wal_utils: Add function to fetch size of an archive element

This is useful if combined with the function to get a full data chunk
from the archives. At the same time regression tests are refactored
so as the extension wal_utils is created only once as initialization
for all the other tests.

wal_utils/Makefile

@@ -3,7 +3,7 @@ MODULES = wal_utils
 EXTENSION = wal_utils
 DATA = wal_utils--1.0.sql
 PGFILEDESC = "wal_utils - Set of tools for WAL data"
-REGRESS = parse_wal_history wal_segment_list
+REGRESS = init archive_data parse_wal_history wal_segment_list
 PG_CONFIG = pg_config
 PGXS := $(shell $(PG_CONFIG) --pgxs)

wal_utils/expected/archive_data.out

+-- Sanity check for archive_get_size
+SELECT archive_get_size('../no_parent'); -- error
+ERROR:  reference to parent directory ("..") not allowed
+SELECT archive_get_size('/no_absolute'); -- error
+ERROR:  absolute path not allowed

wal_utils/expected/init.out

+-- Create extension for all the following tests
+CREATE EXTENSION wal_utils;

wal_utils/expected/parse_wal_history.out

--- Install extension for tests
-CREATE EXTENSION wal_utils;
+-- Tests for parsing of WAL history file
 CREATE TABLE history_data (data text);
 -- Note that COPY truncates newlines.
 -- Load history file that succeeds

wal_utils/expected/wal_segment_list.out

--- Install extension for tests
-CREATE EXTENSION wal_utils;
-ERROR:  extension "wal_utils" already exists
+-- Tests for build of WAL segment list for multiple timelines
 -- NULL checks
 SELECT build_wal_segment_list(NULL, '0/0'::pg_lsn, 1, '0/0'::pg_lsn, NULL);
 ERROR:  origin or target data cannot be NULL

wal_utils/sql/archive_data.sql

+-- Sanity check for archive_get_size
+SELECT archive_get_size('../no_parent'); -- error
+SELECT archive_get_size('/no_absolute'); -- error

wal_utils/sql/init.sql

+-- Create extension for all the following tests
+CREATE EXTENSION wal_utils;

wal_utils/sql/parse_wal_history.sql

--- Install extension for tests
-CREATE EXTENSION wal_utils;
+-- Tests for parsing of WAL history file
 CREATE TABLE history_data (data text);

wal_utils/sql/wal_segment_list.sql

--- Install extension for tests
-CREATE EXTENSION wal_utils;
+-- Tests for build of WAL segment list for multiple timelines
 -- NULL checks
 SELECT build_wal_segment_list(NULL, '0/0'::pg_lsn, 1, '0/0'::pg_lsn, NULL);

wal_utils/wal_utils--1.0.sql

@@ -28,3 +28,27 @@ CREATE FUNCTION build_wal_segment_list(
 RETURNS SETOF text
 AS 'MODULE_PATHNAME'
 LANGUAGE C;
+
+-- Set of routines for archive data fetching
+-- Get data from the archive path. The base path where the lookup is
+-- done uses as environment variable PGARCHIVE which points to a local
+-- path where the archives are located. This should be a variable loaded
+-- by Postgres. Note that there is no restriction on the file name that
+-- caller can use here, a segment file could be compressed, and the
+-- archive could be used as well to store some custom metadata. The path
+-- defined cannot be absolute as well.
+-- CREATE FUNCTION archive_get_data(
+-- 	IN filename text,
+-- 	IN begin bigint,
+-- 	IN offset bigint,
+-- 	OUT data bytea)
+-- RETURNS bytea
+-- AS 'MODULE_PATHNAME'
+-- LANGUAGE C STRICT;
+-- Get the size of a file in archives.
+CREATE FUNCTION archive_get_size(
+	IN filename text,
+	OUT size bigint)
+RETURNS bigint
+AS 'MODULE_PATHNAME'
+LANGUAGE C STRICT;

wal_utils/wal_utils.c

@@ -14,6 +14,8 @@
 #include "postgres.h"
 #include "fmgr.h"
+#include <sys/stat.h>
+
 #include "access/timeline.h"
 #include "access/xlog_internal.h"
 #include "catalog/pg_type.h"
@@ -31,6 +33,7 @@ static List *parseTimeLineHistory(char *buffer);
  */
 PG_FUNCTION_INFO_V1(parse_wal_history);
 PG_FUNCTION_INFO_V1(build_wal_segment_list);
+PG_FUNCTION_INFO_V1(archive_get_size);
 /*
  * parseTimeLineHistory
@@ -433,3 +436,73 @@ build_wal_segment_list(PG_FUNCTION_ARGS)
 	return (Datum) 0;
 }
+
+/*
+ * Check the defined file name, looking at if it is an absolute path
+ * and if it contains references to a parent directory. Then build
+ * a full path name using the path defined for the archives which is
+ * enforced by the environment where Postgres is running.
+ */
+static char *
+check_and_build_filepath(char *filename)
+{
+	char	   *filepath;
+	char	   *archive_path;
+
+	if (is_absolute_path(filename))
+		ereport(ERROR,
+				(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+				 (errmsg("absolute path not allowed"))));
+	if (path_contains_parent_reference(filename))
+		ereport(ERROR,
+				(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+				 (errmsg("reference to parent directory (\"..\") not allowed"))));
+
+	archive_path = getenv("PGARCHIVE");
+	if (archive_path == NULL)
+		ereport(ERROR,
+				(errcode(ERRCODE_FDW_INVALID_OPTION_NAME),
+				 errmsg("archive path is not defined"),
+				 errhint("Check value of environment variable %s",
+						 "PGARCHIVE")));
+
+	filepath = (char *) palloc(MAXPGPATH);
+	snprintf(filepath, MAXPGPATH, "%s/%s", archive_path, filename);
+
+	/* length can change here */
+	canonicalize_path(filepath);
+
+	return filepath;
+}
+
+/*
+ * archive_get_size
+ *
+ * Look at a file in the archives whose path is defined by the environment
+ * variable PGARCHIVE and get its size. This is useful when combined with
+ * archive_get_data to evaluate a set of chunks to be used during any
+ * data transfer from the archives.
+ */
+Datum
+archive_get_size(PG_FUNCTION_ARGS)
+{
+	char	   *filename = text_to_cstring(PG_GETARG_TEXT_PP(0));
+	char	   *filepath;
+	struct stat fst;
+
+	if (!superuser())
+		ereport(ERROR,
+				(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
+				 (errmsg("must be superuser to read files"))));
+
+	filepath = check_and_build_filepath(filename);
+	pfree(filename);
+
+	/* get needed information about the file */
+	if (stat(filepath, &fst) < 0)
+		ereport(ERROR,
+				(errcode_for_file_access(),
+				 errmsg("could not stat file \"%s\": %m", filepath)));
+
+	PG_RETURN_INT64((int64) fst.st_size);
+}