git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] fetch: allow adding a filter after initial clone.
@ 2020-05-13 20:00 Xin Li
  2020-05-13 20:43 ` Junio C Hamano
  2020-05-13 23:44 ` brian m. carlson
  0 siblings, 2 replies; 30+ messages in thread
From: Xin Li @ 2020-05-13 20:00 UTC (permalink / raw)
  To: git; +Cc: Xin Li, jrn

Signed-off-by: Xin Li <delphij@google.com>
---
 builtin/fetch.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index 3ae52c015d..e5faa17ecd 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,8 +1790,16 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
+	if (filter_options.choice && !has_promisor_remote()) {
+		char repo_version_string[10];
+
+		xsnprintf(repo_version_string, sizeof(repo_version_string),
+			  "%d", (int)GIT_REPO_VERSION);
+		git_config_set("core.repositoryformatversion",
+			repo_version_string);
+		git_config_set("extensions.partialclone", "origin");
+		promisor_remote_reinit();
+	}
 
 	if (all) {
 		if (argc == 1)
-- 
2.26.2.645.ge9eca65c58-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH] fetch: allow adding a filter after initial clone.
  2020-05-13 20:00 [PATCH] fetch: allow adding a filter after initial clone Xin Li
@ 2020-05-13 20:43 ` Junio C Hamano
  2020-05-13 21:41   ` Xin Li
  2020-05-13 23:44 ` brian m. carlson
  1 sibling, 1 reply; 30+ messages in thread
From: Junio C Hamano @ 2020-05-13 20:43 UTC (permalink / raw)
  To: Xin Li; +Cc: git, jrn

Xin Li <delphij@google.com> writes:

(nothnig).

Can you help readers by describing what this change is about?

This space is reserved for the patch author to describe why this
change is a good idea (if this patch is adding a new feature), what
is already broken without this patch (if this patch is a bugfix),
and why this change is a safe thing to do (if this patch lifts a
limitation we had before that has been protecting us from getting
into a bad state).


> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  builtin/fetch.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/builtin/fetch.c b/builtin/fetch.c
> index 3ae52c015d..e5faa17ecd 100644
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -1790,8 +1790,16 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  	if (depth || deepen_since || deepen_not.nr)
>  		deepen = 1;
>  
> -	if (filter_options.choice && !has_promisor_remote())
> -		die("--filter can only be used when extensions.partialClone is set");
> +	if (filter_options.choice && !has_promisor_remote()) {
> +		char repo_version_string[10];
> +
> +		xsnprintf(repo_version_string, sizeof(repo_version_string),
> +			  "%d", (int)GIT_REPO_VERSION);
> +		git_config_set("core.repositoryformatversion",
> +			repo_version_string);
> +		git_config_set("extensions.partialclone", "origin");
> +		promisor_remote_reinit();
> +	}
>  
>  	if (all) {
>  		if (argc == 1)

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH] fetch: allow adding a filter after initial clone.
  2020-05-13 20:43 ` Junio C Hamano
@ 2020-05-13 21:41   ` Xin Li
  2020-05-13 22:07     ` Junio C Hamano
  2020-05-13 22:18     ` Junio C Hamano
  0 siblings, 2 replies; 30+ messages in thread
From: Xin Li @ 2020-05-13 21:41 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: git, Jonathan Nieder

The proposed change would allow converting an existing git clone to
use partial clones.  For example:

$ git clone --depth=1 https://android.googlesource.com/platform/bionic .
$ git fetch --unshallow --filter=blob:none origin

Previously, to allow this one would have to do the following manually
first; the existing code would handle the rest gracefully:

$ git config core.repositoryFormatVersion 1
$ git config extensions.partialClone origin

And the proposed change would have git do it for the user
automatically, the existing workflow remains otherwise unchanged.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH] fetch: allow adding a filter after initial clone.
  2020-05-13 21:41   ` Xin Li
@ 2020-05-13 22:07     ` Junio C Hamano
  2020-05-13 22:18     ` Junio C Hamano
  1 sibling, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-05-13 22:07 UTC (permalink / raw)
  To: Xin Li; +Cc: git, Jonathan Nieder

Xin Li <delphij@google.com> writes:

> The proposed change would allow converting an existing git clone to
> use partial clones.  For example:
>
> $ git clone --depth=1 https://android.googlesource.com/platform/bionic .
> $ git fetch --unshallow --filter=blob:none origin
>
> Previously, to allow this one would have to do the following manually
> first; the existing code would handle the rest gracefully:
>
> $ git config core.repositoryFormatVersion 1
> $ git config extensions.partialClone origin
>
> And the proposed change would have git do it for the user
> automatically, the existing workflow remains otherwise unchanged.

Please do not explain these things to the list readers.  Instead,
remember to write them in the proposed log message when you are
ready to send an updated patch, which you would probably want to
wait before giving other people to comment on the patch.

Thanks.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH] fetch: allow adding a filter after initial clone.
  2020-05-13 21:41   ` Xin Li
  2020-05-13 22:07     ` Junio C Hamano
@ 2020-05-13 22:18     ` Junio C Hamano
  1 sibling, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-05-13 22:18 UTC (permalink / raw)
  To: Xin Li
  Cc: git, Jonathan Nieder, Derrick Stolee, Christian Couder, Jonathan Tan

Xin Li <delphij@google.com> writes:

> The proposed change would allow converting an existing git clone to
> use partial clones.  For example:
>
> $ git clone --depth=1 https://android.googlesource.com/platform/bionic .
> $ git fetch --unshallow --filter=blob:none origin
>
> Previously, to allow this one would have to do the following manually
> first; the existing code would handle the rest gracefully:
>
> $ git config core.repositoryFormatVersion 1
> $ git config extensions.partialClone origin
>
> And the proposed change would have git do it for the user
> automatically, the existing workflow remains otherwise unchanged.

Actually, do we even know, when we "upgrade" the repository with the
patch, that the 'origin' supports the required protocol extensions
like on-demand fetching of objects?

At "git clone" time, "git clone --filter=..."  against a server that
is not prepared to serve as a lazy clone's backup remote would fail,
so we will never write extensions.partialClone=origin when origin
cannot serve as such. 

But writing these config you suggest to "do manually" is already not
a safe thing to do, without making sure it is supported, no?  It
does not sound like a good idea to do an unsafe thing for the user
automatically.

Added a few folks on Cc: list, who know and care more about partial
clones than I do for input.

Thanks.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH] fetch: allow adding a filter after initial clone.
  2020-05-13 20:00 [PATCH] fetch: allow adding a filter after initial clone Xin Li
  2020-05-13 20:43 ` Junio C Hamano
@ 2020-05-13 23:44 ` brian m. carlson
  2020-05-28  2:53   ` [PATCH v2 0/1] " Xin Li
  2020-05-28  2:54   ` [PATCH v2 1/1] " Xin Li
  1 sibling, 2 replies; 30+ messages in thread
From: brian m. carlson @ 2020-05-13 23:44 UTC (permalink / raw)
  To: Xin Li; +Cc: git, jrn

[-- Attachment #1: Type: text/plain, Size: 3457 bytes --]

On 2020-05-13 at 20:00:40, Xin Li wrote:
> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  builtin/fetch.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/builtin/fetch.c b/builtin/fetch.c
> index 3ae52c015d..e5faa17ecd 100644
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -1790,8 +1790,16 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  	if (depth || deepen_since || deepen_not.nr)
>  		deepen = 1;
>  
> -	if (filter_options.choice && !has_promisor_remote())
> -		die("--filter can only be used when extensions.partialClone is set");
> +	if (filter_options.choice && !has_promisor_remote()) {
> +		char repo_version_string[10];
> +
> +		xsnprintf(repo_version_string, sizeof(repo_version_string),
> +			  "%d", (int)GIT_REPO_VERSION);
> +		git_config_set("core.repositoryformatversion",
> +			repo_version_string);
> +		git_config_set("extensions.partialclone", "origin");

Some things stood out to me here.  One, is this setting up the
repository if it's not already created?  If so, we'd probably want to
use one of the appropriate functions in setup.c.  Even if we're just
changing it, we should probably use a helper function.

Two, it isn't necessarily safe to automatically change the repository
version.  Keys that start with "extensions." are not special in format
version 0, but they are in format version 1.  I can technically have an
"extensions.tomatoSalad" in version 0 without any special meaning or
negative consequences, but as soon as we change to version 1, Git will
refuse to read it, since it doesn't know about the tomatoSalad extension
and in v1 unknown extensions are fatal.

My example may sound silly, but since extensions can be set in the
global config, users could well rely on v0 repositories ignoring them
and having them automatically turned on for their v1 repositories.  (I'm
thinking of the future extensions.objectFormat as something somebody
might try to do here, as dangerous as it may be.)

These aren't insurmountable problems, but they are things we'd need to
check for before just changing the repository version, so we'd want to
stuff some code in setup.c to handle this case.

Third, I'm not sure that "origin" is always the value we want to use
here.  At a previous employer, the upstream remote was called
"upstream", and your personal fork was called "origin", so I'd have
wanted upstream here..  We'd probably want to use whatever remote the
user is already using in this case, and gracefully handle the URL case
if that isn't allowed here (which it may be; I'm not that familiar with
partial clone).

I also agree with Junio's assessment that you'd probably want to explain
more about this feature in the commit message.  For example, I'd want to
know what this patch does and have some idea of how I might invoke this
feature, why it's safe to change the repository version, how one sets
the remote for fetch, and pretty much answers to all the other things I
asked here.  Even if I understand these things now, that doesn't mean a
future developer will in six months' time, and mentioning these things
in the commit message helps leave a note to them that you considered (or
did not consider, as the case may be) certain issues and helps them
understand the code as you saw and wrote it.
-- 
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 263 bytes --]

^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v2 0/1] fetch: allow adding a filter after initial clone
  2020-05-13 23:44 ` brian m. carlson
@ 2020-05-28  2:53   ` Xin Li
  2020-05-28  2:54   ` [PATCH v2 1/1] " Xin Li
  1 sibling, 0 replies; 30+ messages in thread
From: Xin Li @ 2020-05-28  2:53 UTC (permalink / raw)
  To: git; +Cc: Xin Li, jrn, iankaz

Previously, to retroactively add filter to an existing (shallow) clone
one would have to manually change the repository configuration to make
git to believe that there was an existing promisor, like:

  git config core.repositoryFormatVersion 1
  git config extensions.partialClone origin
  git fetch --unshallow --filter=blob:none origin

Because the code can already set up promisor, it would be safer and more
convenient to just do that in git itself.

This version of change will also prevent the code from making damaging
repository upgrades (when non-standard extensions exists) as pointed out
by earlier reviewers.

Xin Li (1):
  fetch: allow adding a filter after initial clone.

 builtin/fetch.c               |  3 ---
 list-objects-filter-options.c |  3 ++-
 repository.h                  |  6 ++++++
 setup.c                       | 30 ++++++++++++++++++++++++++++++
 t/t0410-partial-clone.sh      | 21 +++++++++++++++++++++
 5 files changed, 59 insertions(+), 4 deletions(-)

-- 
2.27.0.rc0.183.gde8f92d652-goog


^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-13 23:44 ` brian m. carlson
  2020-05-28  2:53   ` [PATCH v2 0/1] " Xin Li
@ 2020-05-28  2:54   ` Xin Li
  2020-05-28  3:28     ` Jonathan Nieder
  2020-05-28 15:04     ` [PATCH v2 1/1] " Junio C Hamano
  1 sibling, 2 replies; 30+ messages in thread
From: Xin Li @ 2020-05-28  2:54 UTC (permalink / raw)
  To: git; +Cc: Xin Li

Retroactively adding filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.

Previously this is possible by manually amending the repository
configuration to make git think there is an existing promisor. Because
the code already does most of the hard work, it's safer for git to
just perform the configuration change automatically instead.

Instead of bailing out immediately when no promisor is available, make
the code check more specific issue (extension became special in
repository version 1, while it can have any value in version 0, so
upgrade should not happen if the repository have an unsupported
configuration that would render it invalid if we upgraded).

