Shhht! Don’t tell anyone, the secret command to get the code is:

bzr branch lp:~rapache-devel/rapache/rapache-stage0

Some people asked how to help.

Well, the first way to help is to open new bugs (even for feature requests, just open a bug and explain your idea). You can do it at https://launchpad.net/rapache .

The second way is actually… to solve bugs, or indicate a way to solve them. I’ve tagged a bunch of them as ‘need-help’. You cand find them here: https://bugs.launchpad.net/rapache/+bugs?field.tag=need-help. For this kind of bugs, we don’t need a full patch, you can just send a proof of concept indicating how to solve the bug. That would save us a nice amount of googling .

I forgot to say: thank you thank you thank you.

Thank you thank you thanks to you all for having told me that rapache gui sucked. Shame on you . I tried to re-organize the main window at the best of my skills. We also had to wake up Luana and ask her to revamp our application icon.

Here is the result of all the effort:

As you can see Rapache can now take care of vhost definitions only present into /etc/apache2/sites-enabled and not present inside sites-available (the correct way would be to create the vhost configuration file inside sites-available and then link to it from sites-enabled in order to enabling it).

How do Rapache and Ubuntu relate ?

My very first aim with Rapache was to help webdev beginners and webdesigner (oh, they hate configuration files so much) to quickly be able to handle apache in the correct way. Giving support in ubuntu irc support channels I noticed many people came there asking how to set up apache and, even easy as it is creating a new virtual host, helping them required a lot of time each.

As for now, the primary goal of apache is to get in a decent shape soon enough to be proposed for Intrepid universe repositories. In order to be able to accomplish such goal, we should focus on a realistic set of features and test them enough to make everything working smoothly. It’s of course better a small working program than a big buggy ball of mud.

That said: which features do you think we still miss in order to make sense for a repo inclusion ? Would you open some bug in launchpad and tell us about what we need to do ? Help us hit the duck .

Come say “hello!”

Last post I forgot to tell we have an irc channel on freenode. Feel free to drop in #rapache to greet us and talk to us.

What is Rapache ?

Rapache is a simple Apache administration tool. It offers a GTK interface to allow extremely rapid VirtualHost creation and manipulation.

Favicons add a nice touch to the list

Personally, as a web developer, I felt the need of a simple tool to manage virtualhosts on my localhost for a long time, just because creating them at hand is way too time consuming than I’d like. Then, right after stumbling on this and reading users comments (some may be biased, I know), I decided to make such a tool by myself.

Funnily enough, me and Emgent had the same idea at the same time so, when he announced the project, I got in touch with him and we decided to join forces.

What to expect

Personally I would have not announced the project before having already put together the basic program. We divided our roadmap in basically 3 stages, the first being my original previous goal.

• Stage 0: basic functionality on localhost. Virtual hosts and Apache modules enabling/disabling/modification. Aimed to web developers.
• Stage 1: Localization, more polished Gui. Functional separation before GUI and Libraries (i.e. a number of handy commands you can use from commandline). A number of handy wizards to handle common configuration issue. More Apache configuration options available from the GUI
• Stage 2: Sky’s the limit . Handling remote servers via SSH. Server bookmarking.

What not to expect

At the moment we’re relying on debian way to handle Apache configuration. So do not expect to be able to run Rapache on a non-debian system at lease before Stage 3, maybe not even afterwards. That’s mainly because the debian way of manage Apache confs is much more easy to handle than the upstream’s one.

In a way we’re creating Rapache for Ubuntu itself (and Debian, by reflex) rather than for Linux in general. Despite of that we’re open for people willing to help us hacking it to work on non-debian systems either.

Where to find us

• Launchpad project page: https://launchpad.net/rapache
• Messy wiki page: https://wiki.ubuntu.com/Rapache

Xdump was one debugging tool I had to develop for the reason that in some situations var_dump() simply is not enough.

Here’s the overview of xdump.

Of course one could say that as good programmer you should never ever findy yourself in such a situation which requires something more powerful than var_dump() or print_r(). I can understand this kind of point-of-view, but I found that real world not always is like that.

Today at work, for example, I had to add some functionality to a zen-cart installation. This is something var_dump simply cannot manage (without driving you crazy, I mean).

Download

I will try to make download available as soon as I can.

Download available on google code

Download

Download zipped source from Google Code

Or browse svn repository

Download source from svn using: svn checkout http://xdump.googlecode.com/svn/trunk/ xdump

Brief story

One of our previous sites (about 50k page-views per day) was using a little php script for resizing dinamically images to thumbnail size. Since we didn’t knew in advance the size we was gonna need, resize was made at run-time.

