fetch: use bundle URIs when having creationToken heuristic
PATCH 1/1 fetch: use bundle URIs when having creationToken heuristic
At the moment, bundle URIs are only used on git-clone(1). For a clone the use of bundle URI is trivial, because repository is empty so downloading bundles will never result in downloading objects that are in the repository already.
For git-fetch(1), this more complicated to use bundle URI. We want to avoid downloading bundles that only contains objects that are in the local repository already.
One way to achieve this is possible when the "creationToken" heuristic is used for bundle URIs. With this heuristic, the server sends along a creationToken for each bundle. When a client downloads bundles with such creationToken, the highest known value is written to fetch.bundleCreationToken in the git-config. The next time bundles are advertised by the server, bundles with a lower creationToken value are ignored. This was already implemented by 7903efb717 (bundle-uri:
download in creationToken order, 2023-01-31) in fetch_bundles_by_token().
Using the creationToken heuristic is optional, but without it the client has no idea which bundles are new and which only have objects the client already has.
With this knowledge, make git-fetch(1) get bundle URI information from the server, but only use bundle URIs if the creationToken heuristic is used.
The code added in builtin/fetch.c is very similar to the code in builtin/clone.c, so while at it make sure we always clean up the bundles list advertised by the server.
--- builtin/clone.c | 13 +++++----- builtin/fetch.c | 17 +++++++++++++ t/t5584-fetch-bundle-uri.sh | 49 +++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+), 6 deletions(-) create mode 100755 t/t5584-fetch-bundle-uri.sh
builtin/clone.c +7 -6
@@ -1406,9 +1406,9 @@ int cmd_clone(int argc, const char **argv, const char *prefix)git_config_set_gently ("fetch.bundleuri" , bundle_uri); }else {/* -* Populate transport->got_remote_bundle_uri and -* transport->bundle_uri. We might get nothing. -*/ +* Populate transport->bundles. We might get nothing or fail +* trying, but clone can continue without bundles as well. +*/ transport_get_remote_bundle_uri (transport); if (transport->bundles &&
@@ -1419,10 +1419,11 @@ int cmd_clone(int argc, const char **argv, const char *prefix)else if (fetch_bundle_list (the_repository, transport->bundles))warning (_ ("failed to fetch advertised bundles" )); - }else { -clear_bundle_list (transport->bundles); -FREE_AND_NULL (transport->bundles); } + +clear_bundle_list (transport->bundles); +if (transport->bundles) +FREE_AND_NULL (transport->bundles); }if (refs)
builtin/fetch.c +17 -0
@@ -1694,6 +1694,23 @@ static int do_fetch(struct transport *transport, retcode =1 ; } +/* +* Populate transport->bundles. We might get nothing or fail +* trying, but fetch can continue without bundles as well. +*/ +transport_get_remote_bundle_uri (transport); + +if (transport->bundles && +hashmap_get_size (&transport->bundles->bundles) && + (transport->bundles->heuristic ==BUNDLE_HEURISTIC_CREATIONTOKEN )) { +if (fetch_bundle_list (the_repository, transport->bundles)) +warning (_ ("failed to fetch advertised bundles" )); + } + +clear_bundle_list (transport->bundles); +if (transport->bundles) +FREE_AND_NULL (transport->bundles); +if (fetch_and_consume_refs (&display_state, transport, transaction, ref_map, &fetch_head, config)) { retcode =1 ;
t/t5584-fetch-bundle-uri.sh new file +49 -0
49 lines not on this page · raw message
--
2.45.2
Quoted 19 lines · Toon Claes
At the moment, bundle URIs are only used on git-clone(1). For a clone the use of bundle URI is trivial, because repository is empty so downloading bundles will never result in downloading objects that are in the repository already.
For git-fetch(1), this more complicated to use bundle URI. We want to avoid downloading bundles that only contains objects that are in the local repository already.
One way to achieve this is possible when the "creationToken" heuristic is used for bundle URIs. With this heuristic, the server sends along a creationToken for each bundle. When a client downloads bundles with such creationToken, the highest known value is written to
fetch.bundleCreationTokenin the git-config. The next time bundles are advertised by the server, bundles with a lower creationToken value are ignored. This was already implemented by7903efb717(bundle-uri:
download in creationToken order, 2023-01-31) infetch_bundles_by_token().
Quoted 3 lines
Using the creationToken heuristic is optional, but without it the client has no idea which bundles are new and which only have objects the client already has.
The above explains that you can see which bundle is newer than which other ones, but ...
Quoted 3 lines
With this knowledge, make git-fetch(1) get bundle URI information from the server, but only use bundle URIs if the creationToken heuristic is used.
... it does not explain how it helps for the repository that runs "fetch" to tell which bundles do not give any new objects it does not have. "clone" starts from an emptyness, so for it it is enough if it can figure out among N bundles the other end has, which ones are sufficient to cover the entire set of objects the other end has, but "fetch" needs to be able to exclude a lot more---the objects this end has need to be excluded". The logic to do so may need to be explained a bit better than the above to make it more convincing.
On the other hand...
Quoted 3 lines
The code added in
builtin/fetch.cis very similar to the code inbuiltin/clone.c, so while at it make sure we always clean up the bundles list advertised by the server.
... I do understand that clean-up code. Please make it a separate preliminary clean-up patch before the main thing.
If the "clone" and "fetch" codepath uses very similar logic, can the code in "clone" the preliminary clean-up patch would fix be refactored in another separate patch on top of the clean-up patch to a helper function that can be called from outside? And then can "fetch" be taught, on top of those two preliminary patches, to call that helper function to implement what this single patch wanted to do?
Thanks.
Quoted 125 lines
Signed-off-by Toon Claes <[email protected]>--- builtin/clone.c | 13 +++++----- builtin/fetch.c | 17 +++++++++++++ t/t5584-fetch-bundle-uri.sh | 49 +++++++++++++++++++++++++++++++++++++ 3 files changed, 73 insertions(+), 6 deletions(-) create mode 100755 t/t5584-fetch-bundle-uri.shbuiltin/clone.c +7
-6@@ -1406,9 +1406,9 @@ int cmd_clone(int argc, const char **argv, const char *prefix)git_config_set_gently ("fetch.bundleuri" , bundle_uri); }else {/* -* Populate transport->got_remote_bundle_uri and -* transport->bundle_uri. We might get nothing. -*/ +* Populate transport->bundles. We might get nothing or fail +* trying, but clone can continue without bundles as well. +*/ transport_get_remote_bundle_uri (transport); if (transport->bundles &&@@ -1419,10 +1419,11 @@ int cmd_clone(int argc, const char **argv, const char *prefix)else if (fetch_bundle_list (the_repository, transport->bundles))warning (_ ("failed to fetch advertised bundles" )); - }else { -clear_bundle_list (transport->bundles); -FREE_AND_NULL (transport->bundles); } + +clear_bundle_list (transport->bundles); +if (transport->bundles) +FREE_AND_NULL (transport->bundles); }if (refs)builtin/fetch.c +17
-0@@ -1694,6 +1694,23 @@ static int do_fetch(struct transport *transport, retcode =1 ; } +/* +* Populate transport->bundles. We might get nothing or fail +* trying, but fetch can continue without bundles as well. +*/ +transport_get_remote_bundle_uri (transport); + +if (transport->bundles && +hashmap_get_size (&transport->bundles->bundles) && + (transport->bundles->heuristic ==BUNDLE_HEURISTIC_CREATIONTOKEN )) { +if (fetch_bundle_list (the_repository, transport->bundles)) +warning (_ ("failed to fetch advertised bundles" )); + } + +clear_bundle_list (transport->bundles); +if (transport->bundles) +FREE_AND_NULL (transport->bundles); +if (fetch_and_consume_refs (&display_state, transport, transaction, ref_map, &fetch_head, config)) { retcode =1 ;t/t5584-fetch-bundle-uri.sh new file +49
-1@@ -0,0 +1,49 @@ +#!/bin/sh + +test_description='test use of bundle URI in "git fetch"' + +GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main +export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME + +. ./test-lib.sh + +test_expect_success 'set up repos and bundles' ' + git init source && + test_commit -C source A && + git clone --no-local source go-A-to-C && + test_commit -C source B && + git clone --no-local source go-B-to-C && + git clone --no-local source go-B-to-D && + git -C source bundle create B.bundle main && + test_commit -C source C && + git -C source bundle create B-to-C.bundle B..main && + git -C source config uploadpack.advertiseBundleURIs true && + git -C source config bundle.version 1 && + git -C source config bundle.mode all && + git -C source config bundle.heuristic creationToken && + git -C source config bundle.bundle-B.uri "file://$(pwd)/source/B.bundle" && + git -C source config bundle.bundle-B.creationToken 1 && + git -C source config bundle.bundle-B-to-C.uri "file://$(pwd)/source/B-to-C.bundle" && + git -C source config bundle.bundle-B-to-C.creationToken 2 +' + +test_expect_success 'fetches one bundle URI to get up-to-date' ' + git -C go-B-to-C -c transfer.bundleURI=true fetch origin && + test 1 = $(ls go-B-to-C/.git/objects/bundles | wc -l) && + test 2 = $(git -C go-B-to-C config fetch.bundleCreationToken) +' + +test_expect_success 'fetches two bundle URIs to get up-to-date' ' + git -C go-A-to-C -c transfer.bundleURI=true fetch origin && + test 2 = $(ls go-A-to-C/.git/objects/bundles | wc -l) && + test 2 = $(git -C go-A-to-C config fetch.bundleCreationToken) +' + +test_expect_success 'fetches one bundle URI and objects from remote' ' + test_commit -C source D && + git -C go-B-to-D -c transfer.bundleURI=true fetch origin && + test 1 = $(ls go-B-to-D/.git/objects/bundles | wc -l) && + test 2 = $(git -C go-B-to-D config fetch.bundleCreationToken) +' + +test_done --2.45.2
v2 cover fetch: use bundle URIs when having creationToken heuristic
In reply to Toon ClaesThis is version 2 of teaching git-fetch(1) to use bundle URIs. For this second revision I've split up the changes in multliple commits as Junio suggested.