Signed-off-by: Xin Li <delphij@google.com>
---
 builtin/fetch.c               |  3 ---
 list-objects-filter-options.c |  3 ++-
 repository.h                  |  6 ++++++
 setup.c                       | 30 ++++++++++++++++++++++++++++++
 t/t0410-partial-clone.sh      | 21 +++++++++++++++++++++
 5 files changed, 59 insertions(+), 4 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b5788c16bf..3347d578ea 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
-
 	if (all) {
 		if (argc == 1)
 			die(_("fetch --all does not take a repository argument"));
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 256bcfbdfe..6d62b60eac 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -326,7 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		git_config_set("core.repositoryformatversion", "1");
+		if (upgrade_repository_format(the_repository, 1) < 0)
+			die(_("Unable to upgrade repository format to support partial clone"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git a/repository.h b/repository.h
index 6534fbb7b3..f301f6f456 100644
--- a/repository.h
+++ b/repository.h
@@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 
 void prepare_repo_settings(struct repository *r);
 
+/*
+ * Return 1 if upgrade repository format to target_version succeeded,
+ * 0 if no upgrade is necessary.
+ */
+int upgrade_repository_format(struct repository *r, int target_version);
+
 #endif /* REPOSITORY_H */
diff --git a/setup.c b/setup.c
index 65fe5ecefb..84da976e07 100644
--- a/setup.c
+++ b/setup.c
@@ -538,6 +538,36 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
+int upgrade_repository_format(struct repository *r, int target_version)
+{
+	const char *gitdir = get_git_dir();
+	struct strbuf sb = STRBUF_INIT;
+	struct strbuf err = STRBUF_INIT;
+	struct strbuf repo_version = STRBUF_INIT;
+	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
+
+	strbuf_git_common_path(&sb, r, "/config");
+	read_repository_format(&repo_fmt, sb.buf);
+	strbuf_release(&sb);
+
+	if (repo_fmt.version >= target_version)
+		return 0;
+
+	repo_fmt.version = target_version;
+
+	if (verify_repository_format(&repo_fmt, &err) < 0) {
+		warning("Unable to upgrade repository format to %d: %s",
+		    target_version, err.buf);
+		strbuf_release(&err);
+		return -1;
+	}
+
+	strbuf_addf(&repo_version, "%d", target_version);
+	git_config_set("core.repositoryformatversion", repo_version.buf);
+	strbuf_release(&repo_version);
+	return 1;
+}
+
 static void init_repository_format(struct repository_format *format)
 {
 	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index a3988bd4b8..71270d3a53 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -30,6 +30,27 @@ test_expect_success 'extensions.partialclone without filter' '
 	git -C client fetch origin
 '
 
+test_expect_success 'convert shallow clone to partial clone' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client fetch --unshallow --filter="blob:none" &&
+	test_cmp_config -C client true remote.origin.promisor &&
+	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
+	test_cmp_config -C client 1 core.repositoryformatversion
+'
+test_expect_success 'convert shallow clone to partial clone must fail with invalid extension' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client config extensions.sandwidth true &&
+	test_must_fail git -C client fetch --unshallow --filter="blob:none"
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
-- 
2.27.0.rc0.183.gde8f92d652-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-28  2:54   ` [PATCH v2 1/1] " Xin Li
@ 2020-05-28  3:28     ` Jonathan Nieder
  2020-05-28  4:08       ` [PATCH v3] " Xin Li
  2020-05-28 15:04     ` [PATCH v2 1/1] " Junio C Hamano
  1 sibling, 1 reply; 30+ messages in thread
From: Jonathan Nieder @ 2020-05-28  3:28 UTC (permalink / raw)
  To: Xin Li; +Cc: git, brian m. carlson

Hi,

Xin Li wrote:

> Retroactively adding filter can be useful for existing shallow clones as
> they allow users to see earlier change histories without downloading all
> git objects in a regular --unshallow fetch.
>
> Previously this is possible by manually amending the repository
> configuration to make git think there is an existing promisor. Because
> the code already does most of the hard work, it's safer for git to
> just perform the configuration change automatically instead.
>
> Instead of bailing out immediately when no promisor is available, make
> the code check more specific issue (extension became special in
> repository version 1, while it can have any value in version 0, so
> upgrade should not happen if the repository have an unsupported
> configuration that would render it invalid if we upgraded).
>
> Signed-off-by: Xin Li <delphij@google.com>
> ---

nit: the cover letter contains

> Previously, to retroactively add filter to an existing (shallow) clone
> one would have to manually change the repository configuration to make
> git to believe that there was an existing promisor, like:
>
>   git config core.repositoryFormatVersion 1
>   git config extensions.partialClone origin
>   git fetch --unshallow --filter=blob:none origin
>
> Because the code can already set up promisor, it would be safer and more
> convenient to just do that in git itself.
>
> This version of change will also prevent the code from making damaging
> repository upgrades (when non-standard extensions exists) as pointed out
> by earlier reviewers.

I think that would make a good commit message itself.

[...]
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  	if (depth || deepen_since || deepen_not.nr)
>  		deepen = 1;
>  
> -	if (filter_options.choice && !has_promisor_remote())
> -		die("--filter can only be used when extensions.partialClone is set");
> -

Makes sense.

[...]
> --- a/list-objects-filter-options.c
> +++ b/list-objects-filter-options.c
> @@ -326,7 +326,8 @@ void partial_clone_register(
>  
>  	/* Check if it is already registered */
>  	if (!promisor_remote_find(remote)) {
> -		git_config_set("core.repositoryformatversion", "1");
> +		if (upgrade_repository_format(the_repository, 1) < 0)
> +			die(_("Unable to upgrade repository format to support partial clone"));

nit: Git's error messages tend to use lowercase (e.g., "fatal: cannot [etc]"
instead of "fatal: Unable [etc]").

>  		/* Add promisor config for the remote */
>  		cfg_name = xstrfmt("remote.%s.promisor", remote);

not about this patch: By the way, the repository format version bump
is not sufficient to achieve its intended aim: we also need to set an
extensions.* setting to ensure Git is new enough to know about partial
clone.  More discussion about this is in [1] (apologies for not having
finished solving that).  This isn't a regression introduced in this
patch, and this patch does the right thing in the context of the
current code.

[...]
> --- a/repository.h
> +++ b/repository.h
> @@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
>  
>  void prepare_repo_settings(struct repository *r);
>  
> +/*
> + * Return 1 if upgrade repository format to target_version succeeded,
> + * 0 if no upgrade is necessary.
> + */

Probably also worth mentioning that this returns -1 on error.

> +int upgrade_repository_format(struct repository *r, int target_version);
> +
>  #endif /* REPOSITORY_H */
[...]
> +++ b/setup.c
> @@ -538,6 +538,36 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
>  	return 0;
>  }
>  
> +int upgrade_repository_format(struct repository *r, int target_version)
> +{
> +	const char *gitdir = get_git_dir();

Unused variable.

> +	struct strbuf sb = STRBUF_INIT;
> +	struct strbuf err = STRBUF_INIT;
> +	struct strbuf repo_version = STRBUF_INIT;
> +	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
> +
> +	strbuf_git_common_path(&sb, r, "/config");

nit: can leave out the '/' to avoid a double-/.

> +	read_repository_format(&repo_fmt, sb.buf);
> +	strbuf_release(&sb);
> +
> +	if (repo_fmt.version >= target_version)
> +		return 0;
> +
> +	repo_fmt.version = target_version;
> +
> +	if (verify_repository_format(&repo_fmt, &err) < 0) {
> +		warning("Unable to upgrade repository format to %d: %s",

Same nit about capitalization.

> +		    target_version, err.buf);

whitespace nit: this would typically use a tab, to line up with the
paren on the previous line.

> +		strbuf_release(&err);
> +		return -1;
> +	}
> +
> +	strbuf_addf(&repo_version, "%d", target_version);
> +	git_config_set("core.repositoryformatversion", repo_version.buf);

Ah, I think I misled you: the config_set API hasn't learned to take a
struct repository yet, so we should hardcode the_repository in this
function instead of taking a "struct repository" parameter.

> +	strbuf_release(&repo_version);
> +	return 1;
> +}
> +
>  static void init_repository_format(struct repository_format *format)
>  {
>  	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
> diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
> index a3988bd4b8..71270d3a53 100755
> --- a/t/t0410-partial-clone.sh
> +++ b/t/t0410-partial-clone.sh
> @@ -30,6 +30,27 @@ test_expect_success 'extensions.partialclone without filter' '
>  	git -C client fetch origin
>  '
>  
> +test_expect_success 'convert shallow clone to partial clone' '
> +	rm -fr server client &&
> +	test_create_repo server &&
> +	test_commit -C server my_commit 1 &&
> +	test_commit -C server my_commit2 1 &&
> +	git clone --depth=1 "file://$(pwd)/server" client &&
> +	git -C client fetch --unshallow --filter="blob:none" &&
> +	test_cmp_config -C client true remote.origin.promisor &&
> +	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
> +	test_cmp_config -C client 1 core.repositoryformatversion
> +'

nit: Missing blank line.

Is there a different check this test could perform to check the
user-facing behavior instead of how the configuration is encoded?

> +test_expect_success 'convert shallow clone to partial clone must fail with invalid extension' '
> +	rm -fr server client &&
> +	test_create_repo server &&
> +	test_commit -C server my_commit 1 &&
> +	test_commit -C server my_commit2 1 &&
> +	git clone --depth=1 "file://$(pwd)/server" client &&
> +	git -C client config extensions.sandwidth true &&
> +	test_must_fail git -C client fetch --unshallow --filter="blob:none"
> +'
> +
>  test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
>  	rm -rf repo &&
>  	test_create_repo repo &&

With whatever subset of the mentioned changes makes sense,
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>

Thanks.

[1] https://lore.kernel.org/git/20200312230931.GF120942@google.com/.

diff --git i/list-objects-filter-options.c w/list-objects-filter-options.c
index 6d62b60eaca..ce9193c3885 100644
--- i/list-objects-filter-options.c
+++ w/list-objects-filter-options.c
@@ -326,8 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		if (upgrade_repository_format(the_repository, 1) < 0)
-			die(_("Unable to upgrade repository format to support partial clone"));
+		if (upgrade_repository_format(1) < 0)
+			die(_("cannot enable partial clone support"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git i/repository.h w/repository.h
index f301f6f4562..14574c6e627 100644
--- i/repository.h
+++ w/repository.h
@@ -197,9 +197,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 void prepare_repo_settings(struct repository *r);
 
 /*
- * Return 1 if upgrade repository format to target_version succeeded,
- * 0 if no upgrade is necessary.
+ * Upgrade the repository format to target_version.
+ * Returns 1 on success, 0 if no upgrade was necessary, and -1 after
+ * printing a diagnostic on error.
  */
-int upgrade_repository_format(struct repository *r, int target_version);
+int upgrade_repository_format(int target_version);
 
 #endif /* REPOSITORY_H */
diff --git i/setup.c w/setup.c
index 84da976e077..d1f0aff7d30 100644
--- i/setup.c
+++ w/setup.c
@@ -538,15 +538,14 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
-int upgrade_repository_format(struct repository *r, int target_version)
+int upgrade_repository_format(int target_version)
 {
-	const char *gitdir = get_git_dir();
 	struct strbuf sb = STRBUF_INIT;
 	struct strbuf err = STRBUF_INIT;
 	struct strbuf repo_version = STRBUF_INIT;
 	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
 
-	strbuf_git_common_path(&sb, r, "/config");
+	strbuf_git_common_path(&sb, the_repository, "config");
 	read_repository_format(&repo_fmt, sb.buf);
 	strbuf_release(&sb);
 
@@ -556,8 +555,8 @@ int upgrade_repository_format(struct repository *r, int target_version)
 	repo_fmt.version = target_version;
 
 	if (verify_repository_format(&repo_fmt, &err) < 0) {
-		warning("Unable to upgrade repository format to %d: %s",
-		    target_version, err.buf);
+		warning("cannot upgrade repository format to %d: %s",
+			target_version, err.buf);
 		strbuf_release(&err);
 		return -1;
 	}
diff --git i/t/t0410-partial-clone.sh w/t/t0410-partial-clone.sh
index 71270d3a539..d580488330f 100755
--- i/t/t0410-partial-clone.sh
+++ w/t/t0410-partial-clone.sh
@@ -41,6 +41,7 @@ test_expect_success 'convert shallow clone to partial clone' '
 	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
 	test_cmp_config -C client 1 core.repositoryformatversion
 '
+
 test_expect_success 'convert shallow clone to partial clone must fail with invalid extension' '
 	rm -fr server client &&
 	test_create_repo server &&

^ permalink raw reply related	[flat|nested] 30+ messages in thread

* [PATCH v3] fetch: allow adding a filter after initial clone.
  2020-05-28  3:28     ` Jonathan Nieder
@ 2020-05-28  4:08       ` Xin Li
  0 siblings, 0 replies; 30+ messages in thread
From: Xin Li @ 2020-05-28  4:08 UTC (permalink / raw)
  To: git; +Cc: Xin Li, jrn, iankaz, Jonathan Nieder

Retroactively adding filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.

