Fix mtime change <1s bug.

5 files changed, 92 insertions, 71 deletions

diff --git a/bm/BM_HISTORY_1000_100.txt b/bm/BM_HISTORY_1000_100.txt
index a31c179..2885405 100644
--- a/bm/BM_HISTORY_1000_100.txt
+++ b/bm/BM_HISTORY_1000_100.txt
@@ -6,55 +6,55 @@ SNAPSHOT: Commit #20
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.43s |                0.08s
+Time            |                0.35s |                0.09s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1299 |                 2113
-Repo size       |             19384 KB |             22056 KB
+Inodes          |                 1301 |                 2121
+Repo size       |             18868 KB |             22076 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #40
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.42s |                0.11s
+Time            |                0.46s |                0.11s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1339 |                 2908
-Repo size       |             19544 KB |             29020 KB
+Inodes          |                 1341 |                 2929
+Repo size       |             19028 KB |             29136 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #60
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.42s |                0.10s
+Time            |                0.43s |                0.12s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1379 |                 3711
-Repo size       |             19704 KB |             35920 KB
+Inodes          |                 1381 |                 3732
+Repo size       |             19188 KB |             36088 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #80
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.41s |                0.11s
+Time            |                0.45s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1419 |                 4520
-Repo size       |             19864 KB |             43176 KB
+Inodes          |                 1421 |                 4538
+Repo size       |             19348 KB |             43104 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #100
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.42s |                0.11s
+Time            |                0.44s |                0.10s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1459 |                 5324
-Repo size       |             20024 KB |             50232 KB
+Inodes          |                 1462 |                 5346
+Repo size       |             19512 KB |             49980 KB
 -------------------------------------------------------------
 
 TOTAL URN REBASES: 0
diff --git a/bm/BM_REBASE_1000_100.txt b/bm/BM_REBASE_1000_100.txt
index 7875767..7550570 100644
--- a/bm/BM_REBASE_1000_100.txt
+++ b/bm/BM_REBASE_1000_100.txt
@@ -1,62 +1,61 @@
 =============================================================
  REBASE BENCHMARK: 1000 files (100 commits)
  CONDITIONS: Depth=2, Files Mod=5%, Change=50%
- INITIAL RAW DATA SIZE: 16976 KB
 =============================================================
 
 SNAPSHOT: Commit #20
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.29s |                0.05s
+Time            |                0.65s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1578 |                 2334
-Repo size       |             20404 KB |             19380 KB
+Inodes          |                 2272 |                 2282
+Repo size       |             38504 KB |             22700 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #40
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.54s |                0.05s
+Time            |                0.59s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1607 |                 3374
-Repo size       |             20520 KB |             23788 KB
+Inodes          |                 3332 |                 3322
+Repo size       |             59384 KB |             31188 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #60
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.31s |                0.05s
+Time            |                0.57s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1635 |                 4414
-Repo size       |             20632 KB |             28196 KB
+Inodes          |                 4392 |                 4362
+Repo size       |             80264 KB |             39676 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #80
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.29s |                0.05s
+Time            |                0.57s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1664 |                 5454
-Repo size       |             20748 KB |             32596 KB
+Inodes          |                 5452 |                 5402
+Repo size       |            101144 KB |             48156 KB
 -------------------------------------------------------------
 
 SNAPSHOT: Commit #100
 -------------------------------------------------------------
 METRIC          | URN                  | GIT                 
 ----------------+----------------------+---------------------
-Time            |                0.54s |                0.10s
+Time            |                0.57s |                0.08s
 Max RSS         |              0.02 MB |              0.01 MB
 Page faults     |        Maj:0 / Min:0 |        Maj:0 / Min:0
-Inodes          |                 1693 |                 6495
-Repo size       |             20864 KB |             37008 KB
+Inodes          |                 6512 |                 6442
+Repo size       |            122024 KB |             56644 KB
 -------------------------------------------------------------
 
