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

Re: [RFC v2 01/16] Implement a remote helper for svn in C.

From
Junio C Hamano <gitster@pobox.com>
Date
Jul 30, 2012, 16:28 UTC
Message-ID
<7va9yhrz6c.fsf@alter.siamese.dyndns.org>
In-Reply-To
<1343658683-10713-2-git-send-email-florian.achleitner.2.6.31@gmail.com>
Florian Achleitner <florian.achleitner.2.6.31@gmail.com> writes:
Show 39 quoted lines
> Enables basic fetching from subversion repositories. When processing Remote URLs
> starting with svn::, git invokes this remote-helper.
> It starts svnrdump to extract revisions from the subversion repository in the
> 'dump file format', and converts them to a git-fast-import stream using
> the functions of vcs-svn/.
>
> Imported refs are created in a private namespace at refs/svn/<remote-name/master.
> The revision history is imported linearly (no branch detection) and completely,
> i.e. from revision 0 to HEAD.
>
> Signed-off-by: Florian Achleitner <florian.achleitner.2.6.31@gmail.com>
> ---
>  contrib/svn-fe/remote-svn.c |  190 +++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 190 insertions(+)
>  create mode 100644 contrib/svn-fe/remote-svn.c
>
> diff --git a/contrib/svn-fe/remote-svn.c b/contrib/svn-fe/remote-svn.c
> new file mode 100644
> index 0000000..d5c2df8
> --- /dev/null
> +++ b/contrib/svn-fe/remote-svn.c
> @@ -0,0 +1,190 @@
> +
> +#include "cache.h"
> +#include "remote.h"
> +#include "strbuf.h"
> +#include "url.h"
> +#include "exec_cmd.h"
> +#include "run-command.h"
> +#include "svndump.h"
> +#include "argv-array.h"
> +
> +static const char *url;
> +static const char *private_ref;
> +static const char *remote_ref = "refs/heads/master";
> +
> +int cmd_capabilities(struct strbuf *line);
> +int cmd_import(struct strbuf *line);
> +int cmd_list(struct strbuf *line);

How many of these and other symbols are necessary to be visible outside this file?

Show 24 quoted lines
> +typedef int (*input_command_handler)(struct strbuf *);
> +struct input_command_entry {
> +	const char *name;
> +	input_command_handler fct;
> +	unsigned char batchable;	/* whether the command starts or is part of a batch */
> +};
> +
> +static const struct input_command_entry input_command_list[] = {
> +		{ "capabilities", cmd_capabilities, 0 },
> +		{ "import", cmd_import, 1 },
> +		{ "list", cmd_list, 0 },
> +		{ NULL, NULL }
> +};
> +
> +int cmd_capabilities(struct strbuf *line)
> +{
> +	printf("import\n");
> +	printf("refspec %s:%s\n\n", remote_ref, private_ref);
> +	fflush(stdout);
> +	return 0;
> +}
> +
> +static void terminate_batch() {
> +	/* terminate a current batch's fast-import stream */
Style:
	static void terminate_batch(void)
	{
		/* terminate ...
Show 17 quoted lines
> +		printf("done\n");
> +		fflush(stdout);
> +}
> +
> +int cmd_import(struct strbuf *line)
> +{
> +	int code, report_fd;
> +	char *back_pipe_env;
> +	int dumpin_fd;
> +	unsigned int startrev = 0;
> +	struct argv_array svndump_argv = ARGV_ARRAY_INIT;
> +	struct child_process svndump_proc;
> +
> +	/*
> +	 * When the remote-helper is invoked by transport-helper.c it passes the
> +	 * filename of this pipe in the env-var.
> +	 */
s/ it passes/, &/;
> +	back_pipe_env = getenv("GIT_REPORT_FIFO");

Can we name "back pipe", "report fifo" and "report fd" more consistently and descriptively?

What kind of "REPORT" are we talking about here? Is it to carry the contents of 

> +	if (!back_pipe_env) {
> +		die("Cannot get cat-blob-pipe from environment! GIT_REPORT_FIFO has to"
> +				"be set by the caller.");
> +	}
Style: unnecesary {} block around a simple statement.  It is OK to
have such a block early in a series if you add more statements to it
in later steps, but that does not seem to be the case for this patch
series.
Show 6 quoted lines
> +	/*
> +	 * Opening a fifo for reading usually blocks until a writer has opened it too.
> +	 * Opening a fifo for writing usually blocks until a reader has opened it too.
> +	 * Therefore, we open with RDWR on both sides to avoid deadlocks.
> +	 * Especially if there's nothing to do and one pipe end is closed immediately.
> +	 */

This smells somewhat fishy "justification". Are we reading what we wrote to the fifo? Who is expected to come at the other end of the fifo? Is it this process that creates that other process? Perhaps you should open it _after_ spawning the process, telling it to open the same fifo for writing (if you are sitting on the reading end)?

