Debugging PHP applications with PhpStorm
Friday, February 21. 2014
Anybody who has reached a certain level of skill in software development eventually will start thinking about how to make the job easier or be more productive during the day. With computers and programming, the tools you're using is a good starting point.
I've written code with a number of text editors (including vim and notepad), some of them really bad and some of them more suited to the task at hand. IMHO one of the best editors for a software developer is Microsoft's Visual Studio. Since many of my projects involve PHP, that is not the optimal editor for that. For years I used Notepad++, but then I got a recommendation for JetBrains PhpStorm. Since that it has become my weapon of choice for PHP.
For example, debugging the application is very simple operation. PhpStorm has all the client-stuff built into it, the only thing that is needed is Xdebug-extension to the server end. In my case, my web server is on a Linux and I typically work on a Windows 7. The debugging is done over my LAN. There is a very good article on PhpStorm author's web site with name Zero-configuration Web Application Debugging with Xdebug and PhpStorm. I don't think the zero-configuration part is true, but it is almost zero anyway. 
On the server end I have a PHP-fpm running, it has
php_admin_flag[xdebug.remote_enable] = on
php_admin_value[xdebug.remote_host] = my_machine_name
php_admin_value[xdebug.remote_port] = 9000
In the worker configuration. That enables the server side to initiate a DBGP-connection to the given client (PhpStorm) and start sending data there. There reference for all the configuration directives can be found from Xdebug's docs.
On the PhpStorm end all you have to do, is enable the listener:
Also I had to drill a hole into my Windows 7 firewall, to allow my web server to connect into TCP/9000. If you're running a single machine setup, that won't be necessary.
Any settings for the setup can be found in the PhpStorm project:
But I think they're ok as a default. At least I didn't change anything there. If your debugger does not work, you fumbled the settings or your listener is not enabled. On Windows, run this on command prompt to confirm:
PS J:\> netstat -ano
Active Connections
Proto Local Address Foreign Address State PID
TCP 0.0.0.0:9000 0.0.0.0:0 LISTENING ?
It will return a long list of things, but one of the entries must be similar to that.
To make the choice between debugging and not debugging can be made really easy. There are a number of ways of starting a debug session, but for Firefox users there is an excellent add-on The easiest Xdebug. It will add an icon to your add-on bar:
If the bug is green, it will send a special cookie to the server during a page load request:
Cookie: XDEBUG_SESSION=netbeans-xdebug
That will initiate the debugging session for you and PhpStorm will stop on a breakpoint if one is set.
Switching to PhpStorm vastly improved my productivity. I would imagaine, that it will do that for anybody.
Anybody who has reached a certain level of skill in software development eventually will start thinking about how to make the job easier or be more productive during the day. With computers and programming, the tools you're using is a good starting point.
I've written code with a number of text editors (including vim and notepad), some of them really bad and some of them more suited to the task at hand. IMHO one of the best editors for a software developer is Microsoft's Visual Studio. Since many of my projects involve PHP, that is not the optimal editor for that. For years I used Notepad++, but then I got a recommendation for JetBrains PhpStorm. Since that it has become my weapon of choice for PHP.
For example, debugging the application is very simple operation. PhpStorm has all the client-stuff built into it, the only thing that is needed is Xdebug-extension to the server end. In my case, my web server is on a Linux and I typically work on a Windows 7. The debugging is done over my LAN. There is a very good article on PhpStorm author's web site with name Zero-configuration Web Application Debugging with Xdebug and PhpStorm. I don't think the zero-configuration part is true, but it is almost zero anyway.
On the server end I have a PHP-fpm running, it has
php_admin_flag[xdebug.remote_enable] = on
php_admin_value[xdebug.remote_host] = my_machine_name
php_admin_value[xdebug.remote_port] = 9000
In the worker configuration. That enables the server side to initiate a DBGP-connection to the given client (PhpStorm) and start sending data there. There reference for all the configuration directives can be found from Xdebug's docs.