-TOTAL URN REBASES: 273
+TOTAL URN REBASES: 4950
diff --git a/bm/bm_rebase.pl b/bm/bm_rebase.pl
index 3cb0fd5..e89e7f5 100644
--- a/bm/bm_rebase.pl
+++ b/bm/bm_rebase.pl
@@ -1,10 +1,13 @@
 #!/usr/bin/perl
 use strict;
 use warnings;
+use File::Copy;
 use File::Spec;
-use File::Path qw(remove_tree);
+use File::Path qw(remove_tree make_path);
 use File::Find;
+use File::Basename;
 use Cwd qw(getcwd abs_path);
+use Time::HiRes qw(time);
 
 my ($files, $depth, $total_commits, $file_perc, $change_perc) = @ARGV;
 if (!defined $files || !defined $depth || !defined $total_commits || !defined $file_perc || !defined $change_perc) {
@@ -14,15 +17,16 @@ if (!defined $files || !defined $depth || !defined $total_commits || !defined $f
 my $base_dir = getcwd();
 my $urn_bin  = abs_path(File::Spec->catfile("..", "urn")); 
 my $seed_bin = abs_path("seed.pl");
-my $bm_repo  = "bm_repo";
+my $bm_repo  = "sandbox";
 
 my $sample_rate = int($total_commits / 5) || 1;
 my %results;
 my $initial_repo_size = "0 KB";
-
 my $rebase_count = 0;
 my %last_base_hashes;
 
+my $global_tick = 0;
+
 sub get_size {
 	my $dir = shift;
 	return "0 KB" unless -d $dir;
@@ -62,10 +66,7 @@ sub run_rebase_benchmark {
 	remove_tree($bm_repo) if -d $bm_repo;
 	system("perl $seed_bin $files $depth > /dev/null 2>&1");
 	
-	# Capture raw directory size before VCS initialization
-	if ($initial_repo_size eq "0 KB") {
-		$initial_repo_size = get_size($bm_repo);
-	}
+	if ($initial_repo_size eq "0 KB") { $initial_repo_size = get_size($bm_repo); }
 	
 	chdir($bm_repo) or die $!;
 
@@ -84,19 +85,37 @@ sub run_rebase_benchmark {
 	my $num_to_change = int(scalar(@file_list) * ($file_perc / 100)) || 1;
 	my @target_files = @file_list[0 .. $num_to_change - 1];
 
-	for my $i (1 .. $total_commits) {
-		for my $target (@target_files) {
-			open(my $fh, '<', $target) or next;
-			my @lines = <$fh>;
-			close($fh);
+	for my $i (2 .. $total_commits) {
+		my $debug_dir = "/tmp/urn/commit_$i";
+		make_path($debug_dir) unless -d $debug_dir;
 
-			my $lines_to_change = int(scalar(@lines) * ($change_perc / 100)) || 1;
-			splice(@lines, 0, $lines_to_change, "Commit $i: block change\n");
-
-			if (open(my $out, '>', $target)) {
-				print $out @lines;
-				close($out);
-			}
+		for my $target (@target_files) {
+			open(my $fh, '<:raw', $target) or die "Read fail: $target - $!";
+    		my $content = do { local $/; <$fh> };
+    		close($fh);
+
+    		$global_tick++;
+    		my $ts = time();
+    		my $header = "C$i-T$ts-N$global_tick ";
+    		
+    		my $total_bytes = length($content);
+    		my $min_len = length($header);
+    		my $to_change = int($total_bytes * ($change_perc / 100));
+
+    		$to_change = $min_len if $to_change < $min_len;
+    		$to_change = $total_bytes if $to_change > $total_bytes;
+
+    		my $new_segment = substr($header . ("." x $to_change), 0, $to_change);
+    		substr($content, 0, $to_change, $new_segment);
+
+    		open(my $out, '>:raw', $target) or die "Write fail: $target - $!";
+    		print $out $content;
+    		close($out);
+
+			my $target_name = basename($target);
+			copy($target, "$debug_dir/$target_name") or warn "Backup failed: $!";
+    		
+    		utime(undef, undef, $target);
 		}
 
 		my $cmd = ($tool_name eq "URN") 
@@ -104,14 +123,17 @@ sub run_rebase_benchmark {
 			: "git add . && git commit -m 'c$i'";
 
 		if ($i % $sample_rate == 0 || $i == $total_commits) {
-			my $stats = `/usr/bin/time -l sh -c "$cmd" 2>&1 > /dev/null`;
-			my ($real, $rss, $maj, $min) = (0, 0, 0, 0);
-			if ($stats =~ /(\d+\.\d+)\s+real/) { $real = $1; }
-			if ($stats =~ /(\d+)\s+maximum resident set size/) { $rss = sprintf("%.2f MB", $1 / 1024 / 1024); }
-			if ($stats =~ /(\d+)\s+page reclaims/) { $min = $1; }
-			if ($stats =~ /(\d+)\s+page faults/) { $maj = $1; }
+			my $raw_output = `/usr/bin/time -l sh -c "$cmd" 2>&1`;
+
+			if ($tool_name eq "URN") {
+				track_rebases();
+			}
 
