2018-06-16 18:39:54 +02:00
|
|
|
#!/usr/bin/perl
|
|
|
|
|
2019-03-01 07:06:27 +01:00
|
|
|
use IPC::Cmd qw[ run ];
|
|
|
|
|
2018-06-16 18:39:54 +02:00
|
|
|
(@ARGV == 1)
|
|
|
|
or die "ERROR: no map file specified.\n";
|
|
|
|
open(my $file, $ARGV[0])
|
|
|
|
or die "ERROR: could not open file '$ARGV[0]'.\n";
|
|
|
|
|
|
|
|
my $src = 0;
|
|
|
|
my $asm = 0;
|
|
|
|
while (my $line = <$file>)
|
|
|
|
{
|
|
|
|
if ($line =~ /^ \.(\w+)\s+0x[0-9a-f]+\s+(0x[0-9a-f]+) (\w+)\/.+\.o/)
|
|
|
|
{
|
|
|
|
my $section = $1;
|
|
|
|
my $size = hex($2);
|
|
|
|
my $dir = $3;
|
|
|
|
|
|
|
|
if ($section =~ /text/)
|
|
|
|
{
|
|
|
|
if ($dir eq 'src')
|
|
|
|
{
|
|
|
|
$src += $size;
|
|
|
|
}
|
|
|
|
elsif ($dir eq 'asm')
|
|
|
|
{
|
|
|
|
$asm += $size;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-03-03 21:02:26 +01:00
|
|
|
# Note that the grep filters out all branch labels. It also requires a minimum
|
|
|
|
# line length of 5, to filter out a ton of generated symbols (like AcCn). No
|
|
|
|
# settings to nm seem to remove these symbols. Finally, nm prints out a separate
|
|
|
|
# entry for whenever a name appears in a file, not just where it's defined. uniq
|
|
|
|
# removes all the duplicate entries.
|
|
|
|
#
|
|
|
|
#
|
|
|
|
# You'd expect this to take a while, because of uniq. It runs in under a second,
|
|
|
|
# though. Uniq is pretty fast!
|
|
|
|
my $base_cmd = "nm pokeemerald.elf | awk '{print \$3}' | grep '^[^_].\\{4\\}' | uniq";
|
|
|
|
|
|
|
|
# This looks for Unknown_, Unknown_, or sub_, followed by just numbers. Note that
|
|
|
|
# it matches even if stuff precedes the unknown, like sUnknown/gUnknown.
|
|
|
|
my $undoc_cmd = "grep '[Uu]nknown_[0-9a-fA-F]*\\|sub_[0-9a-fA-F]*'";
|
|
|
|
|
|
|
|
# This looks for every symbol with an address at the end of it. Some things are
|
|
|
|
# given a name based on their type / location, but still have an unknown purpose.
|
|
|
|
# For example, FooMap_EventScript_FFFFFFF.
|
|
|
|
my $partial_doc_cmd = "grep '[0-9a-fA-F]\\{6,7\\}'";
|
|
|
|
|
|
|
|
my $count_cmd = "wc -l";