With site’s restyling the number of resized images on every page increased sensibly. As it’s easy to guess server load went sky-high. We tried to replace the package with PhpThumb but the situation got even worse (yes, worse was possible).

What’s wrong with PhpThumb

I have some problems with PhpThumb. Oh yes, I do.

• It’s bloated. Even if PhpThumb is a class, the caching routines are written in a procedural file who reads $_GET superglobal to decide what to do. This means I had to simulate HTTP request by emptying manually $_GET and putting there only values PhpThumb needs. You don’t pass through that file.. you don’t cache.

• Cache working.

  • Does it work? I couldn’t verify, the only thing I know is that the cache directory kept becoming bigger, with apparently no limit. Actually, it became so big to cause server load by itself.

  • I tried disabling cache in the config. Guess what? It kept caching, yes it did! I had to physically delete the cache path to block caching.

  • Cached file retrieving. It’s made using readfile() or header redirects. While that’s better than nothing I think we can do better, much better.

Strategy

The solution I choose is what some call generate on request. This trick consists in creating a dynamic script, making it the default Apache’s ErrorDocument for the current directory and let it handle the content generation.

The flow:

1. Assume we store all downloadable wallpapers on our site are stored into /image/wallpapers directory.

2. The website’s page will contain a non existant image path. Say something like: http://example.com/dynamic/40×30/images/wallpapers/britney.jpg

3. At the time of the first visit to the page the resized image will not be there yet.

4. When the image requests hits Apache, it will be redirected to our custom ErrorDocument (front_controller.php).

5. front_controller.php will take care to read /images/wallpapers/britney.jpg generate our 40×30 nice thumbnail at /dynamic/40×30/images/wallpapers/britney.jpg path.

6. Afterwars the user browser will be redirected to the generated image, making it appear on the page.

7. Every other request to the page will find the image already at the path requested, without needing PHP to do anything.

The Code

Here’s some code boxes to let you know how code looks like.

.htaccess

The only purpose of .htaccess here is to instruct to consider our Front Controller the default error page for the generated images path.

The first time an image is requested the Front Controller manages to create it at the requested path, and to redirect the user to it.

The following requests will find the image already waiting for them. This means the file will be served by Apache itself with no php intermediation. This can lead to performance improvement around the order of magnitude.

ErrorDocument 404 /dynamic/front_controller.php

Front Controller

The purpouse of this file is to include libs, instantiate main class, decide which drivers are to be used and make thing run.

include ( '_libs/logger.php' );
include ( '_libs/cache_on_request.php' );
$FrontController = new GenerateOnRequest();
//decomment the following 2 lines only for debug purpouse
$FrontController->configuration['debug']=true; //prints logs (and hinibits header final)
//$FrontController->configuration['stopGeneration']=true; //checked by drivers, hinibits concrete file generation
//attacching drivers
$FrontController->attachDriver( new ResetController );
$FrontController->attachDriver( new GdImageResizer );
$FrontController->run( $_SERVER[ 'REQUEST_URI' ], dirname(__FILE__ ) );

php, lib-dev

Have you ever tried to var_dump $GLOBALS ?

Even if when I started this lib there were already some classes which provided nice html formatted dump, in the moment I needed one, I couldn’t (my fault) find one.

Xdump started as a bunch of procedural lines of code and slowly evolved in a object-oriented library which can provide nice collapsable html dumps and backtraces, spoof references (to avoid dump bloat and recursions) and show source code lines where dump was invoked as well as display on screen source code lines for every backtrace jump.

Download

Xdump Download Page

Overview

Xdump let’s you display php vars nicely on the screen.

• Jump to examples

• Test cases are provided (requires SimpleTest)

Usage

1. Include “xdump/xdump.php” in your main include file.

2. Use following code:

echo xdump::dump($var);

1. To show a backtrace instead type:

echo xdump::backtrace();

Samples

Dumping get_defined_vars() inside a Zen-Cart template

Licence and contact

• Xdump is MIT Licensed.

• Info, suggestions and stuff to: stefano@stefanoforenza.com

Features

General

• Displays php dumps more nicely

• Dumps are javascript contractible. They occupy little space on the page and open when clicked.

• Display nicely without styles and javascript as well (if you work with them disabled)

• Shows file, line number and source lines around the dump invocation (no more digging around to remove debugging code). Shows source lines for every backtrace jump as well (neat! try it).

• (almost) Plug & Play

Technical

• Handles reference recursions inside objects and arrays (no endless loop). Will signal visually nested references (the parent referenced element will blink orange).

• PHP5 Overloading aware (doesn’t touch get, set etc)

• PHP4 Compatible

• Shows you backtrace info and the chunk of the code that invoked the dump

• Some little effort to limit the size of markup inside the dump. Keep in mind that the dump will anyway be bigger than a vardump