> +	report_fd = open(back_pipe_env, O_RDWR);
> +	if (report_fd < 0) {
> +		die("Unable to open fast-import back-pipe! %s", strerror(errno));
> +	}
Style: (ditto).
Show 7 quoted lines
> +int cmd_list(struct strbuf *line)
> +{
> +	printf("? HEAD\n");
> +	printf("? %s\n\n", remote_ref);
> +	fflush(stdout);
> +	return 0;
> +}

It somehow feels funny that remote_ref (which seems to be hardcoded to "refs/heads/master") and HEAD are listed here, even though the other side is not even in the Git land and the name "master" does not have any significance. The name "refs/heads/master" may be necessary to form the LHS of the refspec in cmd_capabilities(), but it somehow feels more natural to only advertise HEAD here and also futz the refspec to printf("refspec HEAD:%s", private_ref) in cmd_capabilities(). Perhaps you tried it already and it did not work for some reason; I dunno.

Show 10 quoted lines
> +int do_command(struct strbuf *line)
> +{
> +	const struct input_command_entry *p = input_command_list;
> +	static int batch_active;
> +	static struct strbuf batch_command = STRBUF_INIT;
> +	/*
> +	 * import commands can be grouped together in a batch.
> +	 * Batches are ended by \n. If no batch is active the program ends.
> +	 */
> +	if (line->len == 0 ) {
Style: lose the SP before closing parenthesis ")".
Show 5 quoted lines
> +		if (batch_active) {
> +			terminate_batch();
> +			batch_active = 0;
> +			return 0;
> +		}

Is it an error to feed an empty line when batch is not active? How is the error diagnosed? Is the user told about the error, and if so how?

> +		return 1;
> +	}
> +	if (batch_active && strcmp(batch_command.buf, line->buf))
> +		die("Active %s batch interrupted by %s", batch_command.buf, line->buf);

So after issuing "import" that causes batch_active to become true, another command e.g. "list" cannot be issued and will result in this die() unless the batch is concluded by issuing an empty line. Can an "import" be issued while another "import" batch is in effect?

Show 7 quoted lines
> +	for(p = input_command_list; p->name; p++) {
> +		if (!prefixcmp(line->buf, p->name) &&
> +				(strlen(p->name) == line->len || line->buf[strlen(p->name)] == ' ')) {
> +			if (p->batchable) {
> +				batch_active = 1;
> +				strbuf_release(&batch_command);
> +				strbuf_addbuf(&batch_command, line);

Wouldn't it make more sense to get rid of batch_active variable and use the presense of batch_command as the signal for the batch in effect? Your "if (batch_active)" becomes "if (batch_command.len)", and "batch_active = 0" becomes "strbuf_release(&batch_command)".

Show 40 quoted lines
> +			}
> +			return p->fct(line);
> +		}
> +	}
> +	warning("Unknown command '%s'\n", line->buf);
> +	return 0;
> +}
> +
> +int main(int argc, const char **argv)
> +{
> +	struct strbuf buf = STRBUF_INIT;
> +	int nongit;
> +	static struct remote *remote;
> +	const char *url_in;
> +
> +	git_extract_argv0_path(argv[0]);
> +	setup_git_directory_gently(&nongit);
> +	if (argc < 2 || argc > 3) {
> +		usage("git-remote-svn <remote-name> [<url>]");
> +		return 1;
> +	}
> +
> +	remote = remote_get(argv[1]);
> +	url_in = remote->url[0];
> +	if (argc == 3)
> +		url_in = argv[2];
> +
> +	end_url_with_slash(&buf, url_in);
> +	url = strbuf_detach(&buf, NULL);
> +
> +	strbuf_init(&buf, 0);
> +	strbuf_addf(&buf, "refs/svn/%s/master", remote->name);
> +	private_ref = strbuf_detach(&buf, NULL);
> +
> +	while(1) {
> +		if (strbuf_getline(&buf, stdin, '\n') == EOF) {
> +			if (ferror(stdin))
> +				die_errno("Error reading command stream");
> +			else
> +				die_errno("Unexpected end of command stream");

On the else clause, ferror() did not say there was an error. What errno do we see from die_errno() in that case?