- The first patch cleans up some unneccessary code.
- The second patch refactor the code to make it possible to reuse.
- The third patch makes git-fetch(1) use bundle URIs when creationToken heuristic is used. This time I've added some more information about how bundle URIs are handled when they have a creationToken heuristic.
Toon Claes (3):
clone: remove double bundle list clear code
transport: introduce transport_has_remote_bundle_uri()
fetch: use bundle URIs when having creationToken heuristic
builtin/clone.c | 26 ++++++-------------- builtin/fetch.c | 13 ++++++++++ t/helper/test-bundle-uri.c | 2 +- t/t5558-clone-bundle-uri.sh | 28 ++++++++++++++++++++- t/t5584-fetch-bundle-uri.sh | 49 +++++++++++++++++++++++++++++++++++++ transport.c | 14 ++++++++++- transport.h | 7 +++--- 7 files changed, 114 insertions(+), 25 deletions(-) create mode 100755 t/t5584-fetch-bundle-uri.sh
Range-diff against v1:
-: ---------- > 1: fb556a76d3 clone: remove double bundle list clear code
1: 00f6017ab0 ! 2: 0f7c53bbe1 fetch: use bundle URIs when having creationToken heuristic
@@ Metadata
Author: Toon Claes <[email protected]>
## Commit message ##
- fetch: use bundle URIs when having creationToken heuristic
+ transport: introduce transport_has_remote_bundle_uri()
- At the moment, bundle URIs are only used on git-clone(1). For a clone
- the use of bundle URI is trivial, because repository is empty so
- downloading bundles will never result in downloading objects that are in
- the repository already.
+ The public function transport_get_remote_bundle_uri() exists to fetch
+ the bundle URI(s) from the remote. This function is only called from
+ builtin/clone.c (not taking test-tool into account). There it ignores
+ the return value, because it doesn't matter whether the server didn't
+ return any bundles or if it failed trying to fetch them, clone can
+ continue without bundle URIs. After calling it, it checks if anything
+ is collected in the bundle list and starts fetching them.
- For git-fetch(1), this more complicated to use bundle URI. We want to
- avoid downloading bundles that only contains objects that are in the
- local repository already.
-
- One way to achieve this is possible when the "creationToken" heuristic
- is used for bundle URIs. With this heuristic, the server sends along a
- creationToken for each bundle. When a client downloads bundles with such
- creationToken, the highest known value is written to
- `fetch.bundleCreationToken` in the git-config. The next time bundles are
- advertised by the server, bundles with a lower creationToken value are
- ignored. This was already implemented by 7903efb717 (bundle-uri:
- download in creationToken order, 2023-01-31) in
- fetch_bundles_by_token().
-
- Using the creationToken heuristic is optional, but without it the
- client has no idea which bundles are new and which only have objects the
- client already has.
-
- With this knowledge, make git-fetch(1) get bundle URI information from
- the server, but only use bundle URIs if the creationToken heuristic is
- used.
-
- The code added in builtin/fetch.c is very similar to the code in
- builtin/clone.c, so while at it make sure we always clean up the bundles
- list advertised by the server.
+ Add public function transport_has_remote_bundle_uri() instead. This
+ calls the (now made private) transport_get_remote_bundle_uri() function
+ and returns whether any bundle URI is received. This makes reuse of the
+ code easier and avoids code duplication when we add bundle URI support
+ to git-fetch(1).
Signed-off-by: Toon Claes <[email protected]>
## builtin/clone.c ##
@@ builtin/clone.c: int cmd_clone(int argc, const char **argv, const char *prefix)
+ bundle_uri);
+ else if (has_heuristic)
git_config_set_gently("fetch.bundleuri", bundle_uri);
- } else {
- /*
+- } else {
+- /*
- * Populate transport->got_remote_bundle_uri and
- * transport->bundle_uri. We might get nothing.
- */
-+ * Populate transport->bundles. We might get nothing or fail
-+ * trying, but clone can continue without bundles as well.
-+ */
- transport_get_remote_bundle_uri(transport);
-
- if (transport->bundles &&
-@@ builtin/clone.c: int cmd_clone(int argc, const char **argv, const char *prefix)
- else if (fetch_bundle_list(the_repository,
- transport->bundles))
- warning(_("failed to fetch advertised bundles"));
-- } else {
-- clear_bundle_list(transport->bundles);
-- FREE_AND_NULL(transport->bundles);
- }
-+
-+ clear_bundle_list(transport->bundles);
-+ if (transport->bundles)
-+ FREE_AND_NULL(transport->bundles);
+- transport_get_remote_bundle_uri(transport);
+-
+- if (transport->bundles &&
+- hashmap_get_size(&transport->bundles->bundles)) {
+- /* At this point, we need the_repository to match the cloned repo. */
+- if (repo_init(the_repository, git_dir, work_tree))
+- warning(_("failed to initialize the repo, skipping bundle URI"));
+- else if (fetch_bundle_list(the_repository,
+- transport->bundles))
+- warning(_("failed to fetch advertised bundles"));
+- }
++ } else if (transport_has_remote_bundle_uri(transport)) {
++ /* At this point, we need the_repository to match the cloned repo. */
++ if (repo_init(the_repository, git_dir, work_tree))
++ warning(_("failed to initialize the repo, skipping bundle URI"));
++ else if (fetch_bundle_list(the_repository,
++ transport->bundles))
++ warning(_("failed to fetch advertised bundles"));
}
if (refs)
- ## builtin/fetch.c ##
-@@ builtin/fetch.c: static int do_fetch(struct transport *transport,
- retcode = 1;
+ ## t/helper/test-bundle-uri.c ##
+@@ t/helper/test-bundle-uri.c: static int cmd_ls_remote(int argc, const char **argv)
}
-+ /*
-+ * Populate transport->bundles. We might get nothing or fail
-+ * trying, but fetch can continue without bundles as well.
