]> dgit.raspbian.org Git - git-annex.git/commitdiff
comment
authorJoey Hess <joeyh@joeyh.name>
Sat, 19 Oct 2024 20:43:35 +0000 (16:43 -0400)
committerJoey Hess <joeyh@joeyh.name>
Sat, 19 Oct 2024 20:43:35 +0000 (16:43 -0400)
doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_1_fe9acc90c92d068109a5f7ec0be1063c._comment [new file with mode: 0644]
doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_2_be800764eb8ce5a6ed181297b3634621._comment [new file with mode: 0644]

diff --git a/doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_1_fe9acc90c92d068109a5f7ec0be1063c._comment b/doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_1_fe9acc90c92d068109a5f7ec0be1063c._comment
new file mode 100644 (file)
index 0000000..5d26313
--- /dev/null
@@ -0,0 +1,9 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 1"""
+ date="2024-10-19T20:23:21Z"
+ content="""
+Here I see the error after `git-annex get`, but not after `git-annex add`.
+
+git version 2.45.2
+"""]]
diff --git a/doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_2_be800764eb8ce5a6ed181297b3634621._comment b/doc/bugs/git-annex_in_a_sha256_repo_produces_error_messages/comment_2_be800764eb8ce5a6ed181297b3634621._comment
new file mode 100644 (file)
index 0000000..5d05c10
--- /dev/null
@@ -0,0 +1,22 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 2"""
+ date="2024-10-19T20:24:21Z"
+ content="""
+The git command that git-annex is running that outputs this message, 
+in the `git-annex get` case, is:
+
+       git diff 4b825dc642cb6eb9a060e54bf8d69288fbee4904 335cd01e749bcfecc96b9d6d1f0bc3edb3ef2aec042da9a40379414a0f5d2243
+
+Here 4b825dc642cb6eb9a060e54bf8d69288fbee4904 is the empty tree sha1, which is
+hardcoded in git-annex as emptyTree. It must need  to instead use the
+corresponding sha256
+(6ef19b41225c5369f1c104d45d8d85efa9b057b53b14b4b9b939dd74decc5321) when diffing
+against a sha256.
+
+A related problem is that deleteSha is the all 0 sha1, and it turns out that
+using that rather than the sha256 version, in a sha256 repository, fails at
+least when `git update-index --index-info` does it. This is a bit surprising
+given the comment in deleteSha that says git will accept the sha1 in that
+situation. deleteSha is used in a couple of other places too, which will need to be checked.
+"""]]