Without this patch, users can make a clone partial by editing the
repository configuration to convert the remote into a promisor, like:

  git config core.repositoryFormatVersion 1
  git config extensions.partialClone origin
  git fetch --unshallow --filter=blob:none origin

Since the hard part of making this work is already in place and such
edits can be error-prone, teach Git to perform the required configuration
change automatically instead.

Instead of bailing out immediately when no promisor is available, make
the code perform a more precise check for any potential problems
(extensions became special in repository version 1, while it can have
any value in version 0, so upgrade should not happen if the repository
have an unsupported configuration that would render it invalid if we
upgraded).

Signed-off-by: Xin Li <delphij@google.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
---
 builtin/fetch.c               |  3 ---
 list-objects-filter-options.c |  3 ++-
 repository.h                  |  6 ++++++
 setup.c                       | 29 +++++++++++++++++++++++++++++
 t/t0410-partial-clone.sh      | 22 ++++++++++++++++++++++
 5 files changed, 59 insertions(+), 4 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b5788c16bf..3347d578ea 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
-
 	if (all) {
 		if (argc == 1)
 			die(_("fetch --all does not take a repository argument"));
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 256bcfbdfe..3553ad7b0a 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -326,7 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		git_config_set("core.repositoryformatversion", "1");
+		if (upgrade_repository_format(1) < 0)
+			die(_("unable to upgrade repository format to support partial clone"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git a/repository.h b/repository.h
index 6534fbb7b3..40cc12c7cf 100644
--- a/repository.h
+++ b/repository.h
@@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 
 void prepare_repo_settings(struct repository *r);
 
+/*
+ * Return 1 if upgrade repository format to target_version succeeded,
+ * 0 if no upgrade is necessary; returns -1 when upgrade is not possible.
+ */
+int upgrade_repository_format(int target_version);
+
 #endif /* REPOSITORY_H */
diff --git a/setup.c b/setup.c
index 65fe5ecefb..123d3d3433 100644
--- a/setup.c
+++ b/setup.c
@@ -538,6 +538,35 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
+int upgrade_repository_format(int target_version)
+{
+	struct strbuf sb = STRBUF_INIT;
+	struct strbuf err = STRBUF_INIT;
+	struct strbuf repo_version = STRBUF_INIT;
+	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
+
+	strbuf_git_common_path(&sb, the_repository, "config");
+	read_repository_format(&repo_fmt, sb.buf);
+	strbuf_release(&sb);
+
+	if (repo_fmt.version >= target_version)
+		return 0;
+
+	repo_fmt.version = target_version;
+
+	if (verify_repository_format(&repo_fmt, &err) < 0) {
+		warning("unable to upgrade repository format to %d: %s",
+		    target_version, err.buf);
+		strbuf_release(&err);
+		return -1;
+	}
+
+	strbuf_addf(&repo_version, "%d", target_version);
+	git_config_set("core.repositoryformatversion", repo_version.buf);
+	strbuf_release(&repo_version);
+	return 1;
+}
+
 static void init_repository_format(struct repository_format *format)
 {
 	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index a3988bd4b8..d580488330 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -30,6 +30,28 @@ test_expect_success 'extensions.partialclone without filter' '
 	git -C client fetch origin
 '
 
+test_expect_success 'convert shallow clone to partial clone' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client fetch --unshallow --filter="blob:none" &&
+	test_cmp_config -C client true remote.origin.promisor &&
+	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
+	test_cmp_config -C client 1 core.repositoryformatversion
+'
+
+test_expect_success 'convert shallow clone to partial clone must fail with invalid extension' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client config extensions.sandwidth true &&
+	test_must_fail git -C client fetch --unshallow --filter="blob:none"
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
-- 
2.27.0.rc0.183.gde8f92d652-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-28  2:54   ` [PATCH v2 1/1] " Xin Li
  2020-05-28  3:28     ` Jonathan Nieder
@ 2020-05-28 15:04     ` Junio C Hamano
  2020-05-28 17:19       ` Jonathan Nieder
  1 sibling, 1 reply; 30+ messages in thread
From: Junio C Hamano @ 2020-05-28 15:04 UTC (permalink / raw)
  To: Xin Li; +Cc: git

Xin Li <delphij@google.com> writes:

> Instead of bailing out immediately when no promisor is available, make
> the code check more specific issue (extension became special in
> repository version 1, while it can have any value in version 0, so
> upgrade should not happen if the repository have an unsupported
> configuration that would render it invalid if we upgraded).

It probably has to be a lot stronger than that.  In version 0,
extension did not have any meaning, so an existing repository can
safely have "[extension] worktreeConfig=~/hello" as long as it is
version 0 and nobody would bitch about extension.worktreeConfig not
being a boolean; worse yet, "[extension] worktreeConfig=true" in
version 0 repository did not make its secondary worktrees to have
separate configuration, but if we upgrade to version 1 merely
because the version of Git knows what extension.worktreeConfig means,
we broke the repository and its worktrees.

It would be safe to upgrade version 0 repository when there is *no*
existing configuration variable in the "extension" section, but "the
repository have an unsupported configuration" is not a useful or
safe criteria to decide if we should refrain from upgrading, I would
think.


^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-28 15:04     ` [PATCH v2 1/1] " Junio C Hamano
@ 2020-05-28 17:19       ` Jonathan Nieder
  2020-05-28 19:12         ` Xin Li
  0 siblings, 1 reply; 30+ messages in thread
From: Jonathan Nieder @ 2020-05-28 17:19 UTC (permalink / raw)
  To: Junio C Hamano; +Cc: Xin Li, git, brian m. carlson

Junio C Hamano wrote:

> It would be safe to upgrade version 0 repository when there is *no*
> existing configuration variable in the "extension" section,

Yes, that makes sense (also for the case Brian mentioned where someone
may try setting extension.objectFormat, see no effect, and then forget
about it).

Thanks,
Jonathan

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-28 17:19       ` Jonathan Nieder
@ 2020-05-28 19:12         ` Xin Li
  2020-05-28 19:17           ` Jonathan Nieder
  0 siblings, 1 reply; 30+ messages in thread
From: Xin Li @ 2020-05-28 19:12 UTC (permalink / raw)
  To: Jonathan Nieder; +Cc: Junio C Hamano, git, brian m. carlson

On Thu, May 28, 2020 at 10:19 AM Jonathan Nieder <jrnieder@gmail.com> wrote:
>
> Junio C Hamano wrote:
>
> > It would be safe to upgrade version 0 repository when there is *no*
> > existing configuration variable in the "extension" section,
>
> Yes, that makes sense (also for the case Brian mentioned where someone
> may try setting extension.objectFormat, see no effect, and then forget
> about it).

I see that's a good point

I found that the current code doesn't really check
repositoryformatversion for e.g. partialclone, so without setting
repositoryformatversion to 1, just setting extensions.partialclone =
origin will still make fetch --filter work and some test cases are
depending on this, do we want to keep this feature/bug?  I have added
some additional checks to my local git version and found that we would
also need to fix a lot of test cases for worktree, etc. if we would
enforce the check.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v2 1/1] fetch: allow adding a filter after initial clone.
  2020-05-28 19:12         ` Xin Li
@ 2020-05-28 19:17           ` Jonathan Nieder
  2020-05-29  0:04             ` [PATCH v4] " Xin Li
  0 siblings, 1 reply; 30+ messages in thread
From: Jonathan Nieder @ 2020-05-28 19:17 UTC (permalink / raw)
  To: Xin Li; +Cc: Junio C Hamano, git, brian m. carlson

Xin Li wrote:

> I found that the current code doesn't really check
> repositoryformatversion for e.g. partialclone, so without setting
> repositoryformatversion to 1, just setting extensions.partialclone =
> origin will still make fetch --filter work and some test cases are
> depending on this, do we want to keep this feature/bug?

It's orthogonal to this patch, but no, we don't.

Thanks,
Jonathan

^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-28 19:17           ` Jonathan Nieder
@ 2020-05-29  0:04             ` Xin Li
  2020-05-29  0:41               ` Junio C Hamano
  2020-05-29  1:01               ` Jonathan Nieder
  0 siblings, 2 replies; 30+ messages in thread
From: Xin Li @ 2020-05-29  0:04 UTC (permalink / raw)
  To: git; +Cc: Xin Li, jrnieder, gitster, sandals, iankaz


Retroactively adding filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.

Without this patch, users can make a clone partial by editing the
repository configuration to convert the remote into a promisor, like:

  git config core.repositoryFormatVersion 1
  git config extensions.partialClone origin
  git fetch --unshallow --filter=blob:none origin

Since the hard part of making this work is already in place and such
edits can be error-prone, teach Git to perform the required configuration
change automatically instead.

Instead of bailing out immediately when no promisor is available, make
the code perform a more precise check for any potential problems
(extensions became special in repository version 1, while it can have
any value in version 0, so upgrade should not happen if the repository
have an unsupported configuration that would render it invalid if we
upgraded).

Signed-off-by: Xin Li <delphij@google.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
---
 builtin/fetch.c                  |  3 --
 builtin/sparse-checkout.c        |  2 +
 cache.h                          |  2 +
 list-objects-filter-options.c    |  3 +-
 repository.h                     |  6 +++
 setup.c                          | 77 ++++++++++++++++++++++++++++++--
 t/t0410-partial-clone.sh         | 23 ++++++++++
 t/t1090-sparse-checkout-scope.sh |  1 -
 t/t2404-worktree-config.sh       |  4 +-
 t/t5500-fetch-pack.sh            |  1 -
 t/t5702-protocol-v2.sh           |  1 -
 11 files changed, 111 insertions(+), 12 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b5788c16bf..3347d578ea 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
-
 	if (all) {
 		if (argc == 1)
 			die(_("fetch --all does not take a repository argument"));
diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index 95d0882417..95669815d4 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c
@@ -249,6 +249,8 @@ static int set_config(enum sparse_checkout_mode mode)
 {
 	const char *config_path;
 
+	if (upgrade_repository_format(1) < 0)
+		die(_("unable to upgrade repository format to enable worktreeConfig"));
 	if (git_config_set_gently("extensions.worktreeConfig", "true")) {
 		error(_("failed to set extensions.worktreeConfig setting"));
 		return 1;
diff --git a/cache.h b/cache.h
index 0f0485ecfe..66dcd2f219 100644
--- a/cache.h
+++ b/cache.h
@@ -1042,6 +1042,7 @@ struct repository_format {
 	int worktree_config;
 	int is_bare;
 	int hash_algo;
+	int has_extensions;
 	char *work_tree;
 	struct string_list unknown_extensions;
 };
@@ -1056,6 +1057,7 @@ struct repository_format {
 	.version = -1, \
 	.is_bare = -1, \
 	.hash_algo = GIT_HASH_SHA1, \
+	.has_extensions = 0, \
 	.unknown_extensions = STRING_LIST_INIT_DUP, \
 }
 
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 256bcfbdfe..3553ad7b0a 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -326,7 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		git_config_set("core.repositoryformatversion", "1");
+		if (upgrade_repository_format(1) < 0)
+			die(_("unable to upgrade repository format to support partial clone"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git a/repository.h b/repository.h
index 6534fbb7b3..40cc12c7cf 100644
--- a/repository.h
+++ b/repository.h
@@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 
 void prepare_repo_settings(struct repository *r);
 
+/*
+ * Return 1 if upgrade repository format to target_version succeeded,
+ * 0 if no upgrade is necessary; returns -1 when upgrade is not possible.
+ */
+int upgrade_repository_format(int target_version);
+
 #endif /* REPOSITORY_H */
diff --git a/setup.c b/setup.c
index 65fe5ecefb..0759e9f8f9 100644
--- a/setup.c
+++ b/setup.c
@@ -13,6 +13,9 @@ static int work_tree_config_is_bogus;
 static struct startup_info the_startup_info;
 struct startup_info *startup_info = &the_startup_info;
 
+static int verify_repository_format_eligibility(const struct repository_format *,
+    struct strbuf *, int);
+
 /*
  * The input parameter must contain an absolute path, and it must already be
  * normalized.
@@ -455,6 +458,7 @@ static int check_repo_format(const char *var, const char *value, void *vdata)
 	if (strcmp(var, "core.repositoryformatversion") == 0)
 		data->version = git_config_int(var, value);
 	else if (skip_prefix(var, "extensions.", &ext)) {
+		data->has_extensions = 1;
 		/*
 		 * record any known extensions here; otherwise,
 		 * we fall through to recording it as unknown, and
@@ -506,9 +510,15 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 		die("%s", err.buf);
 	}
 
