --- remote-curl.c | 19 ++++++++++++++++++- t/t5550-http-fetch-dumb.sh | 15 +++++++++++++++ 2 files changed, 33 insertions(+), 1 deletion(-)
diff --git a/remote-curl.c b/remote-curl.c index 0b6d7815fd..004b707fdf 100644 --- a/remote-curl.c +++ b/remote-curl.c @@ -266,12 +266,23 @@ static struct ref *parse_git_refs(struct discovery *heads, int for_push) return list; } +/* + * Try to detect the hash algorithm used by the remote repository when using + * the dumb HTTP transport. As dumb transports cannot tell us the object hash + * directly have to derive it from the advertised ref lengths. + */ static const struct git_hash_algo *detect_hash_algo(struct discovery *heads) { const char *p = memchr(heads->buf, '\t', heads->len); int algo; + + /* + * In case the remote has no refs we have no way to reliably determine + * the object hash used by that repository. In that case we simply fall + * back to SHA1, which may or may not be correct. + */ if (!p) - return the_hash_algo; + return &hash_algos[GIT_HASH_SHA1]; algo = hash_algo_by_length((p - heads->buf) / 2); if (algo == GIT_HASH_UNKNOWN) @@ -295,6 +306,12 @@ static struct ref *parse_info_refs(struct discovery *heads) "is this a git repository?", transport_anonymize_url(url.buf)); + /* + * Set the repository's hash algo to whatever we have just detected. + * This ensures that we can correctly parse the remote references. + */ + repo_set_hash_algo(the_repository, hash_algo_by_ptr(options.hash_algo)); + data = heads->buf; start = NULL; mid = data; diff --git a/t/t5550-http-fetch-dumb.sh b/t/t5550-http-fetch-dumb.sh index 4c3b32785d..5f16cbc58d 100755 --- a/t/t5550-http-fetch-dumb.sh +++ b/t/t5550-http-fetch-dumb.sh @@ -55,6 +55,21 @@ test_expect_success 'list refs from outside any repository' ' test_cmp expect actual ' + +test_expect_success 'list detached HEAD from outside any repository' ' + git clone --mirror "$HTTPD_DOCUMENT_ROOT_PATH/repo.git" \ + "$HTTPD_DOCUMENT_ROOT_PATH/repo-detached.git" && + git -C "$HTTPD_DOCUMENT_ROOT_PATH/repo-detached.git" \ + update-ref --no-deref HEAD refs/heads/main && + git -C "$HTTPD_DOCUMENT_ROOT_PATH/repo-detached.git" update-server-info && + cat >expect <<-EOF && + $(git rev-parse main) HEAD + $(git rev-parse main) refs/heads/main + EOF + nongit git ls-remote "$HTTPD_URL/dumb/repo-detached.git" >actual && + test_cmp expect actual +' + test_expect_success 'create password-protected repository' ' mkdir -p "$HTTPD_DOCUMENT_ROOT_PATH/auth/dumb/" && cp -Rf "$HTTPD_DOCUMENT_ROOT_PATH/repo.git" \
-- 2.44.GIT