Setting automatic GIT deployment of PHP project

Question

What I want to do is, to switch from FTP deployment into GIT. I mean, I want to keep automatically keep synced my Bitbucket private repo and my shared webhosting. I googled and found following script to deploy my webserver (based on this article).

// Set these dependant on your BB credentials    
$username = 'username';
$password = 'password';

// Grab the data from BB's POST service and decode
$json = stripslashes($_POST['payload']);
$data = json_decode($json);

// Set some parameters to fetch the correct files
$uri   = $data->repository->absolute_url;
$node  = $data->commits[0]->node;
$files = $data->commits[0]->files;

// Foreach through the files and curl them over    
foreach ($files as $file) {
    if ($file->type == "removed") {
        unlink($file->file);
    } else {
        $url  = "https://api.bitbucket.org/1.0/repositories"
            . $uri . "raw/" .$node ."/" . $file->file;
        $path = $file->file;

        $dirname = dirname($path);
        if (!is_dir($dirname)) {
            mkdir($dirname, 0775, true);
        }
        
        $fp = fopen($path, 'w');
        
        $ch = curl_init($url);
        curl_setopt($ch, CURLOPT_USERPWD, "$username:$password");
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true); 
        curl_setopt($ch, CURLOPT_FILE, $fp);
        
        $data = curl_exec($ch);
        
        curl_close($ch);
        fclose($fp);    
    }   
}

The problem is, this works on simple changesets like 5-10 file change. But when I push the whole project for the first time (for example with 600-700 files and folders) into my bitbucket private profile, this script doesn't work. (just doesn't, no error on errors.log)

What am I missing?

By the way, Can I do something like that:

As we know, Bitbucket can send POST information into an exact url (given by user) directly after a commit has been made. So when deploy.php receives POST, we can get the entire commit as a zip or tar, clean our current files and unzip the new commit into webserver.

Is that possible? If yes then how? Any other good way?

Update

I found the code below for automated deploying php project. The problem is https://bitbucket.org/$username/$reponame/get/tip.zip this url doesnt work on bitbucket private git repo: probably related with authentication (I haven't tested this on public repo) What i need is to get the last commit's zip file and unzip inside my project.

<?

// your Bitbucket username
$username   = "edifreak";

// your Bitbucket repo name
$reponame   = "canvas-game-demo";

// extract to
$dest       = "./"; // leave ./ for relative destination

////////////////////////////////////////////////////////
// Let's get stuff done!

// set higher script timeout (for large repo's or slow servers)
set_time_limit(380);

// download the repo zip file
$repofile = file_get_contents("https://bitbucket.org/$username/$reponame/get/tip.zip");
file_put_contents('tip.zip', $repofile);
unset($repofile);

// unzip
$zip = new ZipArchive;
$res = $zip->open('tip.zip');
if ($res === TRUE) {
    $zip->extractTo('./');
    $zip->close();
} else {
    die('ZIP not supported on this server!');
}

// delete unnecessary .hg files
@unlink("$username-$reponame-tip/.hgignore");
@unlink("$username-$reponame-tip/.hg_archival.txt");

// function to delete all files in a directory recursively
function rmdir_recursively($dir) { 
    if (is_dir($dir)) { 
        $objects = scandir($dir); 
        foreach ($objects as $object) { 
            if ($object != "." && $object != "..") { 
                if (filetype($dir."/".$object) == "dir") rmdir_recursively($dir."/".$object); else unlink($dir."/".$object); 
            } 
        } 
        reset($objects); 
        rmdir($dir); 
    } 
} 

// function to recursively copy the files
function copy_recursively($src, $dest) {
    if (is_dir($src)) {
        if($dest != "./") rmdir_recursively($dest);
        @mkdir($dest);
        $files = scandir($src);
        foreach ($files as $file)
            if ($file != "." && $file != "..") copy_recursively("$src/$file", "$dest/$file"); 
        }
    else if (file_exists($src)) copy($src, $dest);
    rmdir_recursively($src);
}

// start copying the files from extracted repo and delete the old directory recursively
copy_recursively("$username-$reponame-tip", $dest);

// delete the repo zip file
unlink("tip.zip");

// Yep, we're done :)
echo "We're done!";
    
?>

What's in your error log when you run this script? Perhaps it's taking too long or using too much RAM and the httpd is killing it? — cdhowie
– cdhowie, Commented Oct 3, 2012 at 20:48
ssh onto the box and run git pull or git clone? Or create a script that will do that for you? — Justin Wood
– Justin Wood, Commented Oct 3, 2012 at 20:49
@cdhowie As İ said, there is no error.log. Dunno what happens — heron
– heron, Commented Oct 3, 2012 at 20:51
Maybe your script is timing out? add set_time_out(0); at the top of your script. — Justin Wood
– Justin Wood, Commented Oct 3, 2012 at 20:56
I know this may be a dumb question but have you already tried running this script locally? It could be one of several issues, including settings on the host. — beenhere4hours
– beenhere4hours, Commented Oct 4, 2012 at 3:30

Denis Kreshikhin · Accepted Answer · 2012-10-07 01:00:59Z

This solution do not provides authentication:

// download the repo zip file
$repofile = file_get_contents("https://bitbucket.org/$username/$reponame/get/tip.zip");
file_put_contents('tip.zip', $repofile);
unset($repofile);

But curl allows it. So a zip archive can be downloaded from a private repository in same way like in first script.

