Saturday, November 23, 2013

Joomla 3.2 and captcha not appearing

So I activated Joomla 3.2's captcha feature for contact forms.

It turned out it wasn't appearing.

So I followed this tutorial (must do) but still it wasn't working.

It didn't took much to realize I had still to fix some bugs in the lines 23-24 of the file plugins\captcha\recaptcha\recaptcha.php

I changed those lines so they read:

 const RECAPTCHA_API_SERVER = "http://www.google.com/recaptcha/api";
 const RECAPTCHA_API_SECURE_SERVER = "https://www.google.com/recaptcha/api";
 const RECAPTCHA_VERIFY_SERVER = "http://www.google.com";

Bottom line, the "http://" was missing...

And after that it worked like a charm : )


Wednesday, October 30, 2013

scanning exim frozen mails for spam

I had a bunch of frozen emails. I wanted to be sure there was no spam. I did this command:


exim -bp | awk '{if ($3!="") {print "newMessage"; system("exim -Mvb " $3) } }' | awk 'BEGIN{RS="newMessage"} !/different string|I do not want|to include inside|mails/{print $0}' | less

So I ignore mails containing certain words and remain only with those strange mails.

I discovered I had sent no spam, but instead I had received undeliverable mail which mas kept frozen.

Thursday, October 17, 2013

exim commands

Delete queue:

sudo exim -bp | awk '{print $3}' | xargs exim -Mrm


extracting mailing script out of exim queue, using awk

This is a nice command I came to:

exim -bp | awk '{ if ($3 neq '') system("exim -Mvh " $3) }' | grep Script | awk '{ print $3}' | sort

It prints the scripts of my server that have send mail.

It is also useful the mail.log php directive, that logs a lot of information of every script that send mail with mail() in php.

Saturday, September 7, 2013

guide to effectively configure vsftp with chroot'ed users on Amazon EC2

VSFTP is fast and secure FTP server.

It is easy to configure. I needed it to work with Amazon EC2. Here's how.

I took this guide: http://blog.liwen.name/configure-vsftpd-on-amazon-ec2/148

Configure vsftpd on Amazon EC2

There are quite a few FTP server options available for Debian: ProFTPDPure-FTPd and wu-ftpd to name a few. Here we opt forvsftpd (very secure FTP daemon), the default FTP server included in Ubuntu, CentOS, Fedora and some other Linux distributions.

Install

apt-get update apt-get install vsftpd

Configure vsftpd

The configuration file /etc/vsftpd.conf of vsftpd is very well commented. Read it through if you want to, otherwise here are a few changes need to be made in order to get it to work with Amazon EC2. The explanation of these changes are mostly quoted from manpages of vsftpd configure package.
nano /etc/vfstpd.conf
First let’s disable anonymous logins:
anonymous_enable=NO
Enable local logins to allow local users to connect via FTP, this must be enabled for any non-anonymous login to work.
local_enable=YES
Give FTP users write permission:
write_enable=YES
Disnable PORT style data connections with port 20. It makes vsftpd run with slightly less privilege.
connect_from_port_20=NO
Restrict local users (all FTP users) in chroot jails (their home directory):
chroot_local_user=YES
To set proper permissions for files(644) and directories(755):
local_unmask=022
Specify a range of ports for vsftpd to run PASV connections
pasv_max_port=12100
pasv_min_port=12000
After setting up the port range, go to your EC2 console and open the ports specified above, also don’t forget to open the default ftp port 21.
It turns out that vsftpd advises the incomming PASV command the internal IP of EC2 instance, which FTP clients would not be able to resolve. To solve this problem, we explicitly tell vsftp to use our public IP address instead of asking the server for it. If you don’t have an Elastic IP associated with the instance, you will need to enable pasv_addr_resolve and provide your public DNS.
pasv_address=your.public.ip.address
That is all we have to do with vsftpd.conf for now. Next let’s setup our first FTP user.

Setup FTP users

To enable group-based FTP access and also make things more organised, create a dedicated FTP user group.
addgroup ftpusers
Next create our first FTP user:
useradd -d /home/web/your/root/ftp/dir/for/the/user -s /usr/sbin/nologin -g ftpusers devuser
Here we added a new user devuser with home directory /home/web/your/root/ftp/dir/for/the/user, we obviously do not want FTP users to have shell access, -s option sets user’s shell to nologinNote: don’t forget to add nologin into/etc/shells, otherwise FTP users may not be able to login via FTP clients.
echo "/usr/sbin/nologin" >> /etc/shells
Set a password for the user:
passwd devuser
To allow FTP users to read and write files in their chroot jails (home directories), we need to let FTP users take ownership of their home directories and give them proper permission.
chown -R devuser /home/web/your/root/ftp/dir/for/the/user
chmod 775 /home/web/your/root/ftp/dir/for/the/user
Create a userlist for vsftpd and add all FTP users into the list – one user per line:
touch /etc/vsftpd.userlist
nano /etc/vsftpd.userlist
The userlist file should look like this:
devuser
user2
user3
Save /etc/vsftpd.userlist, reopen /etc/vsftpd.conf and add the following lines to the end of the file:
userlist_enable=YES
userlist_file=/etc/vsftpd.userlist
If you only want to allow the users in the userlist to login and deny anyone else, you can also set:
userlist_deny=NO
Now save the file, restart the vsftpd service.
/etc/init.d/vsftpd restart
Phew, that’s it, now you have successfully configured vsftpd on Amazon EC2 instance.