-+ */
-+ transport_get_remote_bundle_uri(transport);
-+
-+ if (transport->bundles &&
-+ hashmap_get_size(&transport->bundles->bundles) &&
-+ (transport->bundles->heuristic == BUNDLE_HEURISTIC_CREATIONTOKEN)) {
-+ if (fetch_bundle_list(the_repository, transport->bundles))
-+ warning(_("failed to fetch advertised bundles"));
-+ }
-+
-+ clear_bundle_list(transport->bundles);
-+ if (transport->bundles)
-+ FREE_AND_NULL(transport->bundles);
-+
- if (fetch_and_consume_refs(&display_state, transport, transaction, ref_map,
- &fetch_head, config)) {
- retcode = 1;
+ transport = transport_get(remote, NULL);
+- if (transport_get_remote_bundle_uri(transport) < 0) {
++ if (!transport_has_remote_bundle_uri(transport)) {
+ error(_("could not get the bundle-uri list"));
+ status = 1;
+ goto cleanup;
- ## t/t5584-fetch-bundle-uri.sh (new) ##
-@@
-+#!/bin/sh
-+
-+test_description='test use of bundle URI in "git fetch"'
-+
-+GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main
-+export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME
-+
-+. ./test-lib.sh
-+
-+test_expect_success 'set up repos and bundles' '
-+ git init source &&
-+ test_commit -C source A &&
-+ git clone --no-local source go-A-to-C &&
-+ test_commit -C source B &&
-+ git clone --no-local source go-B-to-C &&
-+ git clone --no-local source go-B-to-D &&
-+ git -C source bundle create B.bundle main &&
-+ test_commit -C source C &&
-+ git -C source bundle create B-to-C.bundle B..main &&
-+ git -C source config uploadpack.advertiseBundleURIs true &&
-+ git -C source config bundle.version 1 &&
-+ git -C source config bundle.mode all &&
-+ git -C source config bundle.heuristic creationToken &&
-+ git -C source config bundle.bundle-B.uri "file://$(pwd)/source/B.bundle" &&
-+ git -C source config bundle.bundle-B.creationToken 1 &&
-+ git -C source config bundle.bundle-B-to-C.uri "file://$(pwd)/source/B-to-C.bundle" &&
-+ git -C source config bundle.bundle-B-to-C.creationToken 2
-+'
+ ## transport.c ##
+@@ transport.c: int transport_fetch_refs(struct transport *transport, struct ref *refs)
+ return rc;
+ }
+
+-int transport_get_remote_bundle_uri(struct transport *transport)
++static int transport_get_remote_bundle_uri(struct transport *transport)
+ {
+ int value = 0;
+ const struct transport_vtable *vtable = transport->vtable;
+@@ transport.c: int transport_get_remote_bundle_uri(struct transport *transport)
+
+ if (vtable->get_bundle_uri(transport) < 0)
+ return error(_("could not retrieve server-advertised bundle-uri list"));
+
-+test_expect_success 'fetches one bundle URI to get up-to-date' '
-+ git -C go-B-to-C -c transfer.bundleURI=true fetch origin &&
-+ test 1 = $(ls go-B-to-C/.git/objects/bundles | wc -l) &&
-+ test 2 = $(git -C go-B-to-C config fetch.bundleCreationToken)
-+'
++ return 0;
++}
+
-+test_expect_success 'fetches two bundle URIs to get up-to-date' '
-+ git -C go-A-to-C -c transfer.bundleURI=true fetch origin &&
-+ test 2 = $(ls go-A-to-C/.git/objects/bundles | wc -l) &&
-+ test 2 = $(git -C go-A-to-C config fetch.bundleCreationToken)
-+'
++int transport_has_remote_bundle_uri(struct transport *transport)
++{
++ transport_get_remote_bundle_uri(transport);
+
-+test_expect_success 'fetches one bundle URI and objects from remote' '
-+ test_commit -C source D &&
-+ git -C go-B-to-D -c transfer.bundleURI=true fetch origin &&
-+ test 1 = $(ls go-B-to-D/.git/objects/bundles | wc -l) &&
-+ test 2 = $(git -C go-B-to-D config fetch.bundleCreationToken)
-+'
++ if (transport->bundles &&
++ hashmap_get_size(&transport->bundles->bundles))
++ return 1;
+
-+test_done
+ return 0;
+ }
+
+
+ ## transport.h ##
+@@ transport.h: const struct ref *transport_get_remote_refs(struct transport *transport,
+ struct transport_ls_refs_options *transport_options);
+
+ /**
+- * Retrieve bundle URI(s) from a remote. Populates "struct
+- * transport"'s "bundle_uri" and "got_remote_bundle_uri".
++ * Try fetch bundle URI(s) from a remote and returns 1 if one or more
++ * bundle URI(s) are received from the server.
++ * Populates "struct transport"'s "bundles" and "got_remote_bundle_uri".
+ */
+-int transport_get_remote_bundle_uri(struct transport *transport);
++int transport_has_remote_bundle_uri(struct transport *transport);
+
+ /*
+ * Fetch the hash algorithm used by a remote.-: ---------- > 3: db8f303dde fetch: use bundle URIs when having creationToken heuristic --
2.45.2
PATCH v2 1/3 clone: remove double bundle list clear code
The bundle list transport->bundles is filled by transport_get_remote_bundle_uri(). Only when the list is not used, it is cleared right away by calling clear_bundle_list().
This looks like we leak memory allocated for the list when transport->bundles *is* used. But in fact, transport->bundles is cleaned up in transport_disconnect() near the end of cmd_clone().
Remove the double clean up of transport->bundles, and depend solely on transport_disconnect() to take care of it.
Also add a test case that hits this code, but due to other leaks we cannot mark it as leak-free.
--- builtin/clone.c | 3 --- t/t5558-clone-bundle-uri.sh | 28 +++++++++++++++++++++++++++- 2 files changed, 27 insertions(+), 4 deletions(-)
builtin/clone.c +0 -3
@@ -1419,9 +1419,6 @@ int cmd_clone(int argc, const char **argv, const char *prefix)else if (fetch_bundle_list (the_repository, transport->bundles))warning (_ ("failed to fetch advertised bundles" )); - }else { -clear_bundle_list (transport->bundles); -FREE_AND_NULL (transport->bundles); } }
t/t5558-clone-bundle-uri.sh +27 -1
@@ -1,6 +1,6 @@ #!/bin/sh -test_description='test fetching bundles with --bundle-uri' +test_description='test clone with use of bundle-uri' . ./test-lib.sh . "$TEST_DIRECTORY"/lib-bundle.sh
@@ -438,6 +438,32 @@ test_expect_success 'negotiation: bundle list with all wanted commits' ' test_grep ! "clone> want " trace-packet.txt ' +test_expect_success 'bundles advertised by the server' ' + test_when_finished rm -f trace*.txt && + git clone clone-from clone-advertiser && + git -C clone-advertiser config uploadpack.advertiseBundleURIs true && + git -C clone-advertiser config bundle.version 1 && + git -C clone-advertiser config bundle.mode all && + git -C clone-advertiser config bundle.bundle-1.uri "file://$(pwd)/clone-from/bundle-1.bundle" && + git -C clone-advertiser config bundle.bundle-2.uri "file://$(pwd)/clone-from/bundle-2.bundle" && + git -C clone-advertiser config bundle.bundle-3.uri "file://$(pwd)/clone-from/bundle-3.bundle" && + git -C clone-advertiser config bundle.bundle-4.uri "file://$(pwd)/clone-from/bundle-4.bundle" && + + GIT_TRACE_PACKET="$(pwd)/trace-packet.txt" \ + git -c transfer.bundleURI=true clone clone-advertiser clone-advertised && + git -C clone-advertised for-each-ref --format="%(refname)" >refs && + grep "refs/bundles/" refs >actual && + cat >expect <<-\EOF && + refs/bundles/base + refs/bundles/left + refs/bundles/merge + refs/bundles/right + EOF + test_cmp expect actual && + # We already have all needed commits so no "want" needed. + test_grep ! "clone> want " trace-packet.txt +' + ######################################################################### # HTTP tests begin here
--
2.45.2
Re: [PATCH v2 1/3] clone: remove double bundle list clear code
Quoted 25 lines · Toon Claes
The bundle list
transport->bundlesis filled bytransport_get_remote_bundle_uri(). Only when the list is not used, it is cleared right away by callingclear_bundle_list().This looks like we leak memory allocated for the list when
transport->bundles*is* used. But in fact,transport->bundlesis cleaned up intransport_disconnect()near the end ofcmd_clone().Remove the double clean up of
transport->bundles, and depend solely ontransport_disconnect()to take care of it.Also add a test case that hits this code, but due to other leaks we cannot mark it as leak-free.