-			if ($tool_name eq "URN") { track_rebases(); }
+			my ($real, $rss, $maj, $min) = (0, 0, 0, 0);
+			$real = $1 if $raw_output =~ /(\d+\.\d+)\s+real/;
+			$rss = sprintf("%.2f MB", $1 / 1024 / 1024) if $raw_output =~ /(\d+)\s+maximum resident set size/;
+			$min = $1 if $raw_output =~ /(\d+)\s+page reclaims/;
+			$maj = $1 if $raw_output =~ /(\d+)\s+page faults/;
 
 			my $meta = ($tool_name eq "URN") ? ".urn" : ".git";
 			$results{$i}{$tool_name} = {
@@ -124,11 +146,10 @@ sub run_rebase_benchmark {
 			print "   [Commit $i] $tool_name sampled.\n";
 		} else {
 			system("$cmd > /dev/null 2>&1");
-			if ($tool_name eq "URN") { track_rebases(); }
+			track_rebases() if $tool_name eq "URN";
 		}
 	}
 	chdir($base_dir);
-	remove_tree($bm_repo);
 }
 
 $rebase_count = 0;
@@ -141,7 +162,6 @@ open(my $res, '>', $out_file) or die $!;
 print $res "=============================================================\n";
 print $res " REBASE BENCHMARK: $files files ($total_commits commits)\n";
 print $res " CONDITIONS: Depth=$depth, Files Mod=$file_perc%, Change=$change_perc%\n";
-print $res " INITIAL RAW DATA SIZE: $initial_repo_size\n";
 print $res "=============================================================\n\n";
 
 foreach my $i (sort { $a <=> $b } keys %results) {
@@ -158,10 +178,7 @@ foreach my $i (sort { $a <=> $b } keys %results) {
 	printf $res "%-15s | %20s | %20s\n", "Repo size", $u->{size}, $g->{size};
 	print $res "-------------------------------------------------------------\n\n";
 }
-
 print $res "TOTAL URN REBASES: $rebase_count\n";
 close($res);
 
-my $pager = $ENV{PAGER} || 'less';
-exec $pager, $out_file;
-
+exec "less $out_file";
diff --git a/bm/seed.pl b/bm/seed.pl
index 7649b64..81c70dd 100644
--- a/bm/seed.pl
+++ b/bm/seed.pl
@@ -10,7 +10,7 @@ if (!defined $total_files || !defined $max_depth) {
 	die "Usage: perl seed.pl <total_files> <max_depth>\nExample: perl seed.pl 200 20\n";
 }
 
-my $target_root = "bm_repo";
+my $target_root = "sandbox";
 my $files_created = 0;
 
 # Helper to generate random "code-like" text
diff --git a/urn b/urn
index 069b695..c8ddbce 100644
--- a/urn
+++ b/urn
@@ -178,19 +178,24 @@ sub run_add {
 				: $idx_entry->{path} cmp $wrk_entry->{path};
 
 		if ($cmp == 0) {
+			my $idx_mtime = (stat(INDEX))[9]; # Needed if a mtime change <1s 
 			if ($idx_entry->{mtime} == $wrk_entry->{mtime} && 
-				$idx_entry->{size}  == $wrk_entry->{size}) {
+				$idx_entry->{size}  == $wrk_entry->{size} &&
+				$idx_entry->{mtime} != $idx_mtime) { 
 				# No change: Preserve all 3 hashes and metadata
 				printf $out "%-40s\t%-40s\t%-40s\t%-12d\t%-10d\t%s\n", 
 					$idx_entry->{s_hash}, $idx_entry->{c_hash}, $idx_entry->{b_hash},
 					$idx_entry->{mtime}, $idx_entry->{size}, $idx_entry->{path};
 			} else {
 				my $p = $wrk_entry->{path};
+				my $last_hash = $idx_entry->{c_hash};
 				my $current_hash = hash_file_content($p);
-				my $stg_path = File::Spec->catfile(TMP_DIR, $p);
-				make_path(dirname($stg_path));
-				
-				(-l $p) ? symlink(readlink($p), $stg_path) : copy($p, $stg_path);
+
+				if ($last_hash ne $current_hash) {
+					my $stg_path = File::Spec->catfile(TMP_DIR, $p);
+					make_path(dirname($stg_path));
+					(-l $p) ? symlink(readlink($p), $stg_path) : copy($p, $stg_path);
+				}
 
 				# Update staged hash, preserve committed and base hashes
 				printf $out "%-40s\t%-40s\t%-40s\t%-12d\t%-10d\t%s\n",