Friday, August 2, 2013

common pitfalls using phonegap + jquerymobile

I fell in quite a few of these... :

(taken from: http://rip747.wordpress.com/2012/04/19/pitfalls-with-jquery-mobile-and-how-to-over-come-them/)

Pitfalls with jQuery Mobile and how to over come them

Posted in Jqueryjquery mobile by rip747 on April 19, 2012
At work I’ve been tasked with creating a mobile app using jQuery Mobile and Phonegap. Needless to say, I’ve had nothing but issues. Below are some of the pitfalls that I’ve been running into and how to over come them. This post is a work in progress and I will add to it as things come up.
1) When creating additional pages, ONLY include the `data-role=”page”`.
First, let me explain what I mean by data-role=”page”. When you create your index.html for your jQuery Mobile project, you set it up like so:
<!– index.html –>
<!DOCTYPE html>
<html>
<head>
<title>My Project</title>
<meta name=”viewport” content=”width=device-width, initial-scale=1″/>
<link rel=”stylesheet” href=”css/jquery.mobile-1.1.0.min.css” />
<script type=”text/javascript” charset=”utf-8″ src=”js/jquery-1.7.1.min.js”></script>
<script type=”text/javascript” charset=”utf-8″ src=”js/jquery.mobile-1.1.0.min.js”></script>
<script type=”text/javascript” charset=”utf-8″ src=”js/phonegap-1.4.1.js”></script>
</head>
<body>
<div data-role=”page” id=”home”>
<div data-role=”header”>
<h1>Your header</h1>
</div>
<div data-role=”content”>
<p>The content goes here</p>
</div>
<div data-role=”footer”>
<p>The footer goes here</p>
</div>
</div>
</body>
</html>
Notice that we have our opening html tag, our head section with all our stylesheets and javascript includes, our opening body tag, our page div (containing the head, content and footer of the page), our closing body tag and closing html tag. This is standard when creating a html page and something we’re all use to.
Now let’s say you want to create another page called search.html. You would think that you would need to copy the head  section, body and html tags to the search.html page, but THIS IS WRONG! All the search.html page would contain is the page div like so:
<!– search.html –>
<div data-role=”page” id=”search”>
<div data-role=”header”>
<h1>Search</h1>
</div>
<div data-role=”content”>
<p>The form used to search</p>
</div>
<div data-role=”footer”>
<p>The footer goes here</p>
</div>
</div>
The reason for this is that jQuery Mobile will inject the search.html page’s content into the index.html via ajax. Because of this, if you put the html, head, and body tags into your search.html page, they will get duplicated. This can cause all sorts of issues that I won’t even go into. Just remember that you only need the html, head and body tags on the index.html page.
2) Disable ajax caching globally
Supposedly in jQuery Mobile disabled ajax caching of pages in 1.1.0. However, I was still having issue with the pages being cached so this was making development a real pain. Luckily you can disable ajax caching altogether by doing:
<script>
$.ajaxSetup ({
// Disable caching of AJAX responses
cache: false
});
</script>
Just put that in your head section on your index.html page and you should be golden.
3) Multiple events firing
This was a HUGE pain in the ass. You normally see this problem when submitting a form from one page to another. On the calling page, you might have some javascript that is bound to an event (like pageshow) that generates some dynamic content after doing an ajax call. If you look in your Net tab in firebug, you’ll notice that the ajax call will increment with each visit. So on the first visit it fires once, the second visit it fires twice and so on. The reason for this is because, again, of the way jQuery Mobile pulls in pages via ajax. Because it will pull the page in for each visit, it will continually add the code you want to run to the event stack on each visit.
Now I’ve seen people try to solve this by placing the code in the `pageinit` event since that only fires the first time the page in pulled via ajax, but this doesn’t work when you’re having to create dynamic content based on a search string.
The solution is quite simple once you think about it, just put your code on the index.html and delegate using the on() method. so for instance, if you have page with an id of `search` and you want to run an ajax request to get the results, you would do:
$(document).on(‘pageshow’, ‘#search’, function(){
// add code to get the search results via ajax here
});
Personally I would recommend that you put all the javascript code for your app in a js file and include that on your index.html page.
4) When inserting dynamic content into the DOM you must call trigger(‘create’) in order for the framework to apply the styling.
For my app, I have a function that automatically add pagination buttons (previous and next) to the footer. In my template the footer is just defined plainly:
<div data-role=”footer”></div>
In my javascript code, I add the pagination by calling the html() method on the footer object (code is summarized):
