My apologies. That part wasn’t really meant to be posted. I figgered you would remove it. I mainly added it because I just noticed that the last couple of posts had been hanging around for a couple of days, and wanted to find out if this thing was on.

It is. Thanks for unclogging the pipes.

I did send an apology to you privately, but I believe that apologies should be delivered in the same venue as the transgression, so I’ll repeat it here.

You guys might find this script useful. I read up on submodules, figgered out where I was wrong, and adjusted the script accordingly. I also figured out that I like the way Git handles submodules a bit better, after reading this article.

#!/usr/bin/perl

use strict;
use Cwd;

# You can specify a starting directory as the script argument.
if ( $_[0] )
{
chdir ( $_[0] );
}

my $global_indent = 0;
print ( ‘Searching the base project at “‘, cwd(), ‘”‘ );
init_and_update();
print ( “\n” );

exit;

# This function is a recursive function that will traverse a submodule hierarchy,
# and will update them, from the bottom up.
sub init_and_update
{
my @submodules;
# First, you must have submodules.
if ( open ( GITFILE, ‘.gitmodules’ ) )
{
my $heading;

# If so, we parse the .gitmodules file, and get the important parts.
# This is a REAL DUMB parser. It counts on the lines being in a particular order.
# This is Perl. We can do better.
while ( $heading = )
{
my $pathname;
my $url;

# The heading is the submodule header (and name).
chomp ( $heading );
$heading =~ s/^\s+//;
$heading =~ s/\s+$//;

# If we have a submodule…
if ( $heading =~ m/\[submodule / )
{
# Strip off the extra
$heading =~ s/\[submodule "(.*?)"\]/$1/;

# Get the pathname
$pathname = ;
chomp ( $pathname );
$pathname =~ s/^\s+//;
$pathname =~ s/\s+$//;
$pathname =~ s/path = (.*?)$/$1/;

# Get the URL of the origin
$url = ;
chomp ( $url );
$url =~ s/^\s+//;
$url =~ s/\s+$//;
$url =~ s/url = (.*?)$/$1/;

# Add it to our stack.
push @submodules, { ‘submodule’ => $heading, ‘pathname’ => $pathname, ‘url’ => $url } ;
}
}

close ( GITFILE );
}

# Make sure that we got some submodules.
if ( @submodules > 0 )
{
output_indents();
print ( “This directory has the following submodules:” );

$global_indent++;
for my $index ( 0 .. $#submodules )
{
output_indents();
print ( $index + 1, ‘) ‘, $submodules[$index] { ‘submodule’ } );
}
$global_indent–;

# Now, we simply go through the list, recursing all the way…
for my $index ( 0 .. $#submodules )
{
# Recursion
my $start_path = cwd();
chdir ( $submodules[$index] { ‘pathname’ } );
output_indents();
print ( “Looking for submodules under the “, $submodules[$index] { ‘submodule’ }, ” submodule” );
$global_indent++;
init_and_update();
$global_indent–;
chdir ( $start_path );
}

# Lets do our own.
`git submodule init 2>&1`;
`git submodule update 2>&1`;
`git submodule foreach ‘git checkout HEAD’ 2>&1`;
output_indents();
print ( “Updated the submodules in the \”", cwd(), “\” directory” );
}
else
{
output_indents();
print ( “No further submodules under this directory” );
}
}

# This simply helps the user to see the hierarchy of the operation.
sub output_indents
{
print ( “\n” );

for my $index ( 0 .. $global_indent )
{
print ( ” ” );
}
}

Quick typo correction on your new line:

`git submodule foreach ‘git checkout master’ 2>&1 Is in the script.

It should be:

`git submodule foreach ‘git checkout master’ 2>&1`;

I just added one line

[...]
`git submodule init 2>&1`;
`git submodule update 2>&1`;
`git submodule foreach ‘git checkout master’ 2>&1`;