-	repository_format_precious_objects = candidate->precious_objects;
-	set_repository_format_partial_clone(candidate->partial_clone);
-	repository_format_worktree_config = candidate->worktree_config;
+	if (candidate->version >= 1) {
+		repository_format_precious_objects = candidate->precious_objects;
+		set_repository_format_partial_clone(candidate->partial_clone);
+		repository_format_worktree_config = candidate->worktree_config;
+	} else {
+		repository_format_precious_objects = 0;
+		set_repository_format_partial_clone(NULL);
+		repository_format_worktree_config = 0;
+	}
 	string_list_clear(&candidate->unknown_extensions, 0);
 
 	if (repository_format_worktree_config) {
@@ -538,6 +548,34 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
+int upgrade_repository_format(int target_version)
+{
+	struct strbuf sb = STRBUF_INIT;
+	struct strbuf err = STRBUF_INIT;
+	struct strbuf repo_version = STRBUF_INIT;
+	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
+
+	strbuf_git_common_path(&sb, the_repository, "config");
+	read_repository_format(&repo_fmt, sb.buf);
+	strbuf_release(&sb);
+
+	if (repo_fmt.version >= target_version)
+		return 0;
+
+	if (verify_repository_format_eligibility(&repo_fmt, &err,
+	    target_version) < 0) {
+		warning("unable to upgrade repository format from %d to %d: %s",
+		    repo_fmt.version, target_version, err.buf);
+		strbuf_release(&err);
+		return -1;
+	}
+
+	strbuf_addf(&repo_version, "%d", target_version);
+	git_config_set("core.repositoryformatversion", repo_version.buf);
+	strbuf_release(&repo_version);
+	return 1;
+}
+
 static void init_repository_format(struct repository_format *format)
 {
 	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
@@ -562,7 +600,7 @@ void clear_repository_format(struct repository_format *format)
 	init_repository_format(format);
 }
 
-int verify_repository_format(const struct repository_format *format,
+static int verify_repository_format_version(const struct repository_format *format,
 			     struct strbuf *err)
 {
 	if (GIT_REPO_VERSION_READ < format->version) {
@@ -571,6 +609,18 @@ int verify_repository_format(const struct repository_format *format,
 		return -1;
 	}
 
+	return 0;
+}
+
+int verify_repository_format(const struct repository_format *format,
+			     struct strbuf *err)
+{
+	int result;
+
+	result = verify_repository_format_version(format, err);
+	if (result != 0)
+		return (result);
+
 	if (format->version >= 1 && format->unknown_extensions.nr) {
 		int i;
 
@@ -585,6 +635,25 @@ int verify_repository_format(const struct repository_format *format,
 	return 0;
 }
 
+static int verify_repository_format_eligibility(const struct repository_format
+    *format, struct strbuf *err, int target_version)
+{
+	int result;
+
+	result = verify_repository_format_version(format, err);
+	if (result != 0)
+		return (result);
+
+	if (format->version <= 0 && format->has_extensions &&
+	    target_version >= 1) {
+		strbuf_addf(err, _("extensions found for repository version %d"),
+			      format->version);
+		return -1;
+	}
+
+	return 0;
+}
+
 void read_gitfile_error_die(int error_code, const char *path, const char *dir)
 {
 	switch (error_code) {
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index a3988bd4b8..463dc3a8be 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -30,6 +30,29 @@ test_expect_success 'extensions.partialclone without filter' '
 	git -C client fetch origin
 '
 
+test_expect_success 'convert shallow clone to partial clone' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client fetch --unshallow --filter="blob:none" &&
+	test_cmp_config -C client true remote.origin.promisor &&
+	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
+	test_cmp_config -C client 1 core.repositoryformatversion
+'
+
+test_expect_success 'convert shallow clone to partial clone must fail with any extension' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	test_cmp_config -C client 0 core.repositoryformatversion &&
+	git -C client config extensions.partialclone origin &&
+	test_must_fail git -C client fetch --unshallow --filter="blob:none"
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
diff --git a/t/t1090-sparse-checkout-scope.sh b/t/t1090-sparse-checkout-scope.sh
index 40cc004326..f35a73dd20 100755
--- a/t/t1090-sparse-checkout-scope.sh
+++ b/t/t1090-sparse-checkout-scope.sh
@@ -63,7 +63,6 @@ test_expect_success 'in partial clone, sparse checkout only fetches needed blobs
 	git -C server commit -m message &&
 
 	test_config -C client core.sparsecheckout 1 &&
-	test_config -C client extensions.partialclone origin &&
 	echo "!/*" >client/.git/info/sparse-checkout &&
 	echo "/a" >>client/.git/info/sparse-checkout &&
 	git -C client fetch --filter=blob:none origin &&
diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
index 286121d8de..9536d10919 100755
--- a/t/t2404-worktree-config.sh
+++ b/t/t2404-worktree-config.sh
@@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
 '
 
 test_expect_success 'enable worktreeConfig extension' '
+	git config core.repositoryformatversion 1 &&
 	git config extensions.worktreeConfig true &&
-	test_cmp_config true extensions.worktreeConfig
+	test_cmp_config true extensions.worktreeConfig &&
+	test_cmp_config 1 core.repositoryformatversion
 '
 
 test_expect_success 'config is shared as before' '
diff --git a/t/t5500-fetch-pack.sh b/t/t5500-fetch-pack.sh
index 8c54e34ef1..0f5ff25179 100755
--- a/t/t5500-fetch-pack.sh
+++ b/t/t5500-fetch-pack.sh
@@ -999,7 +999,6 @@ fetch_filter_blob_limit_zero () {
 	test_config -C "$SERVER" uploadpack.allowfilter 1 &&
 
 	git clone "$URL" client &&
-	test_config -C client extensions.partialclone origin &&
 
 	test_commit -C "$SERVER" two &&
 
diff --git a/t/t5702-protocol-v2.sh b/t/t5702-protocol-v2.sh
index 5039e66dc4..8b27fad6cd 100755
--- a/t/t5702-protocol-v2.sh
+++ b/t/t5702-protocol-v2.sh
@@ -348,7 +348,6 @@ test_expect_success 'partial fetch' '
 	rm -rf client "$(pwd)/trace" &&
 	git init client &&
 	SERVER="file://$(pwd)/server" &&
-	test_config -C client extensions.partialClone "$SERVER" &&
 
 	GIT_TRACE_PACKET="$(pwd)/trace" git -C client -c protocol.version=2 \
 		fetch --filter=blob:none "$SERVER" master:refs/heads/other &&
-- 
2.27.0.rc0.183.gde8f92d652-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-29  0:04             ` [PATCH v4] " Xin Li
@ 2020-05-29  0:41               ` Junio C Hamano
  2020-05-29 18:00                 ` Junio C Hamano
  2020-05-29  1:01               ` Jonathan Nieder
  1 sibling, 1 reply; 30+ messages in thread
From: Junio C Hamano @ 2020-05-29  0:41 UTC (permalink / raw)
  To: Xin Li; +Cc: git, jrnieder, sandals, iankaz

Xin Li <delphij@google.com> writes:

> Retroactively adding filter can be useful for existing shallow clones as
> they allow users to see earlier change histories without downloading all
> git objects in a regular --unshallow fetch.
>
> Without this patch, users can make a clone partial by editing the
> repository configuration to convert the remote into a promisor, like:
>
>   git config core.repositoryFormatVersion 1
>   git config extensions.partialClone origin
>   git fetch --unshallow --filter=blob:none origin
>
> Since the hard part of making this work is already in place and such
> edits can be error-prone, teach Git to perform the required configuration
> change automatically instead.
>
> Instead of bailing out immediately when no promisor is available, make
> the code perform a more precise check for any potential problems
> (extensions became special in repository version 1, while it can have
> any value in version 0, so upgrade should not happen if the repository
> have an unsupported configuration that would render it invalid if we
> upgraded).

Upgrade from v0 to v1 must follow the more strict "no extension" rule,
not "no unknown ones" rule, so the above description must be corrected.
Perhaps like this?

	... so upgrade from version 0 should not happen if the
	repository has ANY extension.  A repository version 1 and
	later make Git fail if there is any unknown extension, so we
	need to fail an upgrade only if there is any extension that
	is unknown to us).

You can drop the second paragraph about upgrading from version 1 to
a later version if you want, as the only interesting use cases in
practice at this point are upgrading from v0 to v1 and staying at v1.

> Signed-off-by: Xin Li <delphij@google.com>
> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>

I think the updated design looks good.  Let's nitpick some styles ;-)

> diff --git a/builtin/fetch.c b/builtin/fetch.c
> index b5788c16bf..3347d578ea 100644
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  	if (depth || deepen_since || deepen_not.nr)
>  		deepen = 1;
>  
> -	if (filter_options.choice && !has_promisor_remote())
> -		die("--filter can only be used when extensions.partialClone is set");
> -
>  	if (all) {
>  		if (argc == 1)
>  			die(_("fetch --all does not take a repository argument"));
> diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
> index 95d0882417..95669815d4 100644
> --- a/builtin/sparse-checkout.c
> +++ b/builtin/sparse-checkout.c
> @@ -249,6 +249,8 @@ static int set_config(enum sparse_checkout_mode mode)
>  {
>  	const char *config_path;
>  
> +	if (upgrade_repository_format(1) < 0)
> +		die(_("unable to upgrade repository format to enable worktreeConfig"));
>  	if (git_config_set_gently("extensions.worktreeConfig", "true")) {
>  		error(_("failed to set extensions.worktreeConfig setting"));
>  		return 1;

OK.


> diff --git a/cache.h b/cache.h
> index 0f0485ecfe..66dcd2f219 100644
> --- a/cache.h
> +++ b/cache.h
> @@ -1042,6 +1042,7 @@ struct repository_format {
>  	int worktree_config;
>  	int is_bare;
>  	int hash_algo;
> +	int has_extensions;
>  	char *work_tree;
>  	struct string_list unknown_extensions;
>  };
> @@ -1056,6 +1057,7 @@ struct repository_format {
>  	.version = -1, \
>  	.is_bare = -1, \
>  	.hash_algo = GIT_HASH_SHA1, \
> +	.has_extensions = 0, \

I am on the fence between "explicitly initializing to zero value is
pointless, especially when we use .designated_initializer" and
"especially with .designated_initializer, it adds a documentation
value to explicitly initialize a field to its zero value".  Unless
other reviewers weigh in, I am OK to let this stand as-is. 

>  	.unknown_extensions = STRING_LIST_INIT_DUP, \
>  }
>  
> diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
> index 256bcfbdfe..3553ad7b0a 100644
> --- a/list-objects-filter-options.c
> +++ b/list-objects-filter-options.c
> @@ -326,7 +326,8 @@ void partial_clone_register(
>  
>  	/* Check if it is already registered */
>  	if (!promisor_remote_find(remote)) {
> -		git_config_set("core.repositoryformatversion", "1");
> +		if (upgrade_repository_format(1) < 0)
> +			die(_("unable to upgrade repository format to support partial clone"));

OK.

> diff --git a/repository.h b/repository.h
> index 6534fbb7b3..40cc12c7cf 100644
> --- a/repository.h
> +++ b/repository.h
> @@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
>  
>  void prepare_repo_settings(struct repository *r);
>  
> +/*
> + * Return 1 if upgrade repository format to target_version succeeded,
> + * 0 if no upgrade is necessary; returns -1 when upgrade is not possible.
> + */

Do we want to start with "Return" but say "returns" later?  

	Return 1 if ..., 0 if ..., and -1 when upgrade is not possible.

> +int upgrade_repository_format(int target_version);
> +
>  #endif /* REPOSITORY_H */

> +int upgrade_repository_format(int target_version)
> +{
> +	struct strbuf sb = STRBUF_INIT;
> +	struct strbuf err = STRBUF_INIT;
> +	struct strbuf repo_version = STRBUF_INIT;
> +	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
> +
> +	strbuf_git_common_path(&sb, the_repository, "config");
> +	read_repository_format(&repo_fmt, sb.buf);
> +	strbuf_release(&sb);
> +
> +	if (repo_fmt.version >= target_version)
> +		return 0;

OK.  That's "already up-to-date" case.

> +	if (verify_repository_format_eligibility(&repo_fmt, &err,
> +	    target_version) < 0) {

I do not think this _eligibility thing should be a separate helper
function.  One reason is that its name sounds nonsensical ("eligible
for what?  it deserives to be verified for its repo format?"), another
is it makes it unclear what "upgrade" requires by hiding the logic
inside that decides the eligibility for upgrading.  Besides, there
is only one callsite.

Open-coding the gist of the helper like this:

	if (verify_repository_format(&repo_fmt, &err) < 0 ||
	    (!repo_fmt.version && repo_fmt.has_extensions)) {

should make it a lot clearer to see.  If the repository is unusable
by the version of Git we are running already, or the repository is
v0 and has configuration variable(s) in "extensions.*" section, we
refuse to upgrade.

