The bundle-URI subsystem adds support for creation-token heuristics
to help incremental fetches.
* ds/bundle-uri-5:
bundle-uri: test missing bundles with heuristic
bundle-uri: store fetch.bundleCreationToken
fetch: fetch from an external bundle URI
bundle-uri: drop bundle.flag from design doc
clone: set fetch.bundleURI if appropriate
bundle-uri: download in creationToken order
bundle-uri: parse bundle.<id>.creationToken values
bundle-uri: parse bundle.heuristic=creationToken
t5558: add tests for creationToken heuristic
bundle: verify using check_connected()
bundle: test unbundling with incomplete history
When Git verifies a bundle to see if it is safe for unbundling, it first
looks to see if the prerequisite commits are in the object store. This
is an easy way to "fail fast" but it is not a sufficient check for
updating refs that guarantee closure under reachability. There could
still be issues if those commits are not reachable from the repository's
references. The repository only has guarantees that its object store is
closed under reachability for the objects that are reachable from
references.
Thus, the code in verify_bundle() has previously had the additional
check that all prerequisite commits are reachable from repository
references. This is done via a revision walk from all references,
stopping only if all prerequisite commits are discovered or all commits
are walked. This uses a custom walk to verify_bundle().
This check is more strict than what Git applies to fetched pack-files.
In the fetch case, Git guarantees that the new references are closed
under reachability by walking from the new references until walking
commits that are reachable from repository refs. This is done through
the well-used check_connected() method.
To better align with the restrictions required by 'git fetch',
reimplement this check in verify_bundle() to use check_connected(). This
also simplifies the code significantly.
The previous change added a test that verified the behavior of 'git
bundle verify' and 'git bundle unbundle' in this case, and the error
messages looked like this:
error: Could not read <missing-commit>
fatal: Failed to traverse parents of commit <extant-commit>
However, by changing the revision walk slightly within check_connected()
and using its quiet mode, we can omit those messages. Instead, we get
only this message, tailored to describing the current state of the
repository:
error: some prerequisite commits exist in the object store,
but are not connected to the repository's history
(Line break added here for the commit message formatting, only.)
While this message does not include any object IDs, there is no
guarantee that those object IDs would help the user diagnose what is
going on, as they could be separated from the prerequisite commits by
some distance. At minimum, this situation describes the situation in a
more informative way than the previous error messages.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When verifying a bundle, Git checks first that all prerequisite commits
exist in the object store, then adds an additional check: those
prerequisite commits must be reachable from references in the
repository.
This check is stronger than what is checked for refs being added during
'git fetch', which simply guarantees that the new refs have a complete
history up to the point where it intersects with the current reachable
history.
However, we also do not have any tests that check the behavior under
this condition. Create a test that demonstrates its behavior.
In order to construct a broken history, perform a shallow clone of a
repository with a linear history, but whose default branch ('base') has
a single commit, so dropping the shallow markers leaves a complete
history from that reference. However, the 'tip' reference adds a
shallow commit whose parent is missing in the cloned repository. Trying
to unbundle a bundle with the 'tip' as a prerequisite will succeed past
the object store check and move into the reachability check.
The two errors that are reported are of this form:
error: Could not read <missing-commit>
fatal: Failed to traverse parents of commit <present-commit>
These messages are not particularly helpful for the person running the
unbundle command, but they do prevent the command from succeeding.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since aef7d75e58 (builtin/bundle.c: let parse-options parse
subcommands, 2022-08-19) we've been segfaulting if no argument was
provided.
The fix is easy, as all of the "git bundle" subcommands require a
non-option argument we can check that we have arguments left after
calling parse-options().
This makes use of code added in 73c3253d75 (bundle: framework for
options before bundle file, 2019-11-10), before this change that code
has always been unreachable. In 73c3253d75 we'd never reach it as we
already checked "argc < 2" in cmd_bundle() itself.
Then when aef7d75e58 (whose segfault we're fixing here) migrated this
code to the subcommand API it removed that "argc < 2" check, but we
were still checking the wrong "argc" in parse_options_cmd_bundle(), we
need to check the "newargc". The "argc" will always be >= 1, as it
will necessarily contain at least the subcommand name
itself (e.g. "create").