Signed-off-by Toon Claes <[email protected]>--- builtin/clone.c | 3 --- t/t5558-clone-bundle-uri.sh | 28 +++++++++++++++++++++++++++- 2 files changed, 27 insertions(+), 4 deletions(-)builtin/clone.c +0
-0
builtin/clone.c int cmd_clone(int argc, const char **argv, const char *prefix) · line 1422
Quoted 7 lines
builtin/clone.c +0
-3@@ -1419,9 +1419,6 @@ int cmd_clone(int argc, const char **argv, const char *prefix)else if (fetch_bundle_list (the_repository, transport->bundles))warning (_ ("failed to fetch advertised bundles" )); - }else { -clear_bundle_list (transport->bundles); -FREE_AND_NULL (transport->bundles);
There is a small difference that here we also set transport->bundles to NULL, whereas in transport_disconnect() we only do free(...). I don't see any issues because of it though. This cleanup looks good.
[snip]
Re: [PATCH v2 1/3] clone: remove double bundle list clear code
In reply to Toon ClaesQuoted 7 lines · Toon Claes
The bundle list
transport->bundlesis filled bytransport_get_remote_bundle_uri(). Only when the list is not used, it is cleared right away by callingclear_bundle_list().This looks like we leak memory allocated for the list when
transport->bundles*is* used. But in fact,transport->bundlesis cleaned up intransport_disconnect()near the end ofcmd_clone().
The transport->bundles is setup and initialized by transport_get() and populated by transport_get_remote_bundle_uri(). In its current form, we clean up transport->bundles if it is empty immediately after checking the remote for bundles. This is redundant though because the cleanup already occurs as part of transport_disconnect().
Since transport_disconnect() is already responsible for freeing other parts of transport, it makes sense to let it be the one to handle it.
Quoted 32 lines
Remove the double clean up of
transport->bundles, and depend solely ontransport_disconnect()to take care of it.Also add a test case that hits this code, but due to other leaks we cannot mark it as leak-free.
Signed-off-by Toon Claes <[email protected]>--- builtin/clone.c | 3 --- t/t5558-clone-bundle-uri.sh | 28 +++++++++++++++++++++++++++- 2 files changed, 27 insertions(+), 4 deletions(-)builtin/clone.c +0
-3@@ -1419,9 +1419,6 @@ int cmd_clone(int argc, const char **argv, const char *prefix)else if (fetch_bundle_list (the_repository, transport->bundles))warning (_ ("failed to fetch advertised bundles" )); - }else { -clear_bundle_list (transport->bundles); -FREE_AND_NULL (transport->bundles); } }t/t5558-clone-bundle-uri.sh +0
-0
t/t5558-clone-bundle-uri.sh line 3
Quoted 4 lines
t/t5558-clone-bundle-uri.sh +1
-1@@ -1,6 +1,6 @@ #!/bin/sh -test_description='test fetching bundles with --bundle-uri' +test_description='test clone with use of bundle-uri'
I assume this description was changed because the test pertains to clones and not fetches. Might be worth mentioning in the commit message.
Quoted 23 lines
t/t5558-clone-bundle-uri.sh +16
-0. ./test-lib.sh . "$TEST_DIRECTORY"/lib-bundle.sh@@ -438,6 +438,32 @@ test_expect_success 'negotiation: bundle list with all wanted commits' ' test_grep ! "clone> want " trace-packet.txt ' +test_expect_success 'bundles advertised by the server' ' + test_when_finished rm -f trace*.txt && + git clone clone-from clone-advertiser && + git -C clone-advertiser config uploadpack.advertiseBundleURIs true && + git -C clone-advertiser config bundle.version 1 && + git -C clone-advertiser config bundle.mode all && + git -C clone-advertiser config bundle.bundle-1.uri "file://$(pwd)/clone-from/bundle-1.bundle" && + git -C clone-advertiser config bundle.bundle-2.uri "file://$(pwd)/clone-from/bundle-2.bundle" && + git -C clone-advertiser config bundle.bundle-3.uri "file://$(pwd)/clone-from/bundle-3.bundle" && + git -C clone-advertiser config bundle.bundle-4.uri "file://$(pwd)/clone-from/bundle-4.bundle" && + + GIT_TRACE_PACKET="$(pwd)/trace-packet.txt" \ + git -c transfer.bundleURI=true clone clone-advertiser clone-advertised && + git -C clone-advertised for-each-ref --format="%(refname)" >refs && + grep "refs/bundles/" refs >actual && + cat >expect <<-\EOF &&
t/t5558-clone-bundle-uri.sh test_expect_success 'negotiation: bundle list with all wanted commits' ' · line 466
Quoted 10 lines
t/t5558-clone-bundle-uri.sh +10
-0+ refs/bundles/base + refs/bundles/left + refs/bundles/merge + refs/bundles/right + EOF + test_cmp expect actual && + # We already have all needed commits so no "want" needed. + test_grep ! "clone> want " trace-packet.txt +' +
Looks like this test is validating that a clone is retrieving bundles from a remote repository advertising said bundles. We to do so by checking for the refs/bundles/* references which should only exist if the advertised bundles were fetched. This makes sense to me.
From the commit message though, I thought the added test might have something to do with the changes to when transport->bundles was freed. I wonder if it would be a good idea to expland on this further in the commit message. Or since the added test is somewhat unrelated to this patch, maybe it should be a separate patch?
-Justin
Quoted 6 lines
t/t5558-clone-bundle-uri.sh +0
-1######################################################################### # HTTP tests begin here --2.45.2
Re: [PATCH v2 1/3] clone: remove double bundle list clear code
Quoted 5 lines · Justin Tobler
From the commit message though, I thought the added test might have something to do with the changes to when
transport->bundleswas freed. I wonder if it would be a good idea to expland on this further in the commit message. Or since the added test is somewhat unrelated to this patch, maybe it should be a separate patch?
I added this test case because I wanted to debug the code path where the remote advertises bundle URIs, all other test cases use --bundle-uri. This helped me to debug the code and understand where the transport->bundles are freed. Although this test case does not fully touch the code path of the code that's removed, but it was close enough to hack a few extra lines and understand what happens. Since I wrote that test anyway, I felt I might as well submit it.
Signature · 1 line
Toon
PATCH v2 2/3 transport: introduce transport_has_remote_bundle_uri()
In reply to Toon ClaesThe public function transport_get_remote_bundle_uri() exists to fetch the bundle URI(s) from the remote. This function is only called from builtin/clone.c (not taking test-tool into account). There it ignores the return value, because it doesn't matter whether the server didn't return any bundles or if it failed trying to fetch them, clone can continue without bundle URIs. After calling it, it checks if anything is collected in the bundle list and starts fetching them.
Add public function transport_has_remote_bundle_uri() instead. This calls the (now made private) transport_get_remote_bundle_uri() function and returns whether any bundle URI is received. This makes reuse of the code easier and avoids code duplication when we add bundle URI support to git-fetch(1).