Which is slightly different from what you did with the three-way
split of verify_repository_format(), which made the "eligibility"
thing not to care about unknown extensions in a repository v1 and
higher.  I actually think we should refuse to update v1 or v2
repository to v3 with a running Git that knows only about v1
(i.e. the repository before upgrading may or may not be something we
understand, and if we do not understand it, we shouldn't touch it).

> +		warning("unable to upgrade repository format from %d to %d: %s",
> +		    repo_fmt.version, target_version, err.buf);
> +		strbuf_release(&err);
> +		return -1;
> +	}
> + ...

And with the suggested change to eliminate "eligibility" helper,
none of the changes below would become necessary, I would think,
so I won't say things like "we do not say 'if (result != 0)';
instead we just say 'if (result)'" ;-)

Thanks.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-29  0:04             ` [PATCH v4] " Xin Li
  2020-05-29  0:41               ` Junio C Hamano
@ 2020-05-29  1:01               ` Jonathan Nieder
  2020-05-29  6:44                 ` [PATCH v5] " Xin Li
                                   ` (2 more replies)
  1 sibling, 3 replies; 30+ messages in thread
From: Jonathan Nieder @ 2020-05-29  1:01 UTC (permalink / raw)
  To: Xin Li; +Cc: git, gitster, sandals, iankaz, Jonathan Tan

(+cc: Jonathan Tan, partial clone expert)
Xin Li wrote:

> Subject: fetch: allow adding a filter after initial clone.

nit: should be no period at the end of the subject line

> Signed-off-by: Xin Li <delphij@google.com>
> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>

Let's see whether I still stand by that. ;-)

[...]
> --- a/setup.c
> +++ b/setup.c
> @@ -13,6 +13,9 @@ static int work_tree_config_is_bogus;
>  static struct startup_info the_startup_info;
>  struct startup_info *startup_info = &the_startup_info;
>  
> +static int verify_repository_format_eligibility(const struct repository_format *,
> +    struct strbuf *, int);

Odd wrapping, but after Junio's suggestion this goes away.

[...]
> +		warning("unable to upgrade repository format from %d to %d: %s",
> +		    repo_fmt.version, target_version, err.buf);

Same wrapping nit: this should use a tab to line up right after the
parenthesis (Git uses 8-space tabs).

[...]
> --- a/t/t1090-sparse-checkout-scope.sh
> +++ b/t/t1090-sparse-checkout-scope.sh
> @@ -63,7 +63,6 @@ test_expect_success 'in partial clone, sparse checkout only fetches needed blobs
>  	git -C server commit -m message &&
>  
>  	test_config -C client core.sparsecheckout 1 &&
> -	test_config -C client extensions.partialclone origin &&

Nice.

>  	echo "!/*" >client/.git/info/sparse-checkout &&
>  	echo "/a" >>client/.git/info/sparse-checkout &&
>  	git -C client fetch --filter=blob:none origin &&
> diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
> index 286121d8de..9536d10919 100755
> --- a/t/t2404-worktree-config.sh
> +++ b/t/t2404-worktree-config.sh
> @@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
>  '
>  
>  test_expect_success 'enable worktreeConfig extension' '
> +	git config core.repositoryformatversion 1 &&
>  	git config extensions.worktreeConfig true &&

Yes, makes sense.  Does this patch need it, or could this go in a
separate patch?

> -	test_cmp_config true extensions.worktreeConfig
> +	test_cmp_config true extensions.worktreeConfig &&
> +	test_cmp_config 1 core.repositoryformatversion

This (both the existing code and the modified version) is strange: we
just set the config, so why are we checking it?

[...]
> --- a/t/t5500-fetch-pack.sh
> +++ b/t/t5500-fetch-pack.sh
> @@ -999,7 +999,6 @@ fetch_filter_blob_limit_zero () {
>  	test_config -C "$SERVER" uploadpack.allowfilter 1 &&
>  
>  	git clone "$URL" client &&
> -	test_config -C client extensions.partialclone origin &&
>  

Nice.

[...]
> --- a/t/t5702-protocol-v2.sh
> +++ b/t/t5702-protocol-v2.sh
> @@ -348,7 +348,6 @@ test_expect_success 'partial fetch' '
>  	rm -rf client "$(pwd)/trace" &&
>  	git init client &&
>  	SERVER="file://$(pwd)/server" &&
> -	test_config -C client extensions.partialClone "$SERVER" &&
>  
>  	GIT_TRACE_PACKET="$(pwd)/trace" git -C client -c protocol.version=2 \
>  		fetch --filter=blob:none "$SERVER" master:refs/heads/other &&

This test is a bit unusual, since it's fetching by URL instead of from
a remote.  Looks like it comes from

	commit ba95710a3bdcb2a80495b1d93a0e482dd69905e1
	Author: Jonathan Tan <jonathantanmy@google.com>
	Date:   Thu May 3 16:46:56 2018 -0700

	    {fetch,upload}-pack: support filter in protocol v2

Jonathan, thoughts about this one?  Is making it set
extensions.partialClone implicitly like the change above a good
change?

With the nits above and the bits Junio described addressed, this would
indeed be
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>

Thanks.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v5] fetch: allow adding a filter after initial clone
  2020-05-29  1:01               ` Jonathan Nieder
@ 2020-05-29  6:44                 ` Xin Li
  2020-05-29  6:54                 ` [PATCH v4] " Xin Li
  2020-05-29 18:06                 ` Junio C Hamano
  2 siblings, 0 replies; 30+ messages in thread
From: Xin Li @ 2020-05-29  6:44 UTC (permalink / raw)
  To: git; +Cc: Xin Li, jrnieder, gitster, sandals, jonathantanmy, iankaz

Retroactively adding filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.

Without this patch, users can make a clone partial by editing the
repository configuration to convert the remote into a promisor, like:

  git config core.repositoryFormatVersion 1
  git config extensions.partialClone origin
  git fetch --unshallow --filter=blob:none origin

Since the hard part of making this work is already in place and such
edits can be error-prone, teach Git to perform the required configuration
change automatically instead.

Instead of bailing out immediately when no promisor is available, make
the code perform a more precise check for any potential problems
(extensions became special in repository version 1, while it can have
any value in version 0, so upgrade from version 0 should not happen if
the repository has ANY extension).

Signed-off-by: Xin Li <delphij@google.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
---
 builtin/fetch.c                  |  3 ---
 builtin/sparse-checkout.c        |  2 ++
 cache.h                          |  1 +
 list-objects-filter-options.c    |  3 ++-
 repository.h                     |  6 +++++
 setup.c                          | 41 +++++++++++++++++++++++++++++---
 t/t0410-partial-clone.sh         | 23 ++++++++++++++++++
 t/t1090-sparse-checkout-scope.sh |  1 -
 t/t2404-worktree-config.sh       |  4 +++-
 t/t5500-fetch-pack.sh            |  1 -
 t/t5702-protocol-v2.sh           |  1 -
 11 files changed, 75 insertions(+), 11 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b5788c16bf..3347d578ea 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
-
 	if (all) {
 		if (argc == 1)
 			die(_("fetch --all does not take a repository argument"));
diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index 95d0882417..95669815d4 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c
@@ -249,6 +249,8 @@ static int set_config(enum sparse_checkout_mode mode)
 {
 	const char *config_path;
 
+	if (upgrade_repository_format(1) < 0)
+		die(_("unable to upgrade repository format to enable worktreeConfig"));
 	if (git_config_set_gently("extensions.worktreeConfig", "true")) {
 		error(_("failed to set extensions.worktreeConfig setting"));
 		return 1;
diff --git a/cache.h b/cache.h
index 0f0485ecfe..e5885cc9ea 100644
--- a/cache.h
+++ b/cache.h
@@ -1042,6 +1042,7 @@ struct repository_format {
 	int worktree_config;
 	int is_bare;
 	int hash_algo;
+	int has_extensions;
 	char *work_tree;
 	struct string_list unknown_extensions;
 };
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 256bcfbdfe..3553ad7b0a 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -326,7 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		git_config_set("core.repositoryformatversion", "1");
+		if (upgrade_repository_format(1) < 0)
+			die(_("unable to upgrade repository format to support partial clone"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git a/repository.h b/repository.h
index 6534fbb7b3..3c1f7d54bd 100644
--- a/repository.h
+++ b/repository.h
@@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 
 void prepare_repo_settings(struct repository *r);
 
+/*
+ * Return 1 if upgrade repository format to target_version succeeded,
+ * 0 if no upgrade is necessary, and -1 when upgrade is not possible.
+ */
+int upgrade_repository_format(int target_version);
+
 #endif /* REPOSITORY_H */
diff --git a/setup.c b/setup.c
index 65fe5ecefb..73bfa98431 100644
--- a/setup.c
+++ b/setup.c
@@ -455,6 +455,7 @@ static int check_repo_format(const char *var, const char *value, void *vdata)
 	if (strcmp(var, "core.repositoryformatversion") == 0)
 		data->version = git_config_int(var, value);
 	else if (skip_prefix(var, "extensions.", &ext)) {
+		data->has_extensions = 1;
 		/*
 		 * record any known extensions here; otherwise,
 		 * we fall through to recording it as unknown, and
@@ -506,9 +507,15 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 		die("%s", err.buf);
 	}
 
-	repository_format_precious_objects = candidate->precious_objects;
-	set_repository_format_partial_clone(candidate->partial_clone);
-	repository_format_worktree_config = candidate->worktree_config;
+	if (candidate->version >= 1) {
+		repository_format_precious_objects = candidate->precious_objects;
+		set_repository_format_partial_clone(candidate->partial_clone);
+		repository_format_worktree_config = candidate->worktree_config;
+	} else {
+		repository_format_precious_objects = 0;
+		set_repository_format_partial_clone(NULL);
+		repository_format_worktree_config = 0;
+	}
 	string_list_clear(&candidate->unknown_extensions, 0);
 
 	if (repository_format_worktree_config) {
@@ -538,6 +545,34 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
+int upgrade_repository_format(int target_version)
+{
+	struct strbuf sb = STRBUF_INIT;
+	struct strbuf err = STRBUF_INIT;
+	struct strbuf repo_version = STRBUF_INIT;
+	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
+
+	strbuf_git_common_path(&sb, the_repository, "config");
+	read_repository_format(&repo_fmt, sb.buf);
+	strbuf_release(&sb);
+
+	if (repo_fmt.version >= target_version)
+		return 0;
+
+	if (verify_repository_format(&repo_fmt, &err) < 0 ||
+	    (!repo_fmt.version && repo_fmt.has_extensions)) {
+		warning("unable to upgrade repository format from %d to %d: %s",
+		    repo_fmt.version, target_version, err.buf);
+		strbuf_release(&err);
+		return -1;
+	}
+
+	strbuf_addf(&repo_version, "%d", target_version);
+	git_config_set("core.repositoryformatversion", repo_version.buf);
+	strbuf_release(&repo_version);
+	return 1;
+}
+
 static void init_repository_format(struct repository_format *format)
 {
 	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index a3988bd4b8..463dc3a8be 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -30,6 +30,29 @@ test_expect_success 'extensions.partialclone without filter' '
 	git -C client fetch origin
 '
 
+test_expect_success 'convert shallow clone to partial clone' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client fetch --unshallow --filter="blob:none" &&
+	test_cmp_config -C client true remote.origin.promisor &&
+	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
+	test_cmp_config -C client 1 core.repositoryformatversion
+'
+
+test_expect_success 'convert shallow clone to partial clone must fail with any extension' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	test_cmp_config -C client 0 core.repositoryformatversion &&
+	git -C client config extensions.partialclone origin &&
+	test_must_fail git -C client fetch --unshallow --filter="blob:none"
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
diff --git a/t/t1090-sparse-checkout-scope.sh b/t/t1090-sparse-checkout-scope.sh
index 40cc004326..f35a73dd20 100755
--- a/t/t1090-sparse-checkout-scope.sh
+++ b/t/t1090-sparse-checkout-scope.sh
@@ -63,7 +63,6 @@ test_expect_success 'in partial clone, sparse checkout only fetches needed blobs
 	git -C server commit -m message &&
 
 	test_config -C client core.sparsecheckout 1 &&
-	test_config -C client extensions.partialclone origin &&
 	echo "!/*" >client/.git/info/sparse-checkout &&
 	echo "/a" >>client/.git/info/sparse-checkout &&
 	git -C client fetch --filter=blob:none origin &&
diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
index 286121d8de..9536d10919 100755
--- a/t/t2404-worktree-config.sh
+++ b/t/t2404-worktree-config.sh
@@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
 '
 
 test_expect_success 'enable worktreeConfig extension' '
+	git config core.repositoryformatversion 1 &&
 	git config extensions.worktreeConfig true &&
-	test_cmp_config true extensions.worktreeConfig
+	test_cmp_config true extensions.worktreeConfig &&
+	test_cmp_config 1 core.repositoryformatversion
 '
 
 test_expect_success 'config is shared as before' '
diff --git a/t/t5500-fetch-pack.sh b/t/t5500-fetch-pack.sh
index 8c54e34ef1..0f5ff25179 100755
--- a/t/t5500-fetch-pack.sh
+++ b/t/t5500-fetch-pack.sh
@@ -999,7 +999,6 @@ fetch_filter_blob_limit_zero () {
 	test_config -C "$SERVER" uploadpack.allowfilter 1 &&
 
 	git clone "$URL" client &&
-	test_config -C client extensions.partialclone origin &&
 
 	test_commit -C "$SERVER" two &&
 
diff --git a/t/t5702-protocol-v2.sh b/t/t5702-protocol-v2.sh
index 5039e66dc4..8b27fad6cd 100755
--- a/t/t5702-protocol-v2.sh
+++ b/t/t5702-protocol-v2.sh
@@ -348,7 +348,6 @@ test_expect_success 'partial fetch' '
 	rm -rf client "$(pwd)/trace" &&
 	git init client &&
 	SERVER="file://$(pwd)/server" &&
-	test_config -C client extensions.partialClone "$SERVER" &&
 
