From: Junio C Hamano Date: Tue, 27 Oct 2026 00:34:00 GMT Subject: Re: [PATCH] doc: document two-pass promisor remote fetch order Message-ID: In-Reply-To: <20261026211210.1941-1-edsonwasswa54@gmail.com> Edson writes: > ... > The implementation uses a two-pass approach: first try accepted remotes > (those advertised by the server and accepted by the client), then try > all configured promisor remotes. Document this behavior to make the > technical design notes complete. > ... > Documentation/technical/partial-clone.adoc | 4 ++++ > 1 file changed, 4 insertions(+) The log message may accurately describe the motivation and the behavior being documented, but the patch itself appears to be malformed or based on an unpublished tree. > diff --git a/Documentation/technical/partial-clone.adoc b/Documentation/technical/partial-clone.adoc > index a1104930fb..f80346729b 100644 > --- a/Documentation/technical/partial-clone.adoc > +++ b/Documentation/technical/partial-clone.adoc > @@ -199,6 +199,10 @@ until all the objects have been fetched. > When a server advertises promisor remotes via the protocol v2 > "promisor-remote" capability and the client accepts them (see > linkgit:gitprotocol-v2[5]), those accepted remotes are tried first. > +Note that clients only accept server-advertised promisor remotes when > +explicitly configured to do so via the `promisor.acceptFromServer` or > +`promisor.acceptFromServerUrl` options (see linkgit:git-config[1]), > +which prevents arbitrary injection of untrusted remotes by servers. > After trying the accepted remotes, Git then tries the other configured > promisor remotes as described below. Specifically, the commit message states: "The implementation uses a two-pass approach ... Document this behavior to make the technical design notes complete." But the patch text provides the documentation of the two-pass approach as context lines (without '+'), and only adds the security model paragraph: When sending a new topic that is not anywhere in my tree, please make sure it applies to either (1) the last published feature release (currently, Git 2.56.0), or (2) a recent tip of 'master' branch. The former is preferred for a topic to fix a bug in the released version, while the latter is more appropriate for a new feature and behavior. A patch that does not apply to either should be accompanied by a note that explains on what tree it is designed to be applied (see [*] for an example of how it is communicated). Thanks. [Reference] * In https://lore.kernel.org/git/20261023-b4-pks-odb-extension-v3-0-b1bb30a006db@pks.im/ the author explains the base of the topic as: The series is built on top of c51408889c (The 6th batch, 2026-10-16) with the following dependencies merged into it: - ps/odb-files-alternates at 0ba59c8453 (odb/source: drop `read_alternates` callback, 2026-10-08) merged into it. - ps/setup-enforce-repo-passed-to-create-repository-has-no-state (phew, what a mouthful) at 0ba59c8453 (odb/source: drop `read_alternates` callback, 2026-10-08). * The resulting topic, applying the patches on a synthetic base the above instruction describes, looks like this: $ git log --first-parent --oneline master..901d6767e6 901d6767e6 builtin/submodule: propagate "--object-storage-format=" ed8e104fe6 builtin/clone: introduce "--object-storage-format=" 3545b04fde builtin/init: introduce handling of object storage format 590e753455 builtin: expose the object storage format 074bf8b219 setup: allow overriding object storage format via environment 0538b7b345 setup: introduce object storage extension 27dfe6b692 loose: drop unused `repo_write_loose_object_map()` 1cdc1db14b Merge branch 'ps/setup-enforce-repo-passed-to-create-repository-has-no-state' into ps/odb-object-storage-extension 7464b1a063 Merge branch 'ps/odb-files-alternates' into ps/odb-object-storage-extension