--- builtin/clone.c | 23 +++++++---------------- t/helper/test-bundle-uri.c | 2 +- transport.c | 14 +++++++++++++- transport.h | 7 ++++--- 4 files changed, 25 insertions(+), 21 deletions(-)
builtin/clone.c +7 -16
@@ -1404,22 +1404,13 @@ int cmd_clone(int argc, const char **argv, const char *prefix) bundle_uri); else if (has_heuristic) git_config_set_gently("fetch.bundleuri", bundle_uri); - }else { -/* -* Populate transport->got_remote_bundle_uri and -* transport->bundle_uri. We might get nothing. -*/ -transport_get_remote_bundle_uri (transport); - -if (transport->bundles && -hashmap_get_size (&transport->bundles->bundles)) { -/* At this point, we need the_repository to match the cloned repo. */ -if (repo_init (the_repository, git_dir, work_tree)) -warning (_ ("failed to initialize the repo, skipping bundle URI" )); -else if (fetch_bundle_list (the_repository, - transport->bundles)) -warning (_ ("failed to fetch advertised bundles" )); - } + }else if (transport_has_remote_bundle_uri (transport)) { +/* At this point, we need the_repository to match the cloned repo. */ +if (repo_init (the_repository, git_dir, work_tree)) +warning (_ ("failed to initialize the repo, skipping bundle URI" )); +else if (fetch_bundle_list (the_repository, + transport->bundles)) +warning (_ ("failed to fetch advertised bundles" )); }if (refs)
t/helper/test-bundle-uri.c +1 -1
@@ -90,7 +90,7 @@ static int cmd_ls_remote(int argc, const char **argv) } transport =transport_get (remote,NULL ); -if (transport_get_remote_bundle_uri (transport) <0 ) { +if (!transport_has_remote_bundle_uri (transport)) {error (_ ("could not get the bundle-uri list" )); status =1 ; goto cleanup;
transport.c +13 -1
@@ -1536,7 +1536,7 @@ int transport_fetch_refs(struct transport *transport, struct ref *refs)return rc; } -int transport_get_remote_bundle_uri (struct transport *transport) +static int transport_get_remote_bundle_uri (struct transport *transport) {int value =0 ;const struct transport_vtable *vtable = transport->vtable;
@@ -1561,6 +1561,18 @@ int transport_get_remote_bundle_uri(struct transport *transport)if (vtable->get_bundle_uri (transport) <0 )return error (_ ("could not retrieve server-advertised bundle-uri list" )); + +return 0 ; +} + +int transport_has_remote_bundle_uri (struct transport *transport) +{ +transport_get_remote_bundle_uri (transport); + +if (transport->bundles && +hashmap_get_size (&transport->bundles->bundles)) +return 1 ; +return 0 ; }
transport.h +4 -3
@@ -294,10 +294,11 @@ const struct ref *transport_get_remote_refs(struct transport *transport, struct transport_ls_refs_options *transport_options);/** -* Retrieve bundle URI(s) from a remote. Populates "struct -* transport"'s "bundle_uri" and "got_remote_bundle_uri". +* Try fetch bundle URI(s) from a remote and returns 1 if one or more +* bundle URI(s) are received from the server. +* Populates "struct transport"'s "bundles" and "got_remote_bundle_uri". */ -int transport_get_remote_bundle_uri (struct transport *transport); +int transport_has_remote_bundle_uri (struct transport *transport);/* * Fetch the hash algorithm used by a remote.
--
2.45.2
Re: [PATCH v2 2/3] transport: introduce transport_has_remote_bundle_uri()
Toon Claes <[email protected]> writes:
[snip]
Quoted 18 lines
t/helper/test-bundle-uri.c +1
-1@@ -90,7 +90,7 @@ static int cmd_ls_remote(int argc, const char **argv) } transport =transport_get (remote,NULL ); -if (transport_get_remote_bundle_uri (transport) <0 ) { +if (!transport_has_remote_bundle_uri (transport)) {error (_ ("could not get the bundle-uri list" )); status =1 ; goto cleanup;transport.c +0
-0
transport.c int transport_fetch_refs(struct transport *transport, struct ref *refs) · line 1540
Quoted 6 lines
transport.c +1
-1@@ -1536,7 +1536,7 @@ int transport_fetch_refs(struct transport *transport, struct ref *refs)return rc; } -int transport_get_remote_bundle_uri (struct transport *transport) +static int transport_get_remote_bundle_uri (struct transport *transport)
Why make it static?
Quoted 27 lines
transport.c +0
-0{int value =0 ;transport.c +12
-0const struct transport_vtable *vtable = transport->vtable;@@ -1561,6 +1561,18 @@ int transport_get_remote_bundle_uri(struct transport *transport)if (vtable->get_bundle_uri (transport) <0 )return error (_ ("could not retrieve server-advertised bundle-uri list" )); + +return 0 ; +} + +int transport_has_remote_bundle_uri (struct transport *transport) +{ +transport_get_remote_bundle_uri (transport); + +if (transport->bundles && +hashmap_get_size (&transport->bundles->bundles)) +return 1 ; +return 0 ; }transport.h +0
-0
transport.h const struct ref *transport_get_remote_refs(struct transport *transport, · line 297
Quoted 6 lines
transport.h +1
-2@@ -294,10 +294,11 @@ const struct ref *transport_get_remote_refs(struct transport *transport, struct transport_ls_refs_options *transport_options);/** -* Retrieve bundle URI(s) from a remote. Populates "struct -* transport"'s "bundle_uri" and "got_remote_bundle_uri". +* Try fetch bundle URI(s) from a remote and returns 1 if one or more
Nit: s/Try/Try to/
transport.h const struct ref *transport_get_remote_refs(struct transport *transport, · line 302
Quoted 6 lines
transport.h +3
-1+* bundle URI(s) are received from the server. +* Populates "struct transport"'s "bundles" and "got_remote_bundle_uri". */ -int transport_get_remote_bundle_uri (struct transport *transport); +int transport_has_remote_bundle_uri (struct transport *transport);
Shouldn't this now be renamed to transport_has_bundle_uri? Earlier, we were 'getting' bundle URIs from the remote. Now, we are abstracting that away and simply asking, 'do we have bundle URIs'.
Quoted 4 lines
transport.h +0
-1/* * Fetch the hash algorithm used by a remote. -- 2.45.2
Re: [PATCH v2 2/3] transport: introduce transport_has_remote_bundle_uri()
Quoted 8 lines · Karthik Nayak
Quoted 8 lines
transport.c +0
-0@@ -1536,7 +1536,7 @@ int transport_fetch_refs(struct transport *transport, struct ref *refs)return rc; }
Quoted 5 lines
transport.c int transport_fetch_refs(struct transport *transport, struct ref *refs)
Quoted 3 lines
transport.c +1
-1-int transport_get_remote_bundle_uri (struct transport *transport) +static int transport_get_remote_bundle_uri (struct transport *transport)Why make it static?
The reason is rather well described in the proposed log message, I think.
Quoted 10 lines
transport.c int transport_fetch_refs(struct transport *transport, struct ref *refs)
Quoted 6 lines
transport.c +3
-1+* bundle URI(s) are received from the server. +* Populates "struct transport"'s "bundles" and "got_remote_bundle_uri". */ -int transport_get_remote_bundle_uri (struct transport *transport); +int transport_has_remote_bundle_uri (struct transport *transport);Shouldn't this now be renamed to
transport_has_bundle_uri? Earlier, we were 'getting' bundle URIs from the remote. Now, we are abstracting that away and simply asking, 'do we have bundle URIs'.
A good question.
PATCH v2 3/3 fetch: use bundle URIs when having creationToken heuristic
In reply to Toon ClaesAt the moment, bundle URIs are only used by git-clone(1). For a clone the use of bundle URI is trivial, because the repository is empty so downloading bundles will never result in downloading objects that are in the repository already.
For git-fetch(1), this more complicated to use bundle URI. We want to avoid downloading bundles that only contains objects that are in the local repository already.
One way to achieve this is possible when the "creationToken" heuristic is used for bundle URIs. We attempt to download and unbundle the minimum number of bundles by creationToken in decreasing order. If we fail to unbundle (after a successful download) then move to the next non-downloaded bundle and attempt downloading. Once we succeed in applying a bundle, move to the previous unapplied bundle and attempt to unbundle it again. At the end the highest applied creationToken is written to fetch.bundleCreationToken in the git-config. The next time bundles are advertised by the server, bundles with a lower creationToken value are ignored. This was already implemented by 7903efb717 (bundle-uri: download in creationToken order, 2023-01-31) in fetch_bundles_by_token().
Using the creationToken heuristic is optional, but without it the client has no idea which bundles are new, how to sort them, and which only have objects the client already has.
With this knowledge, make git-fetch(1) use bundle URIs from the server, but only when the creationToken heuristic is used.
--- builtin/fetch.c | 13 ++++++++++ t/t5584-fetch-bundle-uri.sh | 49 +++++++++++++++++++++++++++++++++++++ 2 files changed, 62 insertions(+) create mode 100755 t/t5584-fetch-bundle-uri.sh
builtin/fetch.c +13 -0
@@ -1694,6 +1694,19 @@ static int do_fetch(struct transport *transport, retcode =1 ; } +if (transport_has_remote_bundle_uri (transport)) { +/* +* Only use bundle-URIs when they use the creationToken +* heuristic, this allows us to ensure not downloading bundles +* we don't need. You can read the comments in +* fetch_bundles_by_token() to understand how this works. +*/ +if (transport->bundles->heuristic ==BUNDLE_HEURISTIC_CREATIONTOKEN ) { +if (fetch_bundle_list (the_repository, transport->bundles)) +warning (_ ("failed to fetch advertised bundles" )); + } + } +if (fetch_and_consume_refs (&display_state, transport, transaction, ref_map, &fetch_head, config)) { retcode =1 ;
t/t5584-fetch-bundle-uri.sh new file +49 -0
49 lines not on this page · raw message
--
2.45.2
Quoted 6 lines · Toon Claes
At the moment, bundle URIs are only used by git-clone(1). For a clone the use of bundle URI is trivial, because the repository is empty so downloading bundles will never result in downloading objects that are in the repository already.