 	GIT_TRACE_PACKET="$(pwd)/trace" git -C client -c protocol.version=2 \
 		fetch --filter=blob:none "$SERVER" master:refs/heads/other &&
-- 
2.27.0.rc0.183.gde8f92d652-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-29  1:01               ` Jonathan Nieder
  2020-05-29  6:44                 ` [PATCH v5] " Xin Li
@ 2020-05-29  6:54                 ` Xin Li
  2020-05-29 18:06                 ` Junio C Hamano
  2 siblings, 0 replies; 30+ messages in thread
From: Xin Li @ 2020-05-29  6:54 UTC (permalink / raw)
  To: Jonathan Nieder
  Cc: git, Junio C Hamano, brian m. carlson, Ian Kasprzak, Jonathan Tan

Regarding t/t2404-worktree-config.sh:

On Thu, May 28, 2020 at 6:01 PM Jonathan Nieder <jrnieder@gmail.com> wrote:
> Xin Li wrote:
>
> > diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
> > index 286121d8de..9536d10919 100755
> > --- a/t/t2404-worktree-config.sh
> > +++ b/t/t2404-worktree-config.sh
> > @@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
> >  '
> >
> >  test_expect_success 'enable worktreeConfig extension' '
> > +     git config core.repositoryformatversion 1 &&
> >       git config extensions.worktreeConfig true &&
>
> Yes, makes sense.  Does this patch need it, or could this go in a
> separate patch?

Yes, this patch needs setting repositoryformatversion to 1 as we would
no longer recognize extensions.worktreeConfig=true on version 0
repositories.

> > -     test_cmp_config true extensions.worktreeConfig
> > +     test_cmp_config true extensions.worktreeConfig &&
> > +     test_cmp_config 1 core.repositoryformatversion
>
> This (both the existing code and the modified version) is strange: we
> just set the config, so why are we checking it?

The check was mainly to match the existing pattern (which sets
extensions.worktreeConfig and immediately asserts that they were set).
These assertions are not strictly necessary but are harmless, so I
don't feel strongly about keeping or removing them.

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-29  0:41               ` Junio C Hamano
@ 2020-05-29 18:00                 ` Junio C Hamano
  0 siblings, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-05-29 18:00 UTC (permalink / raw)
  To: Xin Li; +Cc: git, jrnieder, sandals, iankaz

Junio C Hamano <gitster@pobox.com> writes:

> Which is slightly different from what you did with the three-way
> split of verify_repository_format(), which made the "eligibility"
> thing not to care about unknown extensions in a repository v1 and
> higher.  I actually think we should refuse to update v1 or v2
> repository to v3 with a running Git that knows only about v1
> (i.e. the repository before upgrading may or may not be something we
> understand, and if we do not understand it, we shouldn't touch it).

It does not change the conclusion, but I think the above sample
situation would not make much sense---a caller that asks this
function to upgrade the repository to v3 when the version of Git it
is linked in does not understand v3 is simply buggy.

But we should still refuse to update v1 to v2 with a version of Git
that understands v2 if the repository has some extension that we do
not know about, so 

 (1) if upgrading from v0, there must be no "extensions.*"; and

 (2) if upgrading from other versions, there must be no
     "extensions.*" we do not recognise.

I suggested would still be the reasonably defensive rule.

Thanks.



^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v4] fetch: allow adding a filter after initial clone.
  2020-05-29  1:01               ` Jonathan Nieder
  2020-05-29  6:44                 ` [PATCH v5] " Xin Li
  2020-05-29  6:54                 ` [PATCH v4] " Xin Li
@ 2020-05-29 18:06                 ` Junio C Hamano
  2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
  2 siblings, 1 reply; 30+ messages in thread
From: Junio C Hamano @ 2020-05-29 18:06 UTC (permalink / raw)
  To: Jonathan Nieder; +Cc: Xin Li, git, sandals, iankaz, Jonathan Tan

Jonathan Nieder <jrnieder@gmail.com> writes:

>> diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
>> index 286121d8de..9536d10919 100755
>> --- a/t/t2404-worktree-config.sh
>> +++ b/t/t2404-worktree-config.sh
>> @@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
>>  '
>>  
>>  test_expect_success 'enable worktreeConfig extension' '
>> +	git config core.repositoryformatversion 1 &&
>>  	git config extensions.worktreeConfig true &&
>
> Yes, makes sense.  Does this patch need it, or could this go in a
> separate patch?

I think it is the consequence of unrelated change in this hunk:

@@ -506,9 +510,15 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 		die("%s", err.buf);
 	}
 
-	repository_format_precious_objects = candidate->precious_objects;
-	set_repository_format_partial_clone(candidate->partial_clone);
-	repository_format_worktree_config = candidate->worktree_config;
+	if (candidate->version >= 1) {
+		repository_format_precious_objects = candidate->precious_objects;
+		set_repository_format_partial_clone(candidate->partial_clone);
+		repository_format_worktree_config = candidate->worktree_config;
+	} else {
+		repository_format_precious_objects = 0;
+		set_repository_format_partial_clone(NULL);
+		repository_format_worktree_config = 0;
+	}
 	string_list_clear(&candidate->unknown_extensions, 0);
 
 	if (repository_format_worktree_config) {

We used to honor extensions.*, as long as the version of Git
understand it, even in a repository whose version is still v0.

I am not sure if this backward incompatible change is necessary for
the purpose of "allow safe upgrade of repository format version"
topic, but as you hinted, it does smell like it belongs to a
separate (and much larger and potentially controversial) patch.

Thanks for carefully reading.


^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v6 0/4] fetch: allow adding a filter after initial clone
  2020-05-29 18:06                 ` Junio C Hamano
@ 2020-06-05  9:10                   ` Xin Li
  2020-06-05  9:10                     ` [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade Xin Li
                                       ` (3 more replies)
  0 siblings, 4 replies; 30+ messages in thread
From: Xin Li @ 2020-06-05  9:10 UTC (permalink / raw)
  To: git
  Cc: Xin Li, Junio C Hamano, Jonathan Nieder, sandals, iankaz, Jonathan Tan

This series of patch makes fetch to allow adding a filter after initial
clone, useful for existing shallow clones as they allow users to see
earlier change histories without downloading all git objects in a regular
--unshallow fetch.

This version of patchset is mostly identical to the v5 draft but splitted
into smaller individual patches.

Please note that the last patch is an incompatible change that would
make extensions stop working for version 0 repositories.  Currently,
unknown extensions are silently ignored for version 0, which means the
user may see undesirable result when upgraded to a new Git version.

Xin Li (4):
  repository: add a helper function to perform repository format upgrade
  fetch: allow adding a filter after initial clone
  sparse-checkout: upgrade repository to version 1 when enabling
    extension
  check_repository_format_gently(): refuse extensions for old
    repositories

 builtin/fetch.c                  |  3 ---
 builtin/sparse-checkout.c        |  2 ++
 cache.h                          |  1 +
 list-objects-filter-options.c    |  3 ++-
 repository.h                     |  6 +++++
 setup.c                          | 41 +++++++++++++++++++++++++++++---
 t/t0410-partial-clone.sh         | 23 ++++++++++++++++++
 t/t1090-sparse-checkout-scope.sh |  1 -
 t/t2404-worktree-config.sh       |  4 +++-
 t/t5500-fetch-pack.sh            |  1 -
 t/t5702-protocol-v2.sh           |  1 -
 11 files changed, 75 insertions(+), 11 deletions(-)

-- 
2.27.0.278.ge193c7cf3a9-goog


^ permalink raw reply	[flat|nested] 30+ messages in thread

* [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade
  2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
@ 2020-06-05  9:10                     ` Xin Li
  2020-06-05 19:12                       ` Junio C Hamano
  2020-06-05  9:10                     ` [PATCH v6 2/4] fetch: allow adding a filter after initial clone Xin Li
                                       ` (2 subsequent siblings)
  3 siblings, 1 reply; 30+ messages in thread
From: Xin Li @ 2020-06-05  9:10 UTC (permalink / raw)
  To: git; +Cc: Xin Li

In version 1 of repository format, "extensions" gained special meaning
and it is safer to avoid upgrading when there are pre-existing
extensions.

Make list-objects-filter to use the helper function instead of setting
repository version directly as a prerequisite of exposing the upgrade
capability.

Signed-off-by: Xin Li <delphij@google.com>
---
 cache.h                       |  1 +
 list-objects-filter-options.c |  3 ++-
 repository.h                  |  6 ++++++
 setup.c                       | 29 +++++++++++++++++++++++++++++
 4 files changed, 38 insertions(+), 1 deletion(-)

diff --git a/cache.h b/cache.h
index 0f0485ecfe..e5885cc9ea 100644
--- a/cache.h
+++ b/cache.h
@@ -1042,6 +1042,7 @@ struct repository_format {
 	int worktree_config;
 	int is_bare;
 	int hash_algo;
+	int has_extensions;
 	char *work_tree;
 	struct string_list unknown_extensions;
 };
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 256bcfbdfe..3553ad7b0a 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -326,7 +326,8 @@ void partial_clone_register(
 
 	/* Check if it is already registered */
 	if (!promisor_remote_find(remote)) {
-		git_config_set("core.repositoryformatversion", "1");
+		if (upgrade_repository_format(1) < 0)
+			die(_("unable to upgrade repository format to support partial clone"));
 
 		/* Add promisor config for the remote */
 		cfg_name = xstrfmt("remote.%s.promisor", remote);
diff --git a/repository.h b/repository.h
index 6534fbb7b3..3c1f7d54bd 100644
--- a/repository.h
+++ b/repository.h
@@ -196,4 +196,10 @@ void repo_update_index_if_able(struct repository *, struct lock_file *);
 
 void prepare_repo_settings(struct repository *r);
 
+/*
+ * Return 1 if upgrade repository format to target_version succeeded,
+ * 0 if no upgrade is necessary, and -1 when upgrade is not possible.
+ */
+int upgrade_repository_format(int target_version);
+
 #endif /* REPOSITORY_H */
diff --git a/setup.c b/setup.c
index 65fe5ecefb..597b41b822 100644
--- a/setup.c
+++ b/setup.c
@@ -455,6 +455,7 @@ static int check_repo_format(const char *var, const char *value, void *vdata)
 	if (strcmp(var, "core.repositoryformatversion") == 0)
 		data->version = git_config_int(var, value);
 	else if (skip_prefix(var, "extensions.", &ext)) {
+		data->has_extensions = 1;
 		/*
 		 * record any known extensions here; otherwise,
 		 * we fall through to recording it as unknown, and
@@ -538,6 +539,34 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 	return 0;
 }
 
+int upgrade_repository_format(int target_version)
+{
+	struct strbuf sb = STRBUF_INIT;
+	struct strbuf err = STRBUF_INIT;
+	struct strbuf repo_version = STRBUF_INIT;
+	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
+
+	strbuf_git_common_path(&sb, the_repository, "config");
+	read_repository_format(&repo_fmt, sb.buf);
+	strbuf_release(&sb);
+
+	if (repo_fmt.version >= target_version)
+		return 0;
+
+	if (verify_repository_format(&repo_fmt, &err) < 0 ||
+	    (!repo_fmt.version && repo_fmt.has_extensions)) {
+		warning("unable to upgrade repository format from %d to %d: %s",
+			repo_fmt.version, target_version, err.buf);
+		strbuf_release(&err);
+		return -1;
+	}
+
+	strbuf_addf(&repo_version, "%d", target_version);
+	git_config_set("core.repositoryformatversion", repo_version.buf);
+	strbuf_release(&repo_version);
+	return 1;
+}
+
 static void init_repository_format(struct repository_format *format)
 {
 	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;
-- 
2.27.0.278.ge193c7cf3a9-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* [PATCH v6 2/4] fetch: allow adding a filter after initial clone
  2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
  2020-06-05  9:10                     ` [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade Xin Li
@ 2020-06-05  9:10                     ` Xin Li
  2020-06-05 19:15                       ` Junio C Hamano
  2020-06-05  9:10                     ` [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension Xin Li
  2020-06-05  9:10                     ` [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories Xin Li
  3 siblings, 1 reply; 30+ messages in thread
From: Xin Li @ 2020-06-05  9:10 UTC (permalink / raw)
  To: git; +Cc: Xin Li

Retroactively adding a filter can be useful for existing shallow clones as
they allow users to see earlier change histories without downloading all
git objects in a regular --unshallow fetch.

