Seems pretty unsafe for me to have that information stored in Google.

While writing some unittests, i was confused by the documentation of php regarding memcache.
I always wondered why there was a pecl/memcache and a pecl/memcached. I assumed one was the client API, the other the server API (or something similar).

BEEEP! WRONG!

Seems that both are libraries providing methods for accessing the memcached (memcache daemon).

You have to choose one! or… use them both… but have to be very aware which one you are using :)

I have chosen to use memcached. Primarly because it’s only a wrapper to libmemcached (which is a big plus in my eyes).

At least… that was what he claimed. I don’t use php on the cli a lot, so just to be sure i wrote a quick test script, which executed flawless.

So I took a look at the script he tried to execute:

echo html_entities_decode(file_get_contents($argv[0]));

So the script did exactly what was asked ;) outputting the content of the script itself, instead of outputting the content of the file given as 1st argument to the script.

hi hi.
He overlooked the fact that the 0-key of the argv array holds the name of the script itself, instead of the 1st argument given on the cli :)

The people of MySQL AB seem to be very happy with the take-over. I can imagine that. It will give for sure some financial ease of mind for the company and his employees and access to a large source of knowledge and resources. Somehow many people, including myself, have a suspicion against “the big corporations”. With the exception of Apple computers ofc :).
Don’t forget however that Sun has a decent and reliable history with Open Source. Open Office and making Java open source to name 2 big achievements of Sun.
I use MySQL now for close to 9 years. I have learned to love this database engine, discovering the more advanced features of SQL step by step. I only have used MySQL directly or interfaced by PHP.
I can only express my hope that Sun will support the continuous development of the MySQL API as used by the MySQLi or mysqlnd extensions of PHP.
Some analysts suspect that Sun is busy to build his “own” alternative to the classic LAMP (Linux, Apache, MySQL, PHP) stack. The Sun-Stack would be SAMJ (Solaris (opensourced!), Apache, MySQL, Java).
People should have the freedom to use and code the language they prefer.
I really hope it won’t be soon that if you want to have the full support for the capabilities of MySQL, you have to use a Java interface to interact with the database engine from a web-environment.

I think that webapplication-coding is by nature more suitable for a procedural style. You have a very clear starting point and end. So the timeline in your script is very linear. Compare this with java, where applets mostly run in a loop waiting for userevents. The fundamental difference in nature (begin->end against looping) explains my struggle to code completely OOP with PHP.

Off course OOP has lots of advantages with PHP (execption handling, code reusability to name a few) but I think when coding webapplications, you will always find yourself coding a more or less procedural timeline within your script.

Being a total version control newbie, I did my googling and reading.
In one sentence: CVS is old and insecure, Subversion is hot and the talk of the town. CVS compares to Subversion as telnet to ssh.
And the fun thing? installing and setting up took me 1,5 hour. Using ports
there are basically two ports to install: Apache2 and Subversion.
Below are links to the pages I used for setting up my subversion server, but here are some hints/notes: (not a howto or tutorial! read the pages!)

Create a svn user and group, and adjust the umask of that user. I used: umask 002.

Build apache2 with WITH_BERKELEYDB=db42.
I have changed httpd.conf to have apache2 runs as svn user and group, this spares you from read/write issues when using Subversion with webdav.

Build Subversion with -DWITH_SVNSERVE_WRAPPER -DWITH_MOD_DAV_SVN.

Su -l to the SVN user, create the repository and import your initial project layout, confiure apache to serve the repository with webDAV and…you’re free to go!

It kicks ass!
I do my webdevelopment on my mac offcourse and I found svnX to be a fantastic client wrapper. 2, 3 hours after installing and setting up my first version control, I found myself wondering how I ever lived (eh..ok developed) without it.

The links I used:
onlamp – Setting up a Secure Subversion Server
the SVN book (free!)

macdev center- making the jump to subversion
onLamp – top ten tips for Subversion users
OS X Subversion client package (you need this for being able to use a gui front-end as svnX.
svnX
Versions Systems link collection

And because I always forget the exact procedure:
Apache+SSL on freeBSD

Recent addition is this article from RedHat, a introduction suited for both CVS users and those new to version control.

