v5-0006-Refactor-split-verify_backup_file-function.patch

From 2206efa24af8882877ec216e0001dee8538e30c1 Mon Sep 17 00:00:00 2001
From: Amul Sul <amul.sul@enterprisedb.com>
Date: Thu, 1 Aug 2024 12:15:26 +0530
Subject: [PATCH v5 06/12] Refactor: split verify_backup_file() function.

Move the manifest entry verification code into a new function as
verify_manifest_entry(). And the total size computation code into
another new function, compute_total_size(), which is called from the
main.
---
 src/bin/pg_verifybackup/pg_verifybackup.c | 76 ++++++++++++++++++-----
 src/bin/pg_verifybackup/pg_verifybackup.h |  3 +
 2 files changed, 62 insertions(+), 17 deletions(-)

diff --git a/src/bin/pg_verifybackup/pg_verifybackup.c b/src/bin/pg_verifybackup/pg_verifybackup.c
index 4e42757c346..ab6bda8c9dc 100644
--- a/src/bin/pg_verifybackup/pg_verifybackup.c
+++ b/src/bin/pg_verifybackup/pg_verifybackup.c
@@ -72,6 +72,7 @@ static void parse_required_wal(verifier_context *context,
 							   char *pg_waldump_path,
 							   char *wal_directory);
 
+static void compute_total_size(verifier_context *context);
 static void usage(void);
 
 static const char *progname;
@@ -250,6 +251,13 @@ main(int argc, char **argv)
 	 */
 	context.manifest = parse_manifest_file(manifest_path);
 
+	/*
+	 * For the progress report, compute the total size of the files to be
+	 * read, which occurs only when checksum verification is enabled.
+	 */
+	if (!context.skip_checksums)
+		compute_total_size(&context);
+
 	/*
 	 * Now scan the files in the backup directory. At this stage, we verify
 	 * that every file on disk is present in the manifest and that the sizes
@@ -614,6 +622,27 @@ verify_backup_file(verifier_context *context, char *relpath, char *fullpath)
 		return;
 	}
 
+	/* Check the backup manifest entry for this file. */
+	m = verify_manifest_entry(context, relpath, sb.st_size);
+
+	/*
+	 * Validate the manifest system identifier, not available in manifest
+	 * version 1.
+	 */
+	if (context->manifest->version != 1 &&
+		strcmp(relpath, "global/pg_control") == 0 &&
+		m != NULL && m->matched && !m->bad)
+		verify_control_file(fullpath, context->manifest->system_identifier);
+}
+
+/*
+ * Verify file and its size entry in the manifest.
+ */
+manifest_file *
+verify_manifest_entry(verifier_context *context, char *relpath, int64 filesize)
+{
+	manifest_file *m;
+
 	/* Check whether there's an entry in the manifest hash. */
 	m = manifest_files_lookup(context->manifest->files, relpath);
 	if (m == NULL)
@@ -621,40 +650,29 @@ verify_backup_file(verifier_context *context, char *relpath, char *fullpath)
 		report_backup_error(context,
 							"\"%s\" is present on disk but not in the manifest",
 							relpath);
-		return;
+		return NULL;
 	}
 
 	/* Flag this entry as having been encountered in the filesystem. */
 	m->matched = true;
 
 	/* Check that the size matches. */
-	if (m->size != sb.st_size)
+	if (m->size != filesize)
 	{
 		report_backup_error(context,
 							"\"%s\" has size %lld on disk but size %zu in the manifest",
-							relpath, (long long int) sb.st_size, m->size);
+							relpath, (long long int) filesize, m->size);
 		m->bad = true;
 	}
 
-	/*
-	 * Validate the manifest system identifier, not available in manifest
-	 * version 1.
-	 */
-	if (context->manifest->version != 1 &&
-		strcmp(relpath, "global/pg_control") == 0)
-		verify_control_file(fullpath, context->manifest->system_identifier);
-
-	/* Update statistics for progress report, if necessary */
-	if (context->show_progress && !context->skip_checksums &&
-		should_verify_checksum(m))
-		context->total_size += m->size;
-
 	/*
 	 * We don't verify checksums at this stage. We first finish verifying that
 	 * we have the expected set of files with the expected sizes, and only
 	 * afterwards verify the checksums. That's because computing checksums may
 	 * take a while, and we'd like to report more obvious problems quickly.
 	 */
+
+	return m;
 }
 
 /*
@@ -817,7 +835,7 @@ verify_file_checksum(verifier_context *context, manifest_file *m,
 
 	/*
 	 * Double-check that we read the expected number of bytes from the file.
-	 * Normally, a file size mismatch would be caught in verify_backup_file
+	 * Normally, a file size mismatch would be caught in verify_manifest_entry
 	 * and this check would never be reached, but this provides additional
 	 * safety and clarity in the event of concurrent modifications or
 	 * filesystem misbehavior.
@@ -988,6 +1006,30 @@ progress_report(verifier_context *context, bool finished)
 	fputc((!finished && isatty(fileno(stderr))) ? '\r' : '\n', stderr);
 }
 
+/*
+ * Compute the total size of backup files for progress reporting.
+ */
+static void
+compute_total_size(verifier_context *context)
+{
+	manifest_data *manifest = context->manifest;
+	manifest_files_iterator it;
+	manifest_file *m;
+	uint64		total_size = 0;
+
+	if (!context->show_progress)
+		return;
+
+	manifest_files_start_iterate(manifest->files, &it);
+	while ((m = manifest_files_iterate(manifest->files, &it)) != NULL)
+	{
+		if (!should_ignore_relpath(context, m->pathname))
+			total_size += m->size;
+	}
+
+	context->total_size = total_size;
+}
+
 /*
  * Print out usage information and exit.
  */
diff --git a/src/bin/pg_verifybackup/pg_verifybackup.h b/src/bin/pg_verifybackup/pg_verifybackup.h
index 90900048547..98c75916255 100644
--- a/src/bin/pg_verifybackup/pg_verifybackup.h
+++ b/src/bin/pg_verifybackup/pg_verifybackup.h
@@ -105,6 +105,9 @@ typedef struct verifier_context
 	uint64		done_size;
 } verifier_context;
 
+extern manifest_file *verify_manifest_entry(verifier_context *context,
+											char *relpath, int64 filesize);
+
 extern void report_backup_error(verifier_context *context,
 								const char *pg_restrict fmt,...)
 			pg_attribute_printf(2, 3);
-- 
2.18.0