Without this patch, users can make a clone partial by editing the
repository configuration to convert the remote into a promisor, like:

  git config core.repositoryFormatVersion 1
  git config extensions.partialClone origin
  git fetch --unshallow --filter=blob:none origin

Since the hard part of making this work is already in place and such
edits can be error-prone, teach Git to perform the required configuration
change automatically instead.

Note that this change does not modify the existing git behavior which
recognizes setting extensions.partialClone without changing
repositoryFormatVersion.

Signed-off-by: Xin Li <delphij@google.com>
---
 builtin/fetch.c                  |  3 ---
 t/t0410-partial-clone.sh         | 12 ++++++++++++
 t/t1090-sparse-checkout-scope.sh |  1 -
 t/t5500-fetch-pack.sh            |  1 -
 t/t5702-protocol-v2.sh           |  1 -
 5 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b5788c16bf..3347d578ea 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
 	if (depth || deepen_since || deepen_not.nr)
 		deepen = 1;
 
-	if (filter_options.choice && !has_promisor_remote())
-		die("--filter can only be used when extensions.partialClone is set");
-
 	if (all) {
 		if (argc == 1)
 			die(_("fetch --all does not take a repository argument"));
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index a3988bd4b8..16ad000382 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -30,6 +30,18 @@ test_expect_success 'extensions.partialclone without filter' '
 	git -C client fetch origin
 '
 
+test_expect_success 'convert shallow clone to partial clone' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	git -C client fetch --unshallow --filter="blob:none" &&
+	test_cmp_config -C client true remote.origin.promisor &&
+	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
+	test_cmp_config -C client 1 core.repositoryformatversion
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
diff --git a/t/t1090-sparse-checkout-scope.sh b/t/t1090-sparse-checkout-scope.sh
index 40cc004326..f35a73dd20 100755
--- a/t/t1090-sparse-checkout-scope.sh
+++ b/t/t1090-sparse-checkout-scope.sh
@@ -63,7 +63,6 @@ test_expect_success 'in partial clone, sparse checkout only fetches needed blobs
 	git -C server commit -m message &&
 
 	test_config -C client core.sparsecheckout 1 &&
-	test_config -C client extensions.partialclone origin &&
 	echo "!/*" >client/.git/info/sparse-checkout &&
 	echo "/a" >>client/.git/info/sparse-checkout &&
 	git -C client fetch --filter=blob:none origin &&
diff --git a/t/t5500-fetch-pack.sh b/t/t5500-fetch-pack.sh
index 8c54e34ef1..0f5ff25179 100755
--- a/t/t5500-fetch-pack.sh
+++ b/t/t5500-fetch-pack.sh
@@ -999,7 +999,6 @@ fetch_filter_blob_limit_zero () {
 	test_config -C "$SERVER" uploadpack.allowfilter 1 &&
 
 	git clone "$URL" client &&
-	test_config -C client extensions.partialclone origin &&
 
 	test_commit -C "$SERVER" two &&
 
diff --git a/t/t5702-protocol-v2.sh b/t/t5702-protocol-v2.sh
index 5039e66dc4..8b27fad6cd 100755
--- a/t/t5702-protocol-v2.sh
+++ b/t/t5702-protocol-v2.sh
@@ -348,7 +348,6 @@ test_expect_success 'partial fetch' '
 	rm -rf client "$(pwd)/trace" &&
 	git init client &&
 	SERVER="file://$(pwd)/server" &&
-	test_config -C client extensions.partialClone "$SERVER" &&
 
 	GIT_TRACE_PACKET="$(pwd)/trace" git -C client -c protocol.version=2 \
 		fetch --filter=blob:none "$SERVER" master:refs/heads/other &&
-- 
2.27.0.278.ge193c7cf3a9-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension
  2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
  2020-06-05  9:10                     ` [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade Xin Li
  2020-06-05  9:10                     ` [PATCH v6 2/4] fetch: allow adding a filter after initial clone Xin Li
@ 2020-06-05  9:10                     ` Xin Li
  2020-06-05 19:21                       ` Junio C Hamano
  2020-06-05  9:10                     ` [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories Xin Li
  3 siblings, 1 reply; 30+ messages in thread
From: Xin Li @ 2020-06-05  9:10 UTC (permalink / raw)
  To: git; +Cc: Xin Li

The 'extensions' configuration variable gets special meaning in the new
repository version, so when enabling the extension we should upgrade the
repository to version 1.

Signed-off-by: Xin Li <delphij@google.com>
---
 builtin/sparse-checkout.c  | 2 ++
 t/t2404-worktree-config.sh | 4 +++-
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
index 95d0882417..95669815d4 100644
--- a/builtin/sparse-checkout.c
+++ b/builtin/sparse-checkout.c
@@ -249,6 +249,8 @@ static int set_config(enum sparse_checkout_mode mode)
 {
 	const char *config_path;
 
+	if (upgrade_repository_format(1) < 0)
+		die(_("unable to upgrade repository format to enable worktreeConfig"));
 	if (git_config_set_gently("extensions.worktreeConfig", "true")) {
 		error(_("failed to set extensions.worktreeConfig setting"));
 		return 1;
diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
index 286121d8de..9536d10919 100755
--- a/t/t2404-worktree-config.sh
+++ b/t/t2404-worktree-config.sh
@@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
 '
 
 test_expect_success 'enable worktreeConfig extension' '
+	git config core.repositoryformatversion 1 &&
 	git config extensions.worktreeConfig true &&
-	test_cmp_config true extensions.worktreeConfig
+	test_cmp_config true extensions.worktreeConfig &&
+	test_cmp_config 1 core.repositoryformatversion
 '
 
 test_expect_success 'config is shared as before' '
-- 
2.27.0.278.ge193c7cf3a9-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories
  2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
                                       ` (2 preceding siblings ...)
  2020-06-05  9:10                     ` [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension Xin Li
@ 2020-06-05  9:10                     ` Xin Li
  2020-06-08 16:59                       ` Junio C Hamano
  3 siblings, 1 reply; 30+ messages in thread
From: Xin Li @ 2020-06-05  9:10 UTC (permalink / raw)
  To: git; +Cc: Xin Li

Previously, extensions were recognized regardless of repository format
version.  If the user sets an undefined "extensions" value on a
repository of version 0 and that value is used by a future git version,
they might get an undesired result.

Because all extensions now also upgrade repository versions, tightening
the check would help avoid this for future extensions.

Signed-off-by: Xin Li <delphij@google.com>
---
 setup.c                  | 12 +++++++++---
 t/t0410-partial-clone.sh | 11 +++++++++++
 2 files changed, 20 insertions(+), 3 deletions(-)

diff --git a/setup.c b/setup.c
index 597b41b822..eb066db6d8 100644
--- a/setup.c
+++ b/setup.c
@@ -507,9 +507,15 @@ static int check_repository_format_gently(const char *gitdir, struct repository_
 		die("%s", err.buf);
 	}
 
-	repository_format_precious_objects = candidate->precious_objects;
-	set_repository_format_partial_clone(candidate->partial_clone);
-	repository_format_worktree_config = candidate->worktree_config;
+	if (candidate->version >= 1) {
+		repository_format_precious_objects = candidate->precious_objects;
+		set_repository_format_partial_clone(candidate->partial_clone);
+		repository_format_worktree_config = candidate->worktree_config;
+	} else {
+		repository_format_precious_objects = 0;
+		set_repository_format_partial_clone(NULL);
+		repository_format_worktree_config = 0;
+	}
 	string_list_clear(&candidate->unknown_extensions, 0);
 
 	if (repository_format_worktree_config) {
diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
index 16ad000382..463dc3a8be 100755
--- a/t/t0410-partial-clone.sh
+++ b/t/t0410-partial-clone.sh
@@ -42,6 +42,17 @@ test_expect_success 'convert shallow clone to partial clone' '
 	test_cmp_config -C client 1 core.repositoryformatversion
 '
 
+test_expect_success 'convert shallow clone to partial clone must fail with any extension' '
+	rm -fr server client &&
+	test_create_repo server &&
+	test_commit -C server my_commit 1 &&
+	test_commit -C server my_commit2 1 &&
+	git clone --depth=1 "file://$(pwd)/server" client &&
+	test_cmp_config -C client 0 core.repositoryformatversion &&
+	git -C client config extensions.partialclone origin &&
+	test_must_fail git -C client fetch --unshallow --filter="blob:none"
+'
+
 test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
 	rm -rf repo &&
 	test_create_repo repo &&
-- 
2.27.0.278.ge193c7cf3a9-goog


^ permalink raw reply related	[flat|nested] 30+ messages in thread

* Re: [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade
  2020-06-05  9:10                     ` [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade Xin Li
@ 2020-06-05 19:12                       ` Junio C Hamano
  0 siblings, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-06-05 19:12 UTC (permalink / raw)
  To: Xin Li; +Cc: git

Xin Li <delphij@google.com> writes:

> In version 1 of repository format, "extensions" gained special meaning
> and it is safer to avoid upgrading when there are pre-existing
> extensions.

I am tempted to suggest s/upgrading/& from version 0/ but if we
assume everybody knows there are only v0 and v1, that may be
unnecessary.  I dunno.

> Make list-objects-filter to use the helper function instead of setting
> repository version directly as a prerequisite of exposing the upgrade
> capability.
>
> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  cache.h                       |  1 +
>  list-objects-filter-options.c |  3 ++-
>  repository.h                  |  6 ++++++
>  setup.c                       | 29 +++++++++++++++++++++++++++++
>  4 files changed, 38 insertions(+), 1 deletion(-)
>
> diff --git a/cache.h b/cache.h
> index 0f0485ecfe..e5885cc9ea 100644
> --- a/cache.h
> +++ b/cache.h
> @@ -1042,6 +1042,7 @@ struct repository_format {
>  	int worktree_config;
>  	int is_bare;
>  	int hash_algo;
> +	int has_extensions;
>  	char *work_tree;
>  	struct string_list unknown_extensions;
>  };
> diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
> index 256bcfbdfe..3553ad7b0a 100644
> --- a/list-objects-filter-options.c
> +++ b/list-objects-filter-options.c
> @@ -326,7 +326,8 @@ void partial_clone_register(
>  
>  	/* Check if it is already registered */
>  	if (!promisor_remote_find(remote)) {
> -		git_config_set("core.repositoryformatversion", "1");
> +		if (upgrade_repository_format(1) < 0)
> +			die(_("unable to upgrade repository format to support partial clone"));

The idea is that builtin/fetch.c calls this before the actual fetch
operation happens, which sounds sensible.


The other caller of this function is in builtin/clone.c; at that
point, we have set up our $GIT_DIR/config file enough to be able to
write the refspec into it, so we should be able to tell what version
number we wrote to the repository.  Could we have written version 0
there in today's code, though (just being curious, as we should be
able to upgrade it to version 1 with this code)?

> +int upgrade_repository_format(int target_version)
> +{
> +	struct strbuf sb = STRBUF_INIT;
> +	struct strbuf err = STRBUF_INIT;
> +	struct strbuf repo_version = STRBUF_INIT;
> +	struct repository_format repo_fmt = REPOSITORY_FORMAT_INIT;
> +
> +	strbuf_git_common_path(&sb, the_repository, "config");
> +	read_repository_format(&repo_fmt, sb.buf);
> +	strbuf_release(&sb);
> +
> +	if (repo_fmt.version >= target_version)
> +		return 0;
> +
> +	if (verify_repository_format(&repo_fmt, &err) < 0 ||
> +	    (!repo_fmt.version && repo_fmt.has_extensions)) {

"If we do not understand the extensions, or if the original is v0
and has any extensions, we shouldn't be upgrading"---makes sense.

