Logaholic.de

Note: If you don’t want to or cannot by company-rule trust Amazon S3, this is probably not what you want to read.

I’m going to show you how one could improve the security of his production environment backups. In this setup the production environment can never harm any old backup. There is another offsite backup location if Amazon S3 should fail – for us being our office, this is also perfect for up-to-date testing/development database snapshots.

We will need two separate Amazon S3 accounts and any s3 console tool (like s3bash).

The production environment, with its own (restricted to put and get files) s3 user, will push backups to our s3 bucket. It is not allowed to delete backups there. It does not have any access to the offsite location. If someone gets access to our production environment, he can not delete our backups on s3 per their acl, and can never harm our offsite backups. This is kind of an one-way solution and represents the “push-strategy“.

The offsite location, with the second, full privileged s3 account, will pull the backups from s3 every night. There are also tools for backup verification and testing-environment updates. This is the “pull-strategy“. The offsite location has access to the production environment for maintenance task and deployment.

Small graphic:

I wanted to mention this setup since i read a blogpost about a worst case scenario:

“A huge flight sim site was hacked and destroyed this weekend  – avsim.com. An important lesson on why off-site backups are critical! They had two servers, and had a backup of A on B, and B on A. Both were taken out.”

Downsides:

• Access to our office network/offsite backups would be bad.
• Bruteforcing/getting access to our administrative S3 Account would be bad.
• you have to trust Amazon with your data

Conclusion:

Always combine different backup strategies and test your backups. One day you will need them!

I am currently building an AIR application with JavaScript (jQuery + jQuery UI), one function downloads zipped files and has to extract the contents somewhere to the local filesystem. Since the files may contain subfolders, i didn’t find any working example supporting subfolders in the docs or via google.

I found an Adobe Tutorial for basic Zip files (see Sources), and after reading the zip specs there was just one small change to their code to get it working with subfolders.

The spec says that entries in the zip file which are directories just end with an ‘/’. Skipping them is the whole magic. (see line 48)

Thankfully, the air.FileStream autocreates director ies, so writing the files to the subfolders does not need any change to the code. I added a basepath as parameter to the outFile function, to complete the for me intuitive function unzipFile(sourceFile, targetPath).

Here is the code:

 function unzipFile(sourceFile, targetPath)
{
	var bytes = new air.ByteArray();
	var fileName = new String();
	var flNameLength;
	var xfldLength;
	var offset;
	var compSize;
	var uncompSize;
	var compMethod;
	var signature; 

	var zfile = air.File.applicationStorageDirectory.resolvePath(sourceFile);
	var zStream = new air.FileStream();
	zStream.open(zfile, air.FileMode.READ);
	bytes.endian = air.Endian.LITTLE_ENDIAN;

	while (zStream.position < zfile.size)
    {
		// read fixed metadata portion of local file header
        zStream.readBytes(bytes, 0, 30);

		bytes.position = 0;
        signature = bytes.readInt();
        // if no longer reading data files, quit
        if (signature != 0x04034b50)
        {
            break;
        }

		bytes.position = 8;
        compMethod = bytes.readByte();  // store compression method (8 == Deflate)

		offset = 0;    // stores length of variable portion of metadata
        bytes.position = 26;  // offset to file name length
        flNameLength = bytes.readShort();    // store file name
        offset += flNameLength;     // add length of file name
        bytes.position = 28;    // offset to extra field length
        xfldLength = bytes.readShort();
        offset += xfldLength;    // add length of extra field

		// read variable length bytes between fixed-length header and compressed file data
        zStream.readBytes(bytes, 30, offset);

		bytes.position = 30;
        fileName = bytes.readUTFBytes(flNameLength); // read file name 

		if (fileName.substr(fileName.length - 1, 1) != '/')
		{
			bytes.position = 18;
			compSize = bytes.readUnsignedInt(); // store size of compressed portion
			bytes.position = 22; // offset to uncompressed size
			uncompSize = bytes.readUnsignedInt(); // store uncompressed size 

			// read compressed file to offset 0 of bytes; for uncompressed files
			// the compressed and uncompressed size is the same
			zStream.readBytes(bytes, 0, compSize);

			if (compMethod == 8 ) // if file is compressed, uncompress
			{
				bytes.uncompress(air.CompressionAlgorithm.DEFLATE);
			}
			outFile(targetPath, fileName, bytes); // call outFile() to write out the file
		}
	}
}