|
|
|
|
|
|
|
|
# It sucks that we have to run this three times, but I can't figure out how to get
|
|
|
|
# stdin working for subcommands in perl while still having a timeout. It's decently
|
|
|
|
# fast anyway.
|
2019-03-01 07:06:27 +01:00
|
|
|
my $total_syms_as_string;
|
|
|
|
(run (
|
2019-03-03 21:02:26 +01:00
|
|
|
command => "$base_cmd | $count_cmd",
|
2019-03-01 07:09:53 +01:00
|
|
|
buffer => \$total_syms_as_string,
|
2019-03-03 21:02:26 +01:00
|
|
|
timeout => 60
|
2019-03-01 07:06:27 +01:00
|
|
|
))
|
|
|
|
or die "ERROR: Error while getting all symbols: $?";
|
|
|
|
|
|
|
|
my $undocumented_as_string;
|
|
|
|
(run (
|
2019-03-03 21:02:26 +01:00
|
|
|
command => "$base_cmd | $undoc_cmd | $count_cmd",
|
2019-03-01 07:09:53 +01:00
|
|
|
buffer => \$undocumented_as_string,
|
2019-03-03 21:02:26 +01:00
|
|
|
timeout => 60
|
2019-03-01 07:06:27 +01:00
|
|
|
))
|
|
|
|
or die "ERROR: Error while filtering for undocumented symbols: $?";
|
|
|
|
|
2019-03-03 21:02:26 +01:00
|
|
|
my $partial_documented_as_string;
|
|
|
|
(run (
|
|
|
|
command => "$base_cmd | $partial_doc_cmd | $count_cmd",
|
|
|
|
buffer => \$partial_documented_as_string,
|
|
|
|
timeout => 60
|
|
|
|
))
|
|
|
|
or die "ERROR: Error while filtering for partial symbols: $?";
|
|
|
|
|
|
|
|
# Performing addition on a string converts it to a number. Any string that fails
|
|
|
|
# to convert to a number becomes 0. So if our converted number is 0, but our string
|
|
|
|
# is nonzero, then the conversion was an error.
|
2019-03-01 07:06:27 +01:00
|
|
|
my $undocumented = $undocumented_as_string + 0;
|
|
|
|
(($undocumented != 0) and ($undocumented_as_string ne "0"))
|
|
|
|
or die "ERROR: Cannot convert string to num: '$undocumented_as_string'";
|
|
|
|
|
2019-03-03 21:02:26 +01:00
|
|
|
my $partial_documented = $partial_documented_as_string + 0;
|
|
|
|
(($partial_documented != 0) and ($partial_documented_as_string ne "0"))
|
|
|
|
or die "ERROR: Cannot convert string to num: '$partial_documented_as_string'";
|
|
|
|
|
2019-03-01 07:06:27 +01:00
|
|
|
my $total_syms = $total_syms_as_string + 0;
|
|
|
|
(($total_syms != 0) and ($total_syms_as_string ne "0"))
|
|
|
|
or die "ERROR: Cannot convert string to num: '$total_syms_as_string'";
|
|
|
|
|
|
|
|
($total_syms != 0)
|
|
|
|
or die "ERROR: No symbols found.";
|
|
|
|
|
2018-06-16 18:39:54 +02:00
|
|
|
my $total = $src + $asm;
|
|
|
|
my $srcPct = sprintf("%.4f", 100 * $src / $total);
|
|
|
|
my $asmPct = sprintf("%.4f", 100 * $asm / $total);
|
2019-03-01 07:06:27 +01:00
|
|
|
|
2019-03-03 21:02:26 +01:00
|
|
|
# partial_documented is double-counting the unknown_* and sub_* symbols.
|
|
|
|
$partial_documented = $partial_documented - $undocumented;
|
|
|
|
|
|
|
|
my $documented = $total_syms - ($undocumented + $partial_documented);
|
2019-03-01 07:06:27 +01:00
|
|
|
my $docPct = sprintf("%.4f", 100 * $documented / $total_syms);
|
2019-03-03 21:02:26 +01:00
|
|
|
my $partialPct = sprintf("%.4f", 100 * $partial_documented / $total_syms);
|
2019-03-01 07:06:27 +01:00
|
|
|
my $undocPct = sprintf("%.4f", 100 * $undocumented / $total_syms);
|
|
|
|
|
2018-06-16 18:39:54 +02:00
|
|
|
print "$total total bytes of code\n";
|
|
|
|
print "$src bytes of code in src ($srcPct%)\n";
|
|
|
|
print "$asm bytes of code in asm ($asmPct%)\n";
|
2019-03-01 07:06:27 +01:00
|
|
|
print "\n";
|
|
|
|
print "$total_syms total symbols\n";
|
|
|
|
print "$documented symbols documented ($docPct%)\n";
|
2019-03-03 21:02:26 +01:00
|
|
|
print "$partial_documented symbols partially documented ($partialPct%)\n";
|
2019-03-01 07:06:27 +01:00
|
|
|
print "$undocumented symbols undocumented ($undocPct%)\n";
|