As an aside, this could be safely squashed into this, but let's not do
that for this minimal segfault fix, as it's an unrelated refactoring:
--- a/builtin/bundle.c
+++ b/builtin/bundle.c
@@ -55,13 +55,12 @@ static int parse_options_cmd_bundle(int argc,
const char * const usagestr[],
const struct option options[],
char **bundle_file) {
- int newargc;
- newargc = parse_options(argc, argv, NULL, options, usagestr,
+ argc = parse_options(argc, argv, NULL, options, usagestr,
PARSE_OPT_STOP_AT_NON_OPTION);
- if (!newargc)
+ if (!argc)
usage_with_options(usagestr, options);
*bundle_file = prefix_filename(prefix, argv[0]);
- return newargc;
+ return argc;
}
static int cmd_bundle_create(int argc, const char **argv, const char *prefix) {
Reported-by: Hubert Jasudowicz <hubertj@stmcyber.pl>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Tested-by: Hubert Jasudowicz <hubertj@stmcyber.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous change moved the 'filter' capability to the end of the 'git
bundle verify' output. Now, add the 'object-format' capability to the
output, when it exists.
This change makes 'git bundle verify' output the hash used in all cases,
even if the capability is not in the bundle. This means that v2 bundles
will always output that they use "sha1". This might look noisy to some
users, but it does simplify the implementation and the test strategy for
this feature.
Since 'verify' ends early when a prerequisite commit is missing, we need
to insert this hash message carefully into our expected test output
throughout t6020.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'filter' capability was added in 105c6f14a (bundle: parse filter
capability, 2022-03-09), but was added in a strange place in the 'git
bundle verify' output.
The tests for this show output like the following:
The bundle contains these 2 refs:
<COMMIT1> <REF1>
<COMMIT2> <REF2>
The bundle uses this filter: blob:none
The bundle records a complete history.
This looks very odd if we have a thin bundle that contains boundary
commits instead of a complete history:
The bundle contains these 2 refs:
<COMMIT1> <REF1>
<COMMIT2> <REF2>
The bundle uses this filter: blob:none
The bundle requires these 2 refs:
<COMMIT3>
<COMMIT4>
This separation between tip refs and boundary refs is unfortunate. Move
the filter capability output to the end of the output. Update the
documentation to match.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Users can create a new repository using 'git clone <bundle-file>'. The
new "@filter" capability for bundles means that we can generate a bundle
that does not contain all reachable objects, even if the header has no
negative commit OIDs.
It is feasible to think that we could make a filtered bundle work with
the command
git clone --filter=$filter --bare <bundle-file>
or possibly replacing --bare with --no-checkout. However, this requires
having some repository-global config that specifies the specified object
filter and notifies Git about the existence of promisor pack-files.
Without a remote, that is currently impossible.
As a stop-gap, parse the bundle header during 'git clone' and die() with
a helpful error message instead of the current behavior of failing due
to "missing objects".
Most of the existing logic for handling bundle clones actually happens
in fetch-pack.c, but that logic is the same as if the user specified
'git fetch <bundle>', so we want to avoid failing to fetch a filtered
bundle when in an existing repository that has the proper config set up
for at least one remote.
Carefully comment around the test that this is not the desired long-term
behavior of 'git clone' in this case, but instead that we need to do
more work before that is possible.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to have a valid pack-file after unbundling a bundle that has
the 'filter' capability, we need to generate a .promisor file. The
bundle does not promise _where_ the objects can be found, but we can
expect that these bundles will be unbundled in repositories with
appropriate promisor remotes that can find those missing objects.
Use the 'git index-pack --promisor=<message>' option to create this
.promisor file. Add "from-bundle" as the message to help anyone diagnose
issues with these promisor packs.
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A previous change allowed Git to parse bundles with the 'filter'
capability. Now, teach Git to create bundles with this option.
Some rearranging of code is required to get the option parsing in the
correct spot. There are now two reasons why we might need capabilities
(a new hash algorithm or an object filter) so that is pulled out into a
place where we can check both at the same time.
The --filter option is parsed as part of setup_revisions(), but it
expected the --objects flag, too. That flag is somewhat implied by 'git
bundle' because it creates a pack-file walking objects, but there is
also a walk that walks the revision range expecting only commits. Make
this parsing work by setting 'revs.tree_objects' and 'revs.blob_objects'
before the call to setup_revisions().
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If a caller to traverse_commit_list() specifies the options for the
--objects flag but does not specify a show_object function pointer, the
result is a segfault. This is currently visible by running 'git bundle
create --objects HEAD'.
We could fix this problem by supplying a no-op callback in
builtin/bundle.c, but that only solves the problem for one builtin,
leaving this segfault open for other callers.
Replace all callers of the show_commit and show_object function pointers
in list-objects.c to call helper functions show_commit() and
show_object() which check that the given context has non-NULL functions
before passing the necessary data. One extra benefit is that it reduces
duplication due to passing ctx->show_data to every caller.
Test that this segfault no longer occurs for 'git bundle'.
Co-authored-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The side-band demultiplexer that is used to display progress output
from the remote end did not clear the line properly when the end of
line hits at a packet boundary, which has been corrected. Also
comes with test clean-ups.