Show 11 quoted lines
> +		}
> +		if (do_command(&buf))
> +			break;
> +		strbuf_reset(&buf);
> +	}
> +
> +	strbuf_release(&buf);
> +	free((void*)url);
> +	free((void*)private_ref);
> +	return 0;
> +}
Previous: Florian AchleitnerNext: Florian Achleitner
Message 46 of 47 in “[RFC 00/16] GSOC remote-svn, rewritten patch series”
  1. Florian AchleitnerJul 26, 2012
  2. 01/16 Implement a remote helper for svn in C.Florian Achleitner, Jul 26, 2012
  3. 02/16 Integrate remote-svn into svn-fe/Makefile.Florian Achleitner, Jul 26, 2012
  4. 03/16 Add svndump_init_fd to allow reading dumps from arbitrary FDs.Florian Achleitner, Jul 26, 2012
  5. 04/16 Add cat-blob report fifo from fast-import to remote-helper.Florian Achleitner, Jul 26, 2012
  6. 05/16 remote-svn, vcs-svn: Enable fetching to private refs.Florian Achleitner, Jul 26, 2012
  7. 06/16 Add a symlink 'git-remote-svn' in base dir.Florian Achleitner, Jul 26, 2012
  8. 07/16 Allow reading svn dumps from files via file:// urls.Florian Achleitner, Jul 26, 2012
  9. 08/16 vcs-svn: add fast_export_note to create notesFlorian Achleitner, Jul 26, 2012
  10. 09/16 Create a note for every imported commit containing svn metadata.Florian Achleitner, Jul 26, 2012
  11. 10/16 When debug==1, start fast-import with "--stats" instead of "--quiet".Florian Achleitner, Jul 26, 2012
  12. 11/16 Add explanatory comment for transport-helpers refs mapping.Florian Achleitner, Jul 26, 2012
  13. 12/16 remote-svn: add incremental import.Florian Achleitner, Jul 26, 2012
  14. 13/16 Add a svnrdump-simulator replaying a dump file for testing.Florian Achleitner, Jul 26, 2012
  15. 14/16 transport-helper: add import|export-marks to fast-import command line.Florian Achleitner, Jul 26, 2012
  16. 15/16 remote-svn: add marks-file regeneration.Florian Achleitner, Jul 26, 2012
  17. 16/16 Add a test script for remote-svn.Florian Achleitner, Jul 26, 2012
  18. Florian AchleitnerJul 26, 2012
  19. Jonathan NiederJul 28, 2012
  20. Jonathan NiederJul 26, 2012
  21. Florian AchleitnerJul 26, 2012
  22. Jonathan NiederJul 26, 2012
  23. Florian AchleitnerJul 26, 2012
  24. 00/16 GSOC remote-svn, rewritten patch seriesFlorian Achleitner, Jul 30, 2012
  25. 01/16 Implement a remote helper for svn in C.Florian Achleitner, Jul 30, 2012
  26. 02/16 Integrate remote-svn into svn-fe/Makefile.Florian Achleitner, Jul 30, 2012
  27. 03/16 Add svndump_init_fd to allow reading dumps from arbitrary FDs.Florian Achleitner, Jul 30, 2012
  28. 04/16 Add cat-blob report fifo from fast-import to remote-helper.Florian Achleitner, Jul 30, 2012
  29. 05/16 remote-svn, vcs-svn: Enable fetching to private refs.Florian Achleitner, Jul 30, 2012
  30. 06/16 Add a symlink 'git-remote-svn' in base dir.Florian Achleitner, Jul 30, 2012
  31. 07/16 Allow reading svn dumps from files via file:// urls.Florian Achleitner, Jul 30, 2012
  32. 08/16 vcs-svn: add fast_export_note to create notesFlorian Achleitner, Jul 30, 2012
  33. 09/16 Create a note for every imported commit containing svn metadata.Florian Achleitner, Jul 30, 2012
  34. 10/16 When debug==1, start fast-import with "--stats" instead of "--quiet".Florian Achleitner, Jul 30, 2012
  35. 11/16 Add explanatory comment for transport-helpers refs mapping.Florian Achleitner, Jul 30, 2012
  36. 12/16 remote-svn: add incremental import.Florian Achleitner, Jul 30, 2012
  37. 13/16 Add a svnrdump-simulator replaying a dump file for testing.Florian Achleitner, Jul 30, 2012
  38. 14/16 transport-helper: add import|export-marks to fast-import command line.Florian Achleitner, Jul 30, 2012
  39. 15/16 remote-svn: add marks-file regeneration.Florian Achleitner, Jul 30, 2012
  40. 16/16 Add a test script for remote-svn.Florian Achleitner, Jul 30, 2012
  41. Jonathan NiederJul 30, 2012
  42. Junio C HamanoJul 30, 2012
  43. Junio C HamanoJul 30, 2012
  44. Jonathan NiederJul 30, 2012
  45. Florian AchleitnerJul 30, 2012
  46. Junio C HamanoJul 30, 2012
  47. Florian AchleitnerJul 31, 2012

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.