var loc = {};
loc.self = $(this);
loc.footer = $(‘[data-role="footer"]‘, loc.self);
loc.footer.html(pagination(loc.params.page, loc.results.PAGES));
The issue was that none of the styling was taking affect on the footer and the prev and next button were showing up as just links. I found in the jQuery Mobile form that you should call page() on the object after altering it, but this really didn’t work. After some more search I found that what I really need to do was call trigger(‘create’) on the object instead. the nice thing is that you can chain this after the html() call and it still works:
loc.footer.html(pagination(loc.params.page, loc.results.PAGES)).trigger(‘create’);
5) When creating a dynamic listview and inserting content into it, you need to call listview() and then listview(‘refresh’) to reapply the styling.
This is basically the same problem as #4 only this time we’re dynamically creating a listview an inserting content into it. In my app, I use Liquid as a templating language as it’s just making life soooo much easier then concatenating javascript strings. So to create my list view, I have it defined in my Liquid template like so:
<script type=”text/liquid” id=”arrestSearchResults-markup”>
{% if RESULTS.size %}
<ul id=”arrestSearchResults-listview” data-role=”listview”>
{% for item in RESULTS %}
<li>
<a href=’{{LINK}}&jms_number={{item.JMS_NUMBER}}’ class=’arrestSearch-details’>
<img src=’{{item.PHOTO_THUMB}}’ />
<h3>{{item.LAST_NAME}}, {{item.FIRST_NAME}} {{item.MIDDLE_NAME}}</h3>
<p>{{item.JAIL}}</p>
</a>
</li>
{% endfor %}
</ul>
{% else %}
<h3>No records found</h3>
{% endif %}
</script>
The target content div is defined plainly, just like my footer is in #4:
<div data-role=”content”></div>
to compile the template and insert it into the content div, I do the following in my code (code is summarized):
var loc = {};
loc.self = $(this);
loc.markup = $(“#arrestSearchResults-markup”).html();
loc.target = $(‘[data-role="content"]‘, loc.self);
// render the markup to the listview
loc.target.html(Liquid.Template.parse(loc.markup).render(assigns));
// refresh the listview
loc.target.find(‘ul’).listview();
loc.target.find(‘ul’).listview(‘refresh’);
Basically what I’m doing is getting the template markup and the content div and putting them into a variables. I then compile the Liquid template and pass in the assigns object that contains the information to render the template.
The key to all of this is the next two lines which finds the unsorted list (‘ul’) which contains the listview a just injected into the content div and calls listview() on it, this tell jQuery Mobile to treat the ul as a listview object. I then call listview(‘refresh’) to have the framework apply the styling to it.
6) When performing validation on a form, the form will still submit.
Here is the setup. You have a form and you’re trying to perform some sort of validation on it when the form is submitted and show the visitor some errors. You tie your validation to the form’s submit event using submit() and include event.preventDefault() good measure when any error occur. However, the form still submits even though errors are through, the event.preventDefault() does prevent the form from not submitting. Heck, you even throw in `return false` hoping the form won’t submit, but it still does.
The issue is that the form is being submitting via ajax and you can’t stop the ajax submission from happening through standard means. The only thing you can do is turn off ajax and submit the form yourself. Now in older versions of the framework, you could turn off ajax for form submissions separately, however in the latest version (1.1.0) you can only turn off ajax globally by setting `ajaxEnabled` to false. This sucks as you most likely want all the ajax goodness, just not on form submission.
The way around the is to add `data-ajax=”false”` to the form:
<form id=”myform” action=”somepage.html” method=”get” data-ajax=”false”>
The will prevent the form from being submitted via ajax. Now the fun part if how in the world are you going to submit the form data to the action page using ajax so you get that nice ajax spinner thingy when you’ve turn ajax off? The answer is manually submit the form data by serializing it and appending it to form’s action attribute. Then use `$.mobile.changePage()` to submit the data via ajax. Below is a little helper function I wrote to do this:
submitForm = function(formid){
var form = $(“#” + formid);
var page = [];
page.push(form.attr(‘action’));
page.push(form.serialize());
$.mobile.changePage(page.join(‘?’));
}
To use, just call submitForm(‘your form id’) and it will handle the submission for you:
submitForm(“myform”);
7) Calling trigger(‘create’) on date-role=”header” has no effect
Though you need to call trigger(‘create’) on the data-role=”content” when adding dynamic content for it to style properly, this doesn’t hold true for data-role=”header”. The solution is to call trigger(‘pagecreate’) instead.