* jx/sideband-cleanup:
test: refactor to use "get_abbrev_oid" to get abbrev oid
test: refactor to use "test_commit" to create commits
test: compare raw output, not mangle tabs and spaces
sideband: don't lose clear-to-eol at packet boundary
Before comparing with the expect file, we used to call function
"make_user_friendly_and_stable_output" to filter out trailing spaces in
output. Ævar recommends using pattern "s/Z$//" to prepare expect file,
and then compare it with raw output.
Since we have fixed the issue of occasionally missing the clear-to-eol
suffix when displaying sideband #2 messages, it is safe and stable to
test against raw output.
Suggested-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ævar reported that the function `make_user_friendly_and_stable_output()`
failed on a i386 box (gcc45) in the gcc farm boxes with error:
sed: couldn't re-allocate memory
It turns out that older versions of bash (4.3) or dash (0.5.7) cannot
evaluate expression like `${A%${A#???????}}` used to get the leading 7
characters of variable A.
Replace the incompatible parameter expansion so that t6020 works on
older version of bash or dash.
Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Removal of GIT_TEST_GETTEXT_POISON continues.
* ab/detox-gettext-tests:
tests: remove most uses of test_i18ncmp
tests: remove last uses of C_LOCALE_OUTPUT
tests: remove most uses of C_LOCALE_OUTPUT
tests: remove last uses of GIT_TEST_GETTEXT_POISON=false
As a follow-up to d162b25f95 (tests: remove support for
GIT_TEST_GETTEXT_POISON, 2021-01-20) remove most uses of test_i18ncmp
via a simple s/test_i18ncmp/test_cmp/g search-replacement.
I'm leaving t6300-for-each-ref.sh out due to a conflict with in-flight
changes between "master" and "seen", as well as the prerequisite
itself due to other changes between "master" and "next/seen" which add
new test_i18ncmp uses.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rename the recently introduced test-bundle-functions.sh to be
consistent with other lib-*.sh files, which is the convention for
these sorts of shared test library functions.
The new test-bundle-functions.sh was introduced in 9901164d81 (test:
add helper functions for git-bundle, 2021-01-11). It was the only
test-*.sh of this nature.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to create an incremental bundle, we need to pass many arguments
to let git-bundle ignore some already packed commits. It will be more
convenient to pass args via stdin. But the current implementation does
not allow us to do this.
This is because args are parsed twice when creating bundle. The first
time for parsing args is in `compute_and_write_prerequisites()` by
running `git-rev-list` command to write prerequisites in bundle file,
and stdin is consumed in this step if "--stdin" option is provided for
`git-bundle`. Later nothing can be read from stdin when running
`setup_revisions()` in `create_bundle()`.
The solution is to parse args once by removing the entire function
`compute_and_write_prerequisites()` and then calling function
`setup_revisions()`. In order to write prerequisites for bundle, will
call `prepare_revision_walk()` and `traverse_commit_list()`. But after
calling `prepare_revision_walk()`, the object array `revs.pending` is
left empty, and the following steps could not work properly with the
empty object array (`revs.pending`). Therefore, make a copy of `revs`
to `revs_copy` for later use right after calling `setup_revisions()`.
The copy of `revs_copy` is not a deep copy, it shares the same objects
with `revs`. The object array of `revs` has been cleared, but objects
themselves are still kept. Flags of objects may change after calling
`prepare_revision_walk()`, we can use these changed flags without
calling the `git rev-list` command and parsing its output like the
former implementation.
Also add testcases for git bundle in t6020, which read args from stdin.
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
`git rev-list` will list one commit for the following command:
$ git rev-list 'main^!'
<tip-commit-of-main-branch>
But providing the same rev-list args to `git bundle`, fail to create
a bundle file.
$ git bundle create - 'main^!'
# v2 git bundle
-<OID> <one-line-message>
fatal: Refusing to create empty bundle.
This is because when removing duplicate objects in function
`object_array_remove_duplicates()`, one unique pending object which has
the same name is deleted by mistake. The revision arg 'main^!' in the
above example is parsed by `handle_revision_arg()`, and at lease two
different objects will be appended to `revs.pending`, one points to the
parent commit of the "main" branch, and the other points to the tip
commit of the "main" branch. These two objects have the same name
"main". Only one object is left with the name "main" after calling the
function `object_array_remove_duplicates()`.
And what's worse, when adding boundary commits into pending list, we use
one-line commit message as names, and the arbitory names may surprise
git-bundle.
Only comparing objects themselves (".item") is also not good enough,
because user may want to create a bundle with two identical objects but
with different reference names, such as: "HEAD" and "refs/heads/main".
Add new function `contains_object()` which compare both the address and
the name of the object.
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move git-bundle related functions from t5510 to a library, and this lib
will be shared with a new testcase t6020 which finds a known breakage of
"git-bundle".
Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