This is the outFile function which writes the uncompressed files to the local file system.

 function outFile(baseDir, fileName, data)
{
    var outFile = air.File.applicationStorageDirectory;
    outFile = outFile.resolvePath(baseDir+air.File.separator+fileName);
    var outStream = new air.FileStream();
    outStream.open(outFile, air.FileMode.WRITE);
    outStream.writeBytes(data, 0, data.length);
    outStream.close();
}

Sources:

• Adobe Air 1.1 – ByteArray example: Reading a .zip file
• PKWARE – .Zip File Format Specification

1. Run > External Tools > External Tools Configurations…

2. Add new (External Program) Configuration

3. Main Tab:

• Location: Point to your Adobe Air SDK\bin\adl.exe
• Working Directory: Point to your project folder (i did this via variable in the screenshot)
• Arguments: your main application configuration, usually application.xml

4. Finished. Run!

If you don’t see the “Program” entry in 2), check if your eclipse filters it here:

After installing phpUnderControl I wanted to use my existing apache, running with ssl, to proxy requests to the “java stuff”. After failing with various mod_proxy reverse proxying attempts I learned how to use mod_rewrite for this purpose.

ProxyPreserveHost on
RewriteEngine on

RequestHeader Set Proxy-keysize 512
RequestHeader Set Proxy-ip %{REMOTE_ADDR}e
RequestHeader Set Host example.org:443

RewriteRule ^/$ /cruisecontrol/ [R,L]

RewriteRule ^/cruisecontrol$ /cruisecontrol/ [R,L]
RewriteRule ^/cruisecontrol/(.*) http://localhost:8080/cruisecontrol/$1 [P,L]

RewriteRule ^/dashboard$ /dashboard/ [R,L]
RewriteRule ^/dashboard/(.*) http://localhost:8080/dashboard/$1 [P,L]

Unfortunately, phpUnderControl didn’t get the baseURL right with this setup. I didn’t want to dive into mod_jk yet, which I had read should also work (or be the more clean/generic solution), so i poked at the phpUnderControl code with grep. And now it just works fine.

You have to edit the files error.jsp, index.jsp, main.jsp and old_index.jsp in /opt/cruisecontrol/webapps/cruisecontrol/: Just place your https-url in the baseURL strings (https://example.org/cruisecontrol/).

The CruiseControl dashboard (/dashboard) does not need any tweaking for https.

One could even add any apache authentication module in front of phpUnderControl now

Do you use HTTP Authentication anywhere in your webapplication? Today I discovered something that could also be useful for you.

The plusserver member control panel uses HTTP Authentication, and the errorpage replacement for the HTTP 401 authentication failed page (press cancel on the login) is replaced with a simple form with two elements: username and a button “receive lost password”.

This is the most useful errorpage-replacement i have ever seen yet You could also add a link to a faq for common questions/problems there.

This is a small review of Daniel ’smacks’ Harringtons beta php5 oop mvc framework “simples”. He sent me this comment yesterday:

Hi Karsten,

dein Beitrag ist zwar schon etwas her, steht aber bei Google zum Thema Micro-Framework ziemlich weit oben. Da ich gerade ebenfalls in der Situation war und etwas einfaches wie Sinatra für PHP gesucht und nichts ansprechendes gefunden habe, hab ich mich letztendlich hingesetzt und selber etwas zusammengestrickt.

Es hört auf den Namen “simples”, bietet Routing, MVC und ein bischen mehr, bietet aber z.B. keine Datenbank-Unterstützung an und kommt mit recht wenig Code aus. Da es bis jetzt noch keine Doku gibt, lohnt es sich auch mal im Code zu stöbern.

Feedback ist immer willkommen!
http://github.com/smacks/simples/tree/master

Ps. Das Ganze bitte noch als Beta betrachten. Für den produktiven Einsatz ist es bisher noch nicht gedacht.

He says that in his search for a lightweight php mvc framework, he didn’t find anything that satisfied his needs, and so he started to write his own. It is called “simples”, is still beta, has no comments, no documentation, no database support, but offers mvc, routing, “and some more”. It is not meant for production usage at the moment.