For git-fetch(1), this more complicated to use bundle URI. We want to
Nit: s/this/it is/
Quoted 5 lines
avoid downloading bundles that only contains objects that are in the local repository already.
One way to achieve this is possible when the "creationToken" heuristic is used for bundle URIs. We attempt to download and unbundle the minimum
This first sentence reads a bit weird. Perhaps, "One way to achieve this is to restrict the usage of bundle URIs to the 'creationToken' heuristic for git-fetch(1)."?
Quoted 11 lines
number of bundles by creationToken in decreasing order. If we fail to unbundle (after a successful download) then move to the next non-downloaded bundle and attempt downloading. Once we succeed in applying a bundle, move to the previous unapplied bundle and attempt to unbundle it again. At the end the highest applied creationToken is written to
fetch.bundleCreationTokenin the git-config. The next time bundles are advertised by the server, bundles with a lower creationToken value are ignored. This was already implemented by7903efb717(bundle-uri: download in creationToken order, 2023-01-31) infetch_bundles_by_token().
Quoted 7 lines
Using the creationToken heuristic is optional, but without it the client has no idea which bundles are new, how to sort them, and which only have objects the client already has.
With this knowledge, make git-fetch(1) use bundle URIs from the server, but only when the creationToken heuristic is used.
I would say that apart from all this it is also worth noting that bundle URIs are still opt-in from the client side.
Rest of the patch looks good to me.
Thanks
[snip]
Quoted 4 lines · Toon Claes
One way to achieve this is possible when the "creationToken" heuristic is used for bundle URIs. We attempt to download and unbundle the minimum number of bundles by creationToken in decreasing order. If we fail to unbundle (after a successful download) then move to the next
Quoted 8 lines
non-downloaded bundle and attempt downloading. Once we succeed in applying a bundle, move to the previous unapplied bundle and attempt to unbundle it again. At the end the highest applied creationToken is written to
fetch.bundleCreationTokenin the git-config. The next time bundles are advertised by the server, bundles with a lower creationToken value are ignored. This was already implemented by7903efb717(bundle-uri: download in creationToken order, 2023-01-31) infetch_bundles_by_token().
I think Junio essentially asked this already, but I'm still missing the bigger picture here. When the "creationToken" heuristic is applied, the effect of your change is that we'll always favor bundle URIs now over performing proper fetches, right?
Now suppose that the server creates new bundled whenever somebody pushes a new change to the default branch. We do not really have information how this bundle is structured. It _could_ be an incremental bundle, and in that case it might be sensible to fetch that bundle. But it could also be that the server generates a full bundle including all objects transitively reachable from that default branch. Now if we started to rely on the "creationToken" heuristic, we would basically end up re-downloading the complete repository, which is a strict regression.
Now that scenario is of course hypothetical. But the problem is that the strategy for how bundle URIs are generated are determined by the hosting provider. So ultimately, I expect that the reality will lie somewhere in between and be different depending on which hosting solution you use.
All of this to me means that the "creationToken" heuristic is not really a good signal, unless I'm missing something about the way it works. Is there any additional signal provided by the server except for the time when the bundle was created? If so, is that information sufficient to determine whether it makes sense for a client to fetch a bundle instead of performing a "proper" fetch? If not, what is the additional info that we would need to make this schema work properly?
So unless I'm missing something, I feel like we need to think bigger and design a heuristic that gives us the information needed. Without such a heuristic, default-enabling may or may not do the right thing, and we have no way to really argue whether it will do as we now depend on server operators to do the right thing.
Patrick
Quoted 4 lines · Patrick Steinhardt
I think Junio essentially asked this already, but I'm still missing the bigger picture here. When the "creationToken" heuristic is applied, the effect of your change is that we'll always favor bundle URIs now over performing proper fetches, right?
Yes, sort of. Bundle URIs are a step before "proper" fetches. A "proper" fetch might still happen after downloading bundles.
(I would rather call them "negotiated fetches" or something, but let's just stick with "proper" fetch for now)
Quoted 9 lines
Now suppose that the server creates new bundled whenever somebody pushes a new change to the default branch. We do not really have information how this bundle is structured. It _could_ be an incremental bundle, and in that case it might be sensible to fetch that bundle. But it could also be that the server generates a full bundle including all objects transitively reachable from that default branch. Now if we started to rely on the "creationToken" heuristic, we would basically end up re-downloading the complete repository, which is a strict regression.
Quoted 4 lines
Now that scenario is of course hypothetical. But the problem is that the strategy for how bundle URIs are generated are determined by the hosting provider. So ultimately, I expect that the reality will lie somewhere in between and be different depending on which hosting solution you use.
That is true. The mechanism of bundle URIs is mostly outside the control of Git itself. It's up to the Git hosting provider how they use it. This gives them a lot of flexibility, like where to store bundles, how incremental bundles are used, and how often bundles are regenerated. But this also comes with a huge responsibility, due to the scenario you describe above.
At GitLab we want to do an initial roll-out of bundle URIs with just one bundle advertised to the client. This bundle is generated every week, and only contains the whole history of default branch (up to the newest commit). But, we'll advertise this bundle with creationToken "1", always. This will cause the client to fetch any bundle for the repository only once. Even when the bundle is updated by the server, the client will not fetch it because the creationToken did not change.
Quoted 4 lines
All of this to me means that the "creationToken" heuristic is not really a good signal, unless I'm missing something about the way it works. Is there any additional signal provided by the server except for the time when the bundle was created?
The Git hosting provider can use the "creationToken" however they want, at the moment it's up to them to decide on a good strategy.
For example, assume you decide to create bundle URIs only for the default branch, then you can choose to use the committer date of the topmost commit of that branch as the creationToken.
Imagine you have the following example commit history (committer date of some commits are indicated with a caret below those commits):
A --- B --- C --- D --- E --- F --- G
^ ^ ^
2024-08-02 2024-08-04 2024-08-09
Today (on 2024-08-02, at this point commits D to G don't exist yet) the Git host decides to create a bundle:
- git bundle create first.bundle main
The server will advertise this bundle with the creationToken being the Unix timestamp of the topmost commit in the bundle:
- first.bundle = 2024-08-02 or in Unix 1722549600
When the client clones/fetches, it will download this bundle and store creationToken 1722549600 (= 2024-08-02) in the repo config as fetch.bundleCreationToken.
A few days later (on 2024-08-04, now D & E are added) the client fetches again, and there are no new bundles, so a "proper" fetch happens.
A week after the first bundle (on 2024-08-09, D to G were added since) the Git host decides to create a new (incremental) bundle:
- git bundle create second.bundle C..main
The server then advertises the following bundles:
- first.bundle = 2024-08-02 or in Unix 1722549600
- second.bundle = 2024-08-09 1723154400
Now when the client fetches again, it sees a new bundle "second.bundle", and will download it. This bundle contains the history from C..G. And you might have noticed, the client already has C..E, while it only needs E..G. This is a slight inefficiency, it's a pitfall of bundle URI I don't think we can avoid. By design bundle URIs are used to pre-generate files for clients to use, so they are never built-to-order. It's almost impossible to avoid clients to download bundles that have objects they already have.
Well, maybe it is possible. But then the question arises, if the current main branch is at G, do you want the client to download a bundle which has a little bit too much objects, or do you want to have the client do a "proper" fetch and skip any bundle?
But all the above assumes there's only one branch in the bundle. With more branches another strategy might be required.
Quoted 4 lines
If so, is that information sufficient to determine whether it makes sense for a client to fetch a bundle instead of performing a "proper" fetch? If not, what is the additional info that we would need to make this schema work properly?
I think my example above indicates the "creationToken" _can_ be sufficient information for the client to determine if downloading a bundle makes sense, but it depends on a well-though-out strategy of the Git host. So one the hand it gives the host the flexibility of using a strategy, on the other hand it puts a lot of responsibility on them.
Quoted 5 lines
So unless I'm missing something, I feel like we need to think bigger and design a heuristic that gives us the information needed. Without such a heuristic, default-enabling may or may not do the right thing, and we have no way to really argue whether it will do as we now depend on server operators to do the right thing.
