git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH 2/5] bundle-uri: create basic file-copy logic

From
Josh Steadmon <steadmon@google.com>
Date
Jul 27, 2022, 22:09 UTC
Message-ID
<YuG3h4AcZjyILvdK@google.com>
In-Reply-To
<7d3159f0d9a29a04b5856982f787f4eddfb91b53.1658781277.git.gitgitgadget@gmail.com>
On 2022.07.25 20:34, Derrick Stolee via GitGitGadget wrote:
Show 15 quoted lines
> From: Derrick Stolee <derrickstolee@github.com>
> 
> Before implementing a way to fetch bundles into a repository, create the
> basic logic. Assume that the URI is actually a file path. Future logic
> will make this more careful to other protocols.
> 
> For now, we also only succeed if the content at the URI is a bundle
> file, not a bundle list. Bundle lists will be implemented in a future
> change.
> 
> Note that the discovery of a temporary filename is slightly racy because
> the odb_mkstemp() relies on the temporary file not existing. With the
> current implementation being limited to file copies, we could replace
> the copy_file() with copy_fd(). The tricky part comes in future changes
> that send the filename to 'git remote-https' and its 'get' capability.
> At that point, we need the file descriptor closed _and_ the file
> unlinked.

Ahh, it looks like this was the point I missed in my previous review. IIUC, we need the file unlinked because http_get_file() will eventually call finalize_object_file() to move a tempfile to the final object name, and that will fail if we have an empty file already in place.