> +		warning("unable to upgrade repository format from %d to %d: %s",
> +			repo_fmt.version, target_version, err.buf);
> +		strbuf_release(&err);
> +		return -1;
> +	}
> +
> +	strbuf_addf(&repo_version, "%d", target_version);
> +	git_config_set("core.repositoryformatversion", repo_version.buf);
> +	strbuf_release(&repo_version);
> +	return 1;
> +}
> +
>  static void init_repository_format(struct repository_format *format)
>  {
>  	const struct repository_format fresh = REPOSITORY_FORMAT_INIT;

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v6 2/4] fetch: allow adding a filter after initial clone
  2020-06-05  9:10                     ` [PATCH v6 2/4] fetch: allow adding a filter after initial clone Xin Li
@ 2020-06-05 19:15                       ` Junio C Hamano
  0 siblings, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-06-05 19:15 UTC (permalink / raw)
  To: Xin Li; +Cc: git

Xin Li <delphij@google.com> writes:

> Retroactively adding a filter can be useful for existing shallow clones as
> they allow users to see earlier change histories without downloading all
> git objects in a regular --unshallow fetch.
>
> Without this patch, users can make a clone partial by editing the
> repository configuration to convert the remote into a promisor, like:
>
>   git config core.repositoryFormatVersion 1
>   git config extensions.partialClone origin
>   git fetch --unshallow --filter=blob:none origin
>
> Since the hard part of making this work is already in place and such
> edits can be error-prone, teach Git to perform the required configuration
> change automatically instead.
>
> Note that this change does not modify the existing git behavior which
> recognizes setting extensions.partialClone without changing
> repositoryFormatVersion.

Hmph, I am not sure which part of the above corresponds to the
removal of "in a repository without extensions.partialClone you
cannot use the --filter option" logic.  With or without this patch,
users can butcher $GIT_DIR/config and pretend that the repository
was created with the pertial clone extension from the beginning, but
this step is not about automating those tasks, it seems.

>
> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  builtin/fetch.c                  |  3 ---
>  t/t0410-partial-clone.sh         | 12 ++++++++++++
>  t/t1090-sparse-checkout-scope.sh |  1 -
>  t/t5500-fetch-pack.sh            |  1 -
>  t/t5702-protocol-v2.sh           |  1 -
>  5 files changed, 12 insertions(+), 6 deletions(-)
>
> diff --git a/builtin/fetch.c b/builtin/fetch.c
> index b5788c16bf..3347d578ea 100644
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -1790,9 +1790,6 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  	if (depth || deepen_since || deepen_not.nr)
>  		deepen = 1;
>  
> -	if (filter_options.choice && !has_promisor_remote())
> -		die("--filter can only be used when extensions.partialClone is set");
> -
>  	if (all) {
>  		if (argc == 1)
>  			die(_("fetch --all does not take a repository argument"));
> diff --git a/t/t0410-partial-clone.sh b/t/t0410-partial-clone.sh
> index a3988bd4b8..16ad000382 100755
> --- a/t/t0410-partial-clone.sh
> +++ b/t/t0410-partial-clone.sh
> @@ -30,6 +30,18 @@ test_expect_success 'extensions.partialclone without filter' '
>  	git -C client fetch origin
>  '
>  
> +test_expect_success 'convert shallow clone to partial clone' '
> +	rm -fr server client &&
> +	test_create_repo server &&
> +	test_commit -C server my_commit 1 &&
> +	test_commit -C server my_commit2 1 &&
> +	git clone --depth=1 "file://$(pwd)/server" client &&
> +	git -C client fetch --unshallow --filter="blob:none" &&
> +	test_cmp_config -C client true remote.origin.promisor &&
> +	test_cmp_config -C client blob:none remote.origin.partialclonefilter &&
> +	test_cmp_config -C client 1 core.repositoryformatversion
> +'
> +
>  test_expect_success 'missing reflog object, but promised by a commit, passes fsck' '
>  	rm -rf repo &&
>  	test_create_repo repo &&
> diff --git a/t/t1090-sparse-checkout-scope.sh b/t/t1090-sparse-checkout-scope.sh
> index 40cc004326..f35a73dd20 100755
> --- a/t/t1090-sparse-checkout-scope.sh
> +++ b/t/t1090-sparse-checkout-scope.sh
> @@ -63,7 +63,6 @@ test_expect_success 'in partial clone, sparse checkout only fetches needed blobs
>  	git -C server commit -m message &&
>  
>  	test_config -C client core.sparsecheckout 1 &&
> -	test_config -C client extensions.partialclone origin &&
>  	echo "!/*" >client/.git/info/sparse-checkout &&
>  	echo "/a" >>client/.git/info/sparse-checkout &&
>  	git -C client fetch --filter=blob:none origin &&
> diff --git a/t/t5500-fetch-pack.sh b/t/t5500-fetch-pack.sh
> index 8c54e34ef1..0f5ff25179 100755
> --- a/t/t5500-fetch-pack.sh
> +++ b/t/t5500-fetch-pack.sh
> @@ -999,7 +999,6 @@ fetch_filter_blob_limit_zero () {
>  	test_config -C "$SERVER" uploadpack.allowfilter 1 &&
>  
>  	git clone "$URL" client &&
> -	test_config -C client extensions.partialclone origin &&
>  
>  	test_commit -C "$SERVER" two &&
>  
> diff --git a/t/t5702-protocol-v2.sh b/t/t5702-protocol-v2.sh
> index 5039e66dc4..8b27fad6cd 100755
> --- a/t/t5702-protocol-v2.sh
> +++ b/t/t5702-protocol-v2.sh
> @@ -348,7 +348,6 @@ test_expect_success 'partial fetch' '
>  	rm -rf client "$(pwd)/trace" &&
>  	git init client &&
>  	SERVER="file://$(pwd)/server" &&
> -	test_config -C client extensions.partialClone "$SERVER" &&
>  
>  	GIT_TRACE_PACKET="$(pwd)/trace" git -C client -c protocol.version=2 \
>  		fetch --filter=blob:none "$SERVER" master:refs/heads/other &&

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension
  2020-06-05  9:10                     ` [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension Xin Li
@ 2020-06-05 19:21                       ` Junio C Hamano
  0 siblings, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-06-05 19:21 UTC (permalink / raw)
  To: Xin Li; +Cc: git

Xin Li <delphij@google.com> writes:

> The 'extensions' configuration variable gets special meaning in the new
> repository version, so when enabling the extension we should upgrade the
> repository to version 1.
>
> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  builtin/sparse-checkout.c  | 2 ++
>  t/t2404-worktree-config.sh | 4 +++-
>  2 files changed, 5 insertions(+), 1 deletion(-)

The other place that "extensions.*" is referred to is in
builtin/init-db.c::initialize_repository_version() and it already
makes sure that extensions.objectformat is set only in a repository
whose verseion is GIT_REPO_VERSION_READ (which is 1---by the way I
suspect we want a better name than _READ for the symbol---it wants
to use the highest version supported by this binary), so with this
patch we covered everything, hopefully.

> diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
> index 95d0882417..95669815d4 100644
> --- a/builtin/sparse-checkout.c
> +++ b/builtin/sparse-checkout.c
> @@ -249,6 +249,8 @@ static int set_config(enum sparse_checkout_mode mode)
>  {
>  	const char *config_path;
>  
> +	if (upgrade_repository_format(1) < 0)
> +		die(_("unable to upgrade repository format to enable worktreeConfig"));
>  	if (git_config_set_gently("extensions.worktreeConfig", "true")) {
>  		error(_("failed to set extensions.worktreeConfig setting"));
>  		return 1;
> diff --git a/t/t2404-worktree-config.sh b/t/t2404-worktree-config.sh
> index 286121d8de..9536d10919 100755
> --- a/t/t2404-worktree-config.sh
> +++ b/t/t2404-worktree-config.sh
> @@ -23,8 +23,10 @@ test_expect_success 'config --worktree without extension' '
>  '
>  
>  test_expect_success 'enable worktreeConfig extension' '
> +	git config core.repositoryformatversion 1 &&
>  	git config extensions.worktreeConfig true &&
> -	test_cmp_config true extensions.worktreeConfig
> +	test_cmp_config true extensions.worktreeConfig &&
> +	test_cmp_config 1 core.repositoryformatversion
>  '
>  
>  test_expect_success 'config is shared as before' '

^ permalink raw reply	[flat|nested] 30+ messages in thread

* Re: [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories
  2020-06-05  9:10                     ` [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories Xin Li
@ 2020-06-08 16:59                       ` Junio C Hamano
  0 siblings, 0 replies; 30+ messages in thread
From: Junio C Hamano @ 2020-06-08 16:59 UTC (permalink / raw)
  To: Xin Li; +Cc: git

Xin Li <delphij@google.com> writes:

> Previously, extensions were recognized regardless of repository format
> version.  If the user sets an undefined "extensions" value on a
> repository of version 0 and that value is used by a future git version,
> they might get an undesired result.
>
> Because all extensions now also upgrade repository versions, tightening
> the check would help avoid this for future extensions.
>
> Signed-off-by: Xin Li <delphij@google.com>
> ---
>  setup.c                  | 12 +++++++++---
>  t/t0410-partial-clone.sh | 11 +++++++++++
>  2 files changed, 20 insertions(+), 3 deletions(-)

Thanks.

People, I think this is the only bit that is potentially
controversial with the risk of breaking existing set-up.  
Its potentially negative effect can be seen in the patch to
test in the previous step.

     test_expect_success 'enable worktreeConfig extension' '
    +	git config core.repositoryformatversion 1 &&
        git config extensions.worktreeConfig true &&

It used to be that, by mistake, you could just say

    git config extensions.worktreeConfig true

in an existing repository (with extra worktrees) and expect that
alone is enough to let you use per-worktree configuration files.

With this patch, that is no longer true, because the extensions.*
will not take effect unless you upgrade core.repositoryformatversion
to 1.

It is the right thing to do in the longer term (we could even
consider it a bugfix), and the workaround is simple enough.

Unless the user is somehow using a configuration variable whose name
begins with "extensions." for whatever random purpose, that is.  If
the user's setup uses extensions.bar, there is no workaround
possible other than renaming the variable to put it outside
extensions.* hierarchy, and update whatever tool that is using that
variable to use the renamed variable.

Comments?

^ permalink raw reply	[flat|nested] 30+ messages in thread

end of thread, other threads:[~2020-06-08 17:00 UTC | newest]

Thread overview: 30+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-05-13 20:00 [PATCH] fetch: allow adding a filter after initial clone Xin Li
2020-05-13 20:43 ` Junio C Hamano
2020-05-13 21:41   ` Xin Li
2020-05-13 22:07     ` Junio C Hamano
2020-05-13 22:18     ` Junio C Hamano
2020-05-13 23:44 ` brian m. carlson
2020-05-28  2:53   ` [PATCH v2 0/1] " Xin Li
2020-05-28  2:54   ` [PATCH v2 1/1] " Xin Li
2020-05-28  3:28     ` Jonathan Nieder
2020-05-28  4:08       ` [PATCH v3] " Xin Li
2020-05-28 15:04     ` [PATCH v2 1/1] " Junio C Hamano
2020-05-28 17:19       ` Jonathan Nieder
2020-05-28 19:12         ` Xin Li
2020-05-28 19:17           ` Jonathan Nieder
2020-05-29  0:04             ` [PATCH v4] " Xin Li
2020-05-29  0:41               ` Junio C Hamano
2020-05-29 18:00                 ` Junio C Hamano
2020-05-29  1:01               ` Jonathan Nieder
2020-05-29  6:44                 ` [PATCH v5] " Xin Li
2020-05-29  6:54                 ` [PATCH v4] " Xin Li
2020-05-29 18:06                 ` Junio C Hamano
2020-06-05  9:10                   ` [PATCH v6 0/4] " Xin Li
2020-06-05  9:10                     ` [PATCH v6 1/4] repository: add a helper function to perform repository format upgrade Xin Li
2020-06-05 19:12                       ` Junio C Hamano
2020-06-05  9:10                     ` [PATCH v6 2/4] fetch: allow adding a filter after initial clone Xin Li
2020-06-05 19:15                       ` Junio C Hamano
2020-06-05  9:10                     ` [PATCH v6 3/4] sparse-checkout: upgrade repository to version 1 when enabling extension Xin Li
2020-06-05 19:21                       ` Junio C Hamano
2020-06-05  9:10                     ` [PATCH v6 4/4] check_repository_format_gently(): refuse extensions for old repositories Xin Li
2020-06-08 16:59                       ` Junio C Hamano

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).