The first thing I noticed is that all of the configuration is done via constants. I don’t think a framework, which will be only one part of an application, should pollute global scope (even constants) too much, if at all. One improvement could be to add a prefix to the configuration (smacks_*), but I highly prefer a config class, which loads its content based on an environment (which itself could be set via ONE prefixed constant) via xml or php files. You can have different configurations for different purposes this way, and just have to switch the environment somewhere.

Another point considering ‘namespaces’ (in a general, not the php 5.3 namespaces way) is, that all “simples” framework classes also don’t have a prefix. You don’t want collisions there.

I also miss autoloading (do I always want to load anything? no.) and a response object.

All in all, the most important point is: Why should I want to use simples? I don’t see one point which is outstanding somehow… yet

Today i’d like to share my very own LAMP (Linux Apache Mysql PHP) development tool list for windows.

I don’t really care if it is Windows XP 32 or Vista 64, if the system supports all the hardware my workstation has (i.e. more than 4gb ram -> Vista 64).

Tier one tools (misc dev tools):

• Mozilla Firefox – browser
  • Firebug – developer plugin: “Firebug integrates with Firefox to put a wealth of web development tools at your fingertips while you browse. You can edit, debug, and monitor CSS, HTML, and JavaScript live in any web page.”
    • Firephp – Firebug extension: “FirePHP enables you to log to your Firebug Console using a simple PHP method call. All data is sent via response headers and will not interfere with the content on your page. FirePHP is ideally suited for AJAX development where clean JSON and XML responses are required.”
    • Yslow – Firebug extension “performance”: “YSlow analyzes web pages and why they’re slow based on Yahoo!’s rules for high performance web sites.”
  • Selenium IDE – developer plugin “test automation”: “Selenium IDE is an integrated development environment for Selenium tests. It is implemented as a Firefox extension, and allows you to record, edit, and debug tests. Selenium IDE includes the entire Selenium Core, allowing you to easily and quickly…”
  • Firegestures – usability plugin: “A customizable mouse gestures extension which enables you to execute various commands and user scripts with five types of gestures.”
• Scite – lightweight, extremely fast text editor, including syntax highlighting, folding, tabbed interface
• Zend Studio/Eclipse (main IDE) – full featured IDE
  • PDT – PHP Development Tools
  • Mylyn – task & planning tools, issue/bug tracker integration
  • Subversive – subversion plugin
• PuTTY – ssh/telnet tool: “PuTTY is a free implementation of Telnet and SSH for Win32 and Unix platforms, along with an xterm terminal emulator.”
• WinSCP – sftp/scp tool: “WinSCP is an open source free SFTP client and FTP client for Windows. Legacy SCP protocol is also supported. Its main function is safe copying of files between a local and a remote computer.”
• Total Commander – file manager, ftp tool, two file windows
• WinRAR – archiver for rar/zip/tar/gz/…
• TrueCrypt – highly secure, portable data storage: “Free open-source disk encryption software for Windows Vista/XP, Mac OS X, and Linux”
• TortoiseSVN – subversion client: “A Subversion client, implemented as a windows shell extension.”
• VMware Workstation – virtual machines
  • LAMP vm
    • Xdebug – debug helpers: “The Xdebug extension helps you debugging your script by providing a lot of valuable debug information.”
    • phpMyAdmin – web-based database administration: “phpMyAdmin is a free software tool written in PHP intended to handle the administration of MySQL over the World Wide Web. phpMyAdmin supports a wide range of operations with MySQL. The most frequently used operations are supported by the user interface (managing databases, tables, fields, relations, indexes, users, permissions, etc), while you still have the ability to directly execute any SQL statement.”