CVS is out, Subversion is in found with digg.com

The real fun is that I have to use windows again. The course uses Borland Jbuilder 8, with a license which only runs on windows XP according to the manual, but I use it on my very old windows98 machine without any problems.
And oh boy, what is an IDE fun. Programming like playing with LEGO.

“You have forgotten a semi-colon right here”.
Oh…thanks :)

Instead of digging through 2500 lines of code where that missing T_SPACE is…
Java is also fun to work with. Very straightforward and easier than C (no memory management / garbage collection needed). I really can tell know where PHP has it’s roots. It really helps me getting some insight in OOP concepts, which are used in PHP 5 as well.

btw…writing this on windows, WordPress really looks better in Firefox as in Safari…damn.

For over a year I glanced at PHP5, but put it aside as too unstable/experimental (a year ago) or too dangerous to migrate due to incompatibility issues.

Some of my projects where however advancing to a point, that procedural programming became to limited. And more and more I found myself looking at (Object Orientated Programming) OOP.
And being a absolute rookie in OOP, I decided that if I had to learn OOP, I wanted it to learn in PHP5 rightaway, instead of starting in PHP4.

Well, 2 weeks ago, I decided that the time has come. I downloaded php5, installed it on my development server and….it just worked?!

A closer look learned that JPgraph -scripts broke and some my legacy code also.
Fortunately, there are two configuration options who both can be set in a htaccess or virtualhost configuration:
zend.ze1_compatibility_mode
and
register_long_arrays

Enabling those two directives in the appropiate virtualhost config made all the scripts 100% working again!(yeah yeah I know, ditch the legacy code and wait for jpgraph 2.0, but we don’t live in a perfect world, do we?)

The last week I find myself coding in PHP5 OOP style for the first time, using the new mysqli extension (prepared statements kick ass!) and enjoying php5 very very much!

Being used to PHP, it took some frustrating tries with javascript, but some hours later I was looking at my first widget.
It basically uses Curl to fetch PHP generated content over a HTTPS connection.
Shown are the last five lines of a access-log from a webapplication.
It works like a charm and refreshes every time dashboard shows up, or with the command+r keycombo.

If Tiger is released and the NDA is lifted, I’ll probably post the widget for download…untill then, some screenshots:

the front, some content blurred for privacy reasons

and the backside, i simply love the animation :)

But running php 4.3.6 and mysql 4.0.20 was getting…unwise.

Upgrading PHP is easy, with the (imho) great installer of entropy.ch.

installer -verboseR -pkg ~/php-4.3.10.pkg -target /

Don’t forget afterwards to restore your httpd.conf and to edit /usr/local/php/httpd.conf.php to add:

AddType application/x-httpd-php .php
AddType application/x-httpd-php .html
AddType application/x-httpd-php-source .phps


MySQL is somewhat more tricky.
The version supplied by Apple is well…historic.
I like to compile MySQL from source (esp. because that takes 8,5 minute on my Xserve :) and install it to /usr/local/mysql-(version).
After that I create a symlink from /usr/local/mysql-(version) to /usr/local/mysql.
Take some care replacing the apple supplied mysql-binaries with symlinks to the up2date /usr/local/mysql/bin/ counterparts.
After this, upgrading mysql from source is painless.

The most recent upgrade:

# cd /usr/local/src
# gnutar -xzvf mysql-4.1.10a.tar.gz
# cd mysql-4.1.10a
# ./configure --with-openssl \
--prefix=/usr/local/mysql-4.1.10_a \
--localstatedir=/usr/local/mysql-4.1.10_a/data\
--mandir=/usr/local/share/man/\
--with-mysqld-user=mysql
# make
# make install
# cd /usr/local/
# mysqladmin -u root -p shutdown
# cp -R mysql/data mysql-4.1.10_a/data
# chown -r mysql:wheel mysql-4.1.10_a/data
# ln -s /usr/local/mysql-4.1.10_a /usr/local/mysql
# mysqld_safe &


That’s all. If something goes wrong, just restore the symlink to the “old” mysql-(version) and you’re safe.
Be aware of the new password hashing mysql uses from 4.1.1 of. I used the –old-password switch untill I’ve upgraded all my clients.