$node = ''; // a node from repo, like c366e96f16...

$fp = fopen($path, 'w');

$ch = curl_init("https://bitbucket.org/$username/$reponame/get/$node.zip");
curl_setopt($ch, CURLOPT_USERPWD, "$username:$password");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true); 
curl_setopt($ch, CURLOPT_FILE, $fp);

$data = curl_exec($ch);

curl_close($ch);
fclose($fp);

I have tested it for my bitbucket account. It's work very well.

If necessary to get last changeset node that we should use bitbucket api GET a list of changesets:

$username = 'login';
$password = 'pass';
$owner = $username; // if user is owner
$repo = 'repo name';

$response = "";
$callback = function($url, $chunk) use (&$response){
    $response .= $chunk;
    return strlen($chunk);
};

$ch = curl_init("https://api.bitbucket.org/1.0/repositories/$owner/$repo/changesets?limit=1");

curl_setopt($ch, CURLOPT_USERPWD, "$username:$password");

curl_setopt($ch, CURLOPT_HEADER, 0);
curl_setopt($ch, CURLOPT_HTTPHEADER, array('User-Agent:Mozilla/5.0'));
curl_setopt($ch, CURLOPT_WRITEFUNCTION, $callback);
curl_exec($ch);
curl_close($ch);

$changesets = json_decode($response, true);
$node = $changesets['changesets'][0]['node'];
$raw_node = $changesets['changesets'][0]['raw_node'];

print($node . PHP_EOL);
print($raw_node . PHP_EOL);

dag · Accepted Answer · 2012-10-07 03:07:25Z

1

I recently discovered Capistrano which is a great tool. It was initially developed for ruby but it's also great in combination with php http://www.davegardner.me.uk/blog/2012/02/13/php-deployment-with-capistrano/

answered Oct 7, 2012 at 3:07

dag

1307 bronze badges

Comments

Tural Ali · Accepted Answer · 2012-10-07 19:16:42Z

Based on your update, replace you php files contents with code below:

<?php
// Set these dependant on your BB credentials    
$username = '';
$password = '';

// your Bitbucket repo name
$reponame   = "";

// extract to
$dest = "./"; // leave ./ for relative destination
// Grab the data from BB's POST service and decode
$json = stripslashes($_POST['payload']);
$data = json_decode($json);

// set higher script timeout (for large repo's or slow servers)
set_time_limit(5000);


// Set some parameters to fetch the correct files
$uri = $data->repository->absolute_url;
$node = $data->commits[0]->node;
$files = $data->commits[0]->files;

// download the repo zip file
$fp = fopen("tip.zip", 'w');

$ch = curl_init("https://bitbucket.org/$username/$reponame/get/$node.zip");
curl_setopt($ch, CURLOPT_USERPWD, "$username:$password");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
curl_setopt($ch, CURLOPT_FILE, $fp);

$data = curl_exec($ch);

curl_close($ch);
fclose($fp);

// unzip
$zip = new ZipArchive;
$res = $zip->open('tip.zip');
if ($res === TRUE) {
    $zip->extractTo('./');
    $zip->close();
} else {
    die('ZIP not supported on this server!');
}

// function to delete all files in a directory recursively
function rmdir_recursively($dir) {
    if (is_dir($dir)) {
        $objects = scandir($dir);
        foreach ($objects as $object) {
            if ($object != "." && $object != "..") {
                if (filetype($dir . "/" . $object) == "dir")
                    rmdir_recursively($dir . "/" . $object); else
                    unlink($dir . "/" . $object);
            }
        }
        reset($objects);
        rmdir($dir);
    }
}

// function to recursively copy the files
function copy_recursively($src, $dest) {
    if (is_dir($src)) {
        if ($dest != "./")
            rmdir_recursively($dest);
        @mkdir($dest);
        $files = scandir($src);
        foreach ($files as $file)
            if ($file != "." && $file != "..")
                copy_recursively("$src/$file", "$dest/$file");
    }
    else if (file_exists($src))
        copy($src, $dest);
    rmdir_recursively($src);
}

// start copying the files from extracted repo and delete the old directory recursively
copy_recursively("$username-$reponame-$node", $dest);

// delete the repo zip file
unlink("tip.zip");
?>

Update

Here are repositories of this script (Modified by Me) on

Collectives™ on Stack Overflow

Setting automatic GIT deployment of PHP project

Update

3 Answers 3

1 Comment

Comments

Update

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

Update

3 Answers 3

1 Comment

Comments

Update

Comments

Your Answer

Sign up or log in

Post as a guest

Related