Wednesday, May 29, 2013

asynchronous php thread (asynchronous cake shell execution)

While cumbersome if not impossible to actually manage threads in PHP, you can always execute in background a php script.

Here's how:

While in a cakephp framework method (of course the important details are the ones in colored background):

$str = 'path/to/cakelib/cake/console/cake -working '.ROOT . DS . APP_DIR.' -app ' . ROOT . DS . APP_DIR . ' my_shell_script > /dev/null 2> '.LOGS.'my_log.log &';

exec($str, $output2, $status);




I used the ideas from this post: http://stackoverflow.com/questions/222414/asynchronous-shell-exec-in-php

This answer:
If it "doesn't care about the output", couldn't the exec to the script be called with the & to background the process?
EDIT - incorporating what @AdamTheHut commented to this post, you can add this to a call to exec:
"> /dev/null 2>/dev/null &"
That will redirect both stdio (first >) and stderr (2>) to /dev/null and run in the background.
There are other ways to do the same thing, but this is the simplest to read.

An alternative to the above double-redirect:
" &> /dev/null &"


Tuesday, May 21, 2013

mysql dump to a file with where clause

This is very useful to export records in a table to a file, to import later...


mysqldump --databases dbname -h hostname --tables tablename --where="condition='1' and condition='2'" -u username -p --no-create-db --no-create-info > myfile.dump

to import

mysql --databases dbname -h hostname --tables tablename -u username -p < myfile.dump

Wednesday, May 15, 2013

Translating month names in CakePHP (i18n)

I found this, and I think it is worth saving.

Taken from: http://www.bravo-kernel.com/2010/12/using-lc_time-with-cakephp/


USING LC_TIME WITH CAKEPHP

CakePHP supports LC_TIME translations since version 1.3 and since it took me some time to completely figure out the logic behind it I am storing my notes here as a mental reminder to self (again).
First things first… make sure you read  the i18n paragraph on LC_TIME in the Book to get some basic understanding of what we are trying to do here.

Preparing for LC_TIME

For __c(), $this->Time->format() and $this->Time->i18nFormat() to work:
  1. create a file called /app/locale/dut/LC_TIME
  2. on your local Linux workstation open /usr/share/i18n/locales/nl_NL
  3. copy everything between LC_TIME and END LC_TIME to the file created in step 1 and save that file
Note: make sure to add the escape_char and comment_char definitions to your LC_TIME file as well or your setup will not be fully functional (see this page for more info):
1
2
comment_char %
escape_char  /

strftime()

There is no need to use php’s strftime() if you consistently stick to using the CakePHP functions mentioned above. However, if you do need to get strftime() up and running you should add the following line to one of your controllers.
1
setlocale(LC_TIME, 'nl_NL.UTF8');
Please note (and understand) that this will use your server’s locales and NOT your manually created Cake LC_TIME file.

Testing your LC_TIME setup

Add the following lines to one of your views (and make sure the TimeHelper is available):
1
2
3
4
5
6
7
8
9
10
$timestamp = time();
 $timestring = $this->Time->format('Y-m-d H:i:s', $timestamp);
 $months = __c('mon', 5 ,true);
 
 pr("Timestamp = $timestamp");
 pr("Timestring = $timestring");
 pr("strftime() translated = " . strftime("%A %e %B %Y", strtotime($timestring)));
 pr("i18nFormat  translated = " . $this->Time->i18nFormat($timestring, "%A %e %B %Y"));
 pr("Time::format translated = " . $this->Time->format($timestring, '%A %e %B %Y'));
 pr($months);
If your setup is fully operational it should display the following LC_TIME translations:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Timestamp = 1292417839
Timestring = 2010-12-15 13:57:19
strftime() translated = Wednesday 15 December 2010
i18nFormat  translated = woensdag 15 december 2010
Time::format translated = woensdag 15 december 2010
Array
(
    [0] => januari
    [1] => februari
    [2] => maart
    [3] => april
    [4] => mei
    [5] => juni
    [6] => juli
    [7] => augustus
    [8] => september
    [9] => oktober
    [10] => november
    [11] => december
)
Note: you might have spotted that the strftime() translation is not translated. This is intentional since I always stick to Cake methods. See the paragraph on strftime() if you do need to use that function.
Enjoy your time translations ;)

cancel script completely on ctrl-c

I found this question interesting: basically how to cancel completely a script and all child processes : You do this by creating a subro...