I think we're absolutely not ready to default-enabling the bundle URI feature in Git. Here at GitLab we're trying pilot bundle URI to figure out how it should/can work. Roll-out to CI runners is our first target and because we have the flexibility to toggle use of bundle URI on the client side, we can test things safely, without affecting any other user.
Talking about heuristics, any heuristic will be a _hack_ if you ask me. If you think about it, the bundles them self should be the single source of truth. In fact, a bundle is a packfile with a header prepended that contains refs and their OID in git-show-ref(1) format, and includes which prerequisite OIDs it has.
So another approach I think of is to have the client partially download bundles (so it pauses the download once they have received the complete header) and make them parse the header to determine if they continue the download. But this approach feels quite complicated to me.
Luckily, with the bundle.heuristic config we can always introduce new values if we discover there is a flaw in the "creationToken" heuristic.
Signature · 1 line
Toon
Quoted 33 lines · Toon claes
Quoted 4 lines · Patrick Steinhardt
I think Junio essentially asked this already, but I'm still missing the bigger picture here. When the "creationToken" heuristic is applied, the effect of your change is that we'll always favor bundle URIs now over performing proper fetches, right?
Yes, sort of. Bundle URIs are a step before "proper" fetches. A "proper" fetch might still happen after downloading bundles.
(I would rather call them "negotiated fetches" or something, but let's just stick with "proper" fetch for now)Quoted 9 lines
Now suppose that the server creates new bundled whenever somebody pushes a new change to the default branch. We do not really have information how this bundle is structured. It _could_ be an incremental bundle, and in that case it might be sensible to fetch that bundle. But it could also be that the server generates a full bundle including all objects transitively reachable from that default branch. Now if we started to rely on the "creationToken" heuristic, we would basically end up re-downloading the complete repository, which is a strict regression.
Quoted 4 lines
Now that scenario is of course hypothetical. But the problem is that the strategy for how bundle URIs are generated are determined by the hosting provider. So ultimately, I expect that the reality will lie somewhere in between and be different depending on which hosting solution you use.
That is true. The mechanism of bundle URIs is mostly outside the control of Git itself. It's up to the Git hosting provider how they use it. This gives them a lot of flexibility, like where to store bundles, how incremental bundles are used, and how often bundles are regenerated. But this also comes with a huge responsibility, due to the scenario you describe above.
Quoted 7 lines
At GitLab we want to do an initial roll-out of bundle URIs with just one bundle advertised to the client. This bundle is generated every week, and only contains the whole history of default branch (up to the newest commit). But, we'll advertise this bundle with creationToken "1", always. This will cause the client to fetch any bundle for the repository only once. Even when the bundle is updated by the server, the client will not fetch it because the creationToken did not change.
Okay. Are details like this documented anywhere in the Git project? Because when we start to build on the preexisting heuristics I think that we should give server operators a guide for how things come together and what the client's expectations can and should be.
Quoted 56 lines
Quoted 4 lines
All of this to me means that the "creationToken" heuristic is not really a good signal, unless I'm missing something about the way it works. Is there any additional signal provided by the server except for the time when the bundle was created?
The Git hosting provider can use the "creationToken" however they want, at the moment it's up to them to decide on a good strategy.
For example, assume you decide to create bundle URIs only for the default branch, then you can choose to use the committer date of the topmost commit of that branch as the creationToken.
Imagine you have the following example commit history (committer date of some commits are indicated with a caret below those commits):
A --- B --- C --- D --- E --- F --- G
^ ^ ^
2024-08-02 2024-08-04 2024-08-09Today (on 2024-08-02, at this point commits D to G don't exist yet) the Git host decides to create a bundle:
- git bundle create first.bundle main
The server will advertise this bundle with the creationToken being the Unix timestamp of the topmost commit in the bundle:
- first.bundle = 2024-08-02 or in Unix 1722549600
When the client clones/fetches, it will download this bundle and store creationToken 1722549600 (= 2024-08-02) in the repo config as
fetch.bundleCreationToken.A few days later (on 2024-08-04, now D & E are added) the client fetches again, and there are no new bundles, so a "proper" fetch happens.
A week after the first bundle (on 2024-08-09, D to G were added since) the Git host decides to create a new (incremental) bundle:
- git bundle create second.bundle C..main
The server then advertises the following bundles:
- first.bundle = 2024-08-02 or in Unix 1722549600
- second.bundle = 2024-08-09 1723154400Now when the client fetches again, it sees a new bundle "second.bundle", and will download it. This bundle contains the history from C..G. And you might have noticed, the client already has C..E, while it only needs E..G. This is a slight inefficiency, it's a pitfall of bundle URI I don't think we can avoid. By design bundle URIs are used to pre-generate files for clients to use, so they are never built-to-order. It's almost impossible to avoid clients to download bundles that have objects they already have.
Quoted 7 lines
Well, maybe it is possible. But then the question arises, if the current main branch is at G, do you want the client to download a bundle which has a little bit too much objects, or do you want to have the client do a "proper" fetch and skip any bundle?
But all the above assumes there's only one branch in the bundle. With more branches another strategy might be required.
Okay. What I'm still missing though is the ability for a client do discern whether bundles are incremental or full.
In the schema you describe, my assumption is that the server operator will eventually regenerate the first bundle to become a full bundle. Which means that if there was any activity in the repository, then the Unix timestamp of that regenerated bundle will likely be bigger than what the client has stored on disk. Consequently, without information that this is in fact a full bundle, the client would end up fetching it. This is a potentially big waste of time and resources.
Quoted 10 lines
Quoted 4 lines
If so, is that information sufficient to determine whether it makes sense for a client to fetch a bundle instead of performing a "proper" fetch? If not, what is the additional info that we would need to make this schema work properly?
I think my example above indicates the "creationToken" _can_ be sufficient information for the client to determine if downloading a bundle makes sense, but it depends on a well-though-out strategy of the Git host. So one the hand it gives the host the flexibility of using a strategy, on the other hand it puts a lot of responsibility on them.
It does put a big burden on the hoster to get it right, that much is true. But I think the bigger problem is that the hoster _cannot_ get it right, because they have no ability to distribute enough information to the client.
For a schema like this to work I think we need to advertise at least two things for each bundle: a unique bundle token, and the token of the base bundle that a bundle may refer to. This would allow us to advertise dependencies between bundles to a client rather easily: a full bundle does not have a base, whereas incremental bundles would point to the previous bundle.
A theoretical flow would thus go like this:
1. The server announces the full bundle A.bundle. The client downloads it and stores the unique token Tok(A) in its config.
2. Subsequent fetches notice that Tok(A) announced by the server and stored in the config remain the same.
3. The server creates incremental bundles B.bundle and C.bundle. B.bundle builds on top of A, C builds on top of B. The server announces all three of those bundles. The client determines that it only has A, so to get to C it needs to fetch both B and C. It then stores Tok(C) in its config.
One scenario not yet covered is when the server regenerates the full bundle to get A' = A + B + C + D. Ideally, the server would temporarily announce a new bundle D = C + D while also announcing that A' and the result from D are equivalent. Like this, the client knows that it does not have to refetch A', but already has its contents and can thus update its config to point to Tok(A').
But even a schema that is more involved like this one is rather suboptimal. Eventually, the equivalence of Tok(A') and D will not be advertised by the server anymore. So the client will not be able to learn about which incremental bundles to fetch anymore, and would be forced to re-download everything.
Quoted 13 lines
Quoted 5 lines
So unless I'm missing something, I feel like we need to think bigger and design a heuristic that gives us the information needed. Without such a heuristic, default-enabling may or may not do the right thing, and we have no way to really argue whether it will do as we now depend on server operators to do the right thing.
I think we're absolutely not ready to default-enabling the bundle URI feature in Git. Here at GitLab we're trying pilot bundle URI to figure out how it should/can work. Roll-out to CI runners is our first target and because we have the flexibility to toggle use of bundle URI on the client side, we can test things safely, without affecting any other user.
Quoted 5 lines
Talking about heuristics, any heuristic will be a _hack_ if you ask me. If you think about it, the bundles them self should be the single source of truth. In fact, a bundle is a packfile with a header prepended that contains refs and their OID in git-show-ref(1) format, and includes which prerequisite OIDs it has.