Tier two tools (specific dev tools):

• Opera/Safari/Chrome – different browsers for testing
• IrfanView – image viewer for lots of formats, (batch-)converter
• MySQL Workbench 5.0 – erd/db planning tool, reverse engineering for sql create scripts, some smart wizards, database visualisation
• BareTail – tail -f for windows (realtime logfile monitoring)
• WinCachegrind/CachegrindVisualizer/KCachegrind in VM – cachegrind visualisation, inspect profiler (xdebug) output
• CloudBerry S3 Explorer – very nice Amazon S3 managing tool

Tier three tools (communication/support):

• Microsoft Outlook (if Exchange is available) – emails, calendar
• Mozilla Thunderbird – emails, calendar
  • Enigmail – OpenPGP security
• UltraVNC – remote desktop, direct connection
• TeamViewer – remote desktop through firewalls, easy and fast setup – very good for helping “not so experienced” users
• Miranda IM – lightweight IM client (icq, aim, msn, jabber, …)
• mIRC – “the” IRC client
• Skype – voice chat

Tier four tools (music/media)

• Foobar2000 – lightweight audioplayer

Windows Settings:

• fixed taskbar with quicklaunch (icons have always the same order), visible time and no “hide unused icons” stuff

“I’m pleased to announce the Zend Framework 1.8.0 release, the first in our 1.8 series of releases. This release marks the culmination of several long-standing projects, as well as a formalization of many of our recommended practices. There are two major stories in this release: first, the addition of several components designed to provide and promote Rapid Application Development; second, two offerings that make using Zend Framework in the cloud easier.” [1]

Some thoughts, in no particular order:

If you know Amazon S3 (Amazon Simple Storage Soluation, a web-service for storing and receiving files, scalable, fast, safe) then you should have a look at Zend_Service_Amazon_S3. The Zend Framework not only offers a nice object oriented implementation, but also provides a PHP Stream Wrapper. Why is this so nice? Because one could add Amazon S3 support to existing applications by simply prefixing any standard file-operation with ’s3://’. This is the code sample from the documentation:

<?php
require_once 'Zend/Service/Amazon/S3.php';

$s3 = new Zend_Service_Amazon_S3($my_aws_key, $my_aws_secret_key);

$s3->registerStreamWrapper("s3");

mkdir("s3://my-own-bucket");
file_put_contents("s3://my-own-bucket/testdata", "mydata");

echo file_get_contents("s3://my-own-bucket/testdata");

Support for Amazon EC2 (Amazon Elastic Comput Cloud) has also been added (Zend_Service_Amazon_Ec2).

“Amazon EC2 provides a web service to allow launching and managing server instances within Amazon’s data centers. These server instances may be used at any time for any length of time — allowing you to scale your site only when you need to handle extra traffic, or run your services entirely from the EC2 platform.” [1]

Zend Framework jumped the train for “the” cli interface to the framework via Zend_Tool. One could create whole projects, models, controllers, views with it. This makes sense for starters imho. My full featured Zend Studio for Eclipse with customized code templates does this job way better for me. One thing i miss (Agavi has it! ^^) is a phpunit interface and some configuration which tests should be run. In my opinion, just the existance of such an option would encourage more users to think about/actually use unit tests.

Routing now supports translation aware routes, and route chaining capabilites. Those are fetures i know and love from Agavi.

There are loads of other new features (see [1]), which I haven’t checked yet – sometimes simply because they didn’t interest me.

I’m curious if switching our main project on monday to ZF 1.8.0 will break any test ^^.

Sources:
[1] Zend Developer Zone: Zend Framework 1.8.0 Released

Yesterday i found some code i wrote in 2002. I was bored in school, and was asking myself if it would be possible to let PHP play tetris, including some visualisation. There was no AJAX or any well-known JavaScript Framework  at that time (of which i had known), so after some “print a full game table for every move you do” sessions, i finally (ab)used outputbuffer-flushing and some DHTML (dom-manipulation through javascript).

The whole thing is only one ~25kb file and resembles a full tetris playing program

Since it was a fun project in my earlier stages, there are some things missing: “good code”, comments, coding standard (naming schemes), oop, … But it works.

The “AI”-part of the script is a “try and weight every possible move, choose the best” approach – one can tune the formula for this calculation for different results.

aiTris – PHP playing Tetris – Demo – just press the start button

I hereby release the source under the same license as i have chosen for this blog (creative commons share alike), so have fun!

Since logaholic.de was kind of inactive some weeks ago, my fellow co-blogger Pierre decided to go his own way with his own blog again. I’d like to thank you for all your effort for logaholic.de and wish you all the best. May the content be with you, my friend! And let there be pingbacks ^^

By the way, i’d like to point to Pierre’s own PHP Micro-framework entropy, which is in work again. I have seen some magic stuff in there so far, so let’s stay tuned for his release(s)