Show 105 quoted lines
> If we were to keep the file descriptor open for the sake of
> normal file copies, then we would pollute the rest of the code for
> little benefit. This is especially the case because we expect that most
> bundle URI use will be based on HTTPS instead of file copies.
> 
> Signed-off-by: Derrick Stolee <derrickstolee@github.com>
> ---
>  Makefile     |   1 +
>  bundle-uri.c | 104 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  bundle-uri.h |  14 +++++++
>  3 files changed, 119 insertions(+)
>  create mode 100644 bundle-uri.c
>  create mode 100644 bundle-uri.h
> 
> diff --git a/Makefile b/Makefile
> index 1624471badc..7d5f48069ea 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -906,6 +906,7 @@ LIB_OBJS += blob.o
>  LIB_OBJS += bloom.o
>  LIB_OBJS += branch.o
>  LIB_OBJS += bulk-checkin.o
> +LIB_OBJS += bundle-uri.o
>  LIB_OBJS += bundle.o
>  LIB_OBJS += cache-tree.o
>  LIB_OBJS += cbtree.o
> diff --git a/bundle-uri.c b/bundle-uri.c
> new file mode 100644
> index 00000000000..b35babc36aa
> --- /dev/null
> +++ b/bundle-uri.c
> @@ -0,0 +1,104 @@
> +#include "cache.h"
> +#include "bundle-uri.h"
> +#include "bundle.h"
> +#include "object-store.h"
> +#include "refs.h"
> +#include "run-command.h"
> +
> +static int find_temp_filename(struct strbuf *name)
> +{
> +	int fd;
> +	/*
> +	 * Find a temporary filename that is available. This is briefly
> +	 * racy, but unlikely to collide.
> +	 */
> +	fd = odb_mkstemp(name, "bundles/tmp_uri_XXXXXX");
> +	if (fd < 0) {
> +		warning(_("failed to create temporary file"));
> +		return -1;
> +	}
> +
> +	close(fd);
> +	unlink(name->buf);
> +	return 0;
> +}
> +
> +static int copy_uri_to_file(const char *file, const char *uri)
> +{
> +	/* File-based URIs only for now. */
> +	return copy_file(file, uri, 0);
> +}
> +
> +static int unbundle_from_file(struct repository *r, const char *file)
> +{
> +	int result = 0;
> +	int bundle_fd;
> +	struct bundle_header header = BUNDLE_HEADER_INIT;
> +	struct string_list_item *refname;
> +	struct strbuf bundle_ref = STRBUF_INIT;
> +	size_t bundle_prefix_len;
> +
> +	if ((bundle_fd = read_bundle_header(file, &header)) < 0)
> +		return 1;
> +
> +	if ((result = unbundle(r, &header, bundle_fd, NULL)))
> +		return 1;
> +
> +	/*
> +	 * Convert all refs/heads/ from the bundle into refs/bundles/
> +	 * in the local repository.
> +	 */
> +	strbuf_addstr(&bundle_ref, "refs/bundles/");
> +	bundle_prefix_len = bundle_ref.len;
> +
> +	for_each_string_list_item(refname, &header.references) {
> +		struct object_id *oid = refname->util;
> +		struct object_id old_oid;
> +		const char *branch_name;
> +		int has_old;
> +
> +		if (!skip_prefix(refname->string, "refs/heads/", &branch_name))
> +			continue;
> +
> +		strbuf_setlen(&bundle_ref, bundle_prefix_len);
> +		strbuf_addstr(&bundle_ref, branch_name);
> +
> +		has_old = !read_ref(bundle_ref.buf, &old_oid);
> +		update_ref("fetched bundle", bundle_ref.buf, oid,
> +			   has_old ? &old_oid : NULL,
> +			   REF_SKIP_OID_VERIFICATION,
> +			   UPDATE_REFS_MSG_ON_ERR);
> +	}
> +
> +	bundle_header_release(&header);
We still also need to release bundle_ref here, right?
Show 54 quoted lines
> +	return result;
> +}
> +
> +int fetch_bundle_uri(struct repository *r, const char *uri)
> +{
> +	int result = 0;
> +	struct strbuf filename = STRBUF_INIT;
> +
> +	if ((result = find_temp_filename(&filename)))
> +		goto cleanup;
> +
> +	if ((result = copy_uri_to_file(filename.buf, uri))) {
> +		warning(_("failed to download bundle from URI '%s'"), uri);
> +		goto cleanup;
> +	}
> +
> +	if ((result = !is_bundle(filename.buf, 0))) {
> +		warning(_("file at URI '%s' is not a bundle"), uri);
> +		goto cleanup;
> +	}
> +
> +	if ((result = unbundle_from_file(r, filename.buf))) {
> +		warning(_("failed to unbundle bundle from URI '%s'"), uri);
> +		goto cleanup;
> +	}
> +
> +cleanup:
> +	unlink(filename.buf);
> +	strbuf_release(&filename);
> +	return result;
> +}
> diff --git a/bundle-uri.h b/bundle-uri.h
> new file mode 100644
> index 00000000000..8a152f1ef14
> --- /dev/null
> +++ b/bundle-uri.h
> @@ -0,0 +1,14 @@
> +#ifndef BUNDLE_URI_H
> +#define BUNDLE_URI_H
> +
> +struct repository;
> +
> +/**
> + * Fetch data from the given 'uri' and unbundle the bundle data found
> + * based on that information.
> + *
> + * Returns non-zero if no bundle information is found at the given 'uri'.
> + */
> +int fetch_bundle_uri(struct repository *r, const char *uri);
> +
> +#endif
> -- 
> gitgitgadget
> 
Previous: Derrick Stolee via GitGitGadgetNext: Derrick Stolee
Message 8 of 36 in “Bundle URIs II: git clone --bundle-uri”
  1. 0/5 Bundle URIs II: git clone --bundle-uriDerrick Stolee via GitGitGadget, Jul 25, 2022
  2. 1/5 remote-curl: add 'get' capabilityDerrick Stolee via GitGitGadget, Jul 25, 2022
  3. Josh SteadmonJul 27, 2022
  4. Ævar Arnfjörð BjarmasonJul 27, 2022
  5. Derrick StoleeAug 1, 2022
  6. Ævar Arnfjörð BjarmasonAug 1, 2022
  7. 2/5 bundle-uri: create basic file-copy logicDerrick Stolee via GitGitGadget, Jul 25, 2022
  8. Josh SteadmonJul 27, 2022
  9. Derrick StoleeAug 1, 2022
  10. 3/5 clone: add --bundle-uri optionDerrick Stolee via GitGitGadget, Jul 25, 2022
  11. 5/5 clone: --bundle-uri cannot be combined with --depthDerrick Stolee via GitGitGadget, Jul 25, 2022
  12. 4/5 bundle-uri: add support for http(s):// and file://Derrick Stolee via GitGitGadget, Jul 25, 2022
  13. Josh SteadmonJul 27, 2022
  14. Derrick StoleeAug 1, 2022
  15. Josh SteadmonJul 27, 2022
  16. Derrick StoleeAug 1, 2022
  17. 0/5 Bundle URIs II: git clone --bundle-uriDerrick Stolee via GitGitGadget, Aug 2, 2022
  18. 1/5 remote-curl: add 'get' capabilityDerrick Stolee via GitGitGadget, Aug 2, 2022
  19. 2/5 bundle-uri: create basic file-copy logicDerrick Stolee via GitGitGadget, Aug 2, 2022
  20. 3/5 clone: add --bundle-uri optionDerrick Stolee via GitGitGadget, Aug 2, 2022
  21. 4/5 bundle-uri: add support for http(s):// and file://Derrick Stolee via GitGitGadget, Aug 2, 2022
  22. Junio C HamanoAug 2, 2022
  23. Derrick StoleeAug 4, 2022
  24. Junio C HamanoAug 4, 2022
  25. 5/5 clone: --bundle-uri cannot be combined with --depthDerrick Stolee via GitGitGadget, Aug 2, 2022
  26. 0/5 Bundle URIs II: git clone --bundle-uriDerrick Stolee via GitGitGadget, Aug 9, 2022
  27. 1/5 remote-curl: add 'get' capabilityDerrick Stolee via GitGitGadget, Aug 9, 2022
  28. 2/5 bundle-uri: create basic file-copy logicDerrick Stolee via GitGitGadget, Aug 9, 2022
  29. 3/5 clone: add --bundle-uri optionDerrick Stolee via GitGitGadget, Aug 9, 2022
  30. Junio C HamanoAug 22, 2022
  31. Derrick StoleeAug 23, 2022
  32. Junio C HamanoAug 24, 2022
  33. 4/5 bundle-uri: add support for http(s):// and file://Derrick Stolee via GitGitGadget, Aug 9, 2022
  34. 4/5 bundle-uri: add support for http(s):// and file://Teng Long, Aug 29, 2022
  35. Derrick StoleeAug 30, 2022
  36. 5/5 clone: --bundle-uri cannot be combined with --depthDerrick Stolee via GitGitGadget, Aug 9, 2022

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.