Yes, that is true indeed.
Quoted 4 lines
So another approach I think of is to have the client partially download bundles (so it pauses the download once they have received the complete header) and make them parse the header to determine if they continue the download. But this approach feels quite complicated to me.
It certainly is more complicated. But it's also the only mechanism that gives you all the information you need, I guess.
Quoted 2 lines
Luckily, with the bundle.heuristic config we can always introduce new values if we discover there is a flaw in the "creationToken" heuristic.
Well, I'd claim we have already shown that it's flawed. I think before we enable incremental bundle fetches by default, even if it's only with the "creationToken" set, we need to show and document a way for server operators to make use of this feature. In other words, we need to document how they have to generate bundles in a way that does not cause clients to re-download everything whenever a new bundle is announced.
Patrick
Re: [PATCH] builtin/clone: teach git-clone(1) the --ref= argument
In reply to Toon ClaesHi,
My apologies, instantly after sending my v1 I noticed some lines in the test weren't indented with tabs. So I'm sending v2 straight away. Sorry for the noise.
--
Toon
PATCH v2 1/1 builtin/clone: teach git-clone(1) the --ref= argument
Add option --ref to git-clone(1). This enables the user to clone and checkout the given named reference. It's pretty similar to --branch and while --branch takes a branch name or tag name, it doesn't take a fully qualified reference. This allows the user to clone a reference that doesn't start with refs/heads or refs/tags. This can be useful when the server uses custom references.
This new argument can be used in conjuction with --single-branch to only clone the given ref.
--- Documentation/git-clone.txt | 9 ++++- builtin/clone.c | 67 ++++++++++++++++++++++++++++++++----- t/t5612-clone-refspec.sh | 35 +++++++++++++++++++ 3 files changed, 102 insertions(+), 9 deletions(-)
Documentation/git-clone.txt +8 -1
@@ -215,6 +215,13 @@ objects from the source repository into a pack in the cloned repository. `--branch` can also take tags and detaches the HEAD at that commit in the resulting repository. +`--ref` _<name>_:: + This detaches HEAD and makes it point to the commit where the _<name>_ + reference is pointing to. In a non-bare repository, this is the ref that + will be checked out. + Can be used in conjunction with `--single-branch` and `--no-tags` to + clone only the given ref. Cannot be combined with `--branch`. + `-u` _<upload-pack>_:: `--upload-pack` _<upload-pack>_:: When given, and the repository to clone from is accessed
@@ -259,7 +266,7 @@ corresponding `--mirror` and `--no-tags` options instead. `--`[`no-`]`single-branch`:: Clone only the history leading to the tip of a single branch, - either specified by the `--branch` option or the primary + either specified by the `--branch` or `--ref` option or the primary branch remote's `HEAD` points at. Further fetches into the resulting repository will only update the remote-tracking branch for the branch this option was used for the
builtin/clone.c +59 -8
131 lines not on this page · raw message
t/t5612-clone-refspec.sh +35 -0
@@ -17,6 +17,7 @@ test_expect_success 'setup' ' git tag two && echo three >file && git commit -a -m three && + git update-ref refs/some/three HEAD && git checkout -b side && echo four >file && git commit -a -m four &&
@@ -236,4 +237,38 @@ test_expect_success '--single-branch with detached' ' test_must_be_empty actual ' +test_expect_success 'with --ref' ' + git clone --ref=refs/some/three . dir_ref && + git -C dir_ref for-each-ref refs > refs && + sed -e "/HEAD$/d" \ + -e "s|/remotes/origin/|/heads/|" refs >actual && + git for-each-ref refs >expect && + test_cmp expect actual +' + +test_expect_success 'with --ref and --no-tags' ' + git clone --ref=refs/some/three --no-tags . dir_ref_notags && + git -C dir_ref_notags for-each-ref refs > refs && + sed -e "/HEAD$/d" \ + -e "s|/remotes/origin/|/heads/|" refs >actual && + git for-each-ref refs/heads >expect && + git for-each-ref refs/some >>expect && + test_cmp expect actual +' + +test_expect_success '--single-branch with --ref' ' + git clone --single-branch --ref=refs/some/three . dir_single_ref && + git -C dir_single_ref for-each-ref refs > actual && + git for-each-ref refs/some >expect && + git for-each-ref refs/tags >>expect && + test_cmp expect actual +' + +test_expect_success '--single-branch with --ref and --no-tags' ' + git clone --single-branch --ref=refs/some/three --no-tags . dir_single_ref_notags && + git -C dir_single_ref_notags for-each-ref refs > actual && + git for-each-ref refs/some >expect && + test_cmp expect actual +' + test_done
Range-diff against v1:
1: 8ee6a42431 ! 1: 07363b1de5 builtin/clone: teach git-clone(1) the --ref= argument
@@ Commit message
checkout the given named reference. It's pretty similar to --branch and
while --branch takes a branch name or tag name, it doesn't take a fully
qualified reference. This allows the user to clone a reference that
- doesn't start with refs/heads/ or refs/tags. This can be useful when the
+ doesn't start with refs/heads or refs/tags. This can be useful when the
server uses custom references.
- Allow the user to clone a certain ref, similar to --branch.
+ This new argument can be used in conjuction with --single-branch to only
+ clone the given ref.
Signed-off-by: Toon Claes <[email protected]>
@@ Documentation/git-clone.txt: objects from the source repository into a pack in t
+ reference is pointing to. In a non-bare repository, this is the ref that
+ will be checked out.
+ Can be used in conjunction with `--single-branch` and `--no-tags` to
-+ clone only given ref. Cannot be combined with `--branch`.
++ clone only the given ref. Cannot be combined with `--branch`.
+
`-u` _<upload-pack>_::
`--upload-pack` _<upload-pack>_::
@@ builtin/clone.c: static void update_head(const struct ref *our, const struct ref
+ else
+ refs_update_ref(get_main_ref_store(the_repository), msg,
+ "HEAD", &our->old_oid, NULL, REF_NO_DEREF,
-+ UPDATE_REFS_MSG_ON_ERR);
++ UPDATE_REFS_DIE_ON_ERR);
} else if (remote) {
/*
* We know remote HEAD points to a non-branch, or
@@ t/t5612-clone-refspec.sh: test_expect_success '--single-branch with detached' '
+test_expect_success 'with --ref' '
+ git clone --ref=refs/some/three . dir_ref &&
-+ git -C dir_ref for-each-ref refs > refs &&
-+ sed -e "/HEAD$/d" \
++ git -C dir_ref for-each-ref refs > refs &&
++ sed -e "/HEAD$/d" \
+ -e "s|/remotes/origin/|/heads/|" refs >actual &&
+ git for-each-ref refs >expect &&
+ test_cmp expect actual
@@ t/t5612-clone-refspec.sh: test_expect_success '--single-branch with detached' '
+
+test_expect_success 'with --ref and --no-tags' '
+ git clone --ref=refs/some/three --no-tags . dir_ref_notags &&
-+ git -C dir_ref_notags for-each-ref refs > refs &&
-+ sed -e "/HEAD$/d" \
++ git -C dir_ref_notags for-each-ref refs > refs &&
++ sed -e "/HEAD$/d" \
+ -e "s|/remotes/origin/|/heads/|" refs >actual &&
+ git for-each-ref refs/heads >expect &&
+ git for-each-ref refs/some >>expect &&
@@ t/t5612-clone-refspec.sh: test_expect_success '--single-branch with detached' '
+
+test_expect_success '--single-branch with --ref' '
+ git clone --single-branch --ref=refs/some/three . dir_single_ref &&
-+ git -C dir_single_ref for-each-ref refs > actual &&
++ git -C dir_single_ref for-each-ref refs > actual &&
+ git for-each-ref refs/some >expect &&
+ git for-each-ref refs/tags >>expect &&
+ test_cmp expect actual
@@ t/t5612-clone-refspec.sh: test_expect_success '--single-branch with detached' '
+
+test_expect_success '--single-branch with --ref and --no-tags' '
+ git clone --single-branch --ref=refs/some/three --no-tags . dir_single_ref_notags &&
-+ git -C dir_single_ref_notags for-each-ref refs > actual &&
++ git -C dir_single_ref_notags for-each-ref refs > actual &&
+ git for-each-ref refs/some >expect &&
+ test_cmp expect actual
+'--
2.46.0