On the PhpStorm end all you have to do, is enable the listener:
Also I had to drill a hole into my Windows 7 firewall, to allow my web server to connect into TCP/9000. If you're running a single machine setup, that won't be necessary.
Any settings for the setup can be found in the PhpStorm project:
But I think they're ok as a default. At least I didn't change anything there. If your debugger does not work, you fumbled the settings or your listener is not enabled. On Windows, run this on command prompt to confirm:
PS J:\> netstat -ano
Active Connections
Proto Local Address Foreign Address State PID
TCP 0.0.0.0:9000 0.0.0.0:0 LISTENING ?
It will return a long list of things, but one of the entries must be similar to that.
To make the choice between debugging and not debugging can be made really easy. There are a number of ways of starting a debug session, but for Firefox users there is an excellent add-on The easiest Xdebug. It will add an icon to your add-on bar:
If the bug is green, it will send a special cookie to the server during a page load request:
Cookie: XDEBUG_SESSION=netbeans-xdebug
That will initiate the debugging session for you and PhpStorm will stop on a breakpoint if one is set.
Switching to PhpStorm vastly improved my productivity. I would imagaine, that it will do that for anybody.
Zend Framework 2: Abusing ZeDb to manage multiple connections
Tuesday, February 18. 2014
Zend Framework 2: Disable layout
Thursday, February 13. 2014
Managing PostgreSQL 8.x permissions to limit application user's access
Wednesday, February 5. 2014
I was working with a legacy project with PostgreSQL 8 installation. A typical software developer simply does not care about DBA enough to think more than once about the permissions setup. The thinking is that for the purpose of writing lines of working code which executes really nice SQL-queries a user with lots of power in its sleeves is a good thing. This is something I bump into a lot. It would be a nice eye-opener if every coder would had to investigate a server which has been cracked into once or twice in the early programming career. I'm sure that would improve the quality of code and improve security thinking.
Anyway, the logic for ignoring security is ok for a development box, given the scenario that it is pretty much inaccessible outside the development team. When going to production things always get more complicated. I have witnessed production boxes which are running applications that have been configured to access DB with Admin-permissions. That happens in an environment where any decent programmer/DBA can spot out a number of other ignored things. Thinking about security is both far above the pay-grade and the skill envelope your regular coder possesses.
In an attempt to do things the-right-way(tm), it is a really good idea to create a specific user for accessing the DB. Even better idea is to limit the permissions so, that application user cannot run the classic "; DROP TABLE users; -- " because lacking the permission to drop tables. We still remember Exploits of a Mom, right?
Image courtesy of xkcd.com.
Back to reality... I was on a production PostgreSQL and evaluated the situation. Database has owner of postgres, schema public had owner of postgres, but all the tables, sequences and views where owned by the application user. So any exploit would allow the application user to drop all tables. Not cool, huh! 
To solve this three things are needed: first, owner of the entire schema must be postgres. Second, the application user needs only to have enough permission for CRUD-operations, nothing more. And third, the schema must not allow users to create new items on it. As default everybody can create new tables and sequences, but if somebody really pops your box and can run anything on your DB, creating new items (besides temporary tables) is not a good thing.
On a PostgreSQL 8 something of a trickery is needed. Version 9.0 introduced us the "GRANT ... ALL TABLES IN SCHEMA", but I didn't have that at my disposal. To get around the entire thing I created two SQL-queries which were crafted to output SQL-queries. I could simply copy/paste the output and run it in pgAdmin III query-window. Nice!
The first query to gather all tables and sequences and change the owner to postgres:
SELECT 'ALTER TABLE ' || table_schema || '.' || table_name ||' OWNER TO postgres;'
FROM information_schema.tables
WHERE
table_type = 'BASE TABLE' and
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'ALTER SEQUENCE ' || sequence_schema || '.' || sequence_name ||' OWNER TO postgres;'
FROM information_schema.sequences
WHERE
sequence_schema NOT IN ('pg_catalog', 'information_schema')
It will output something like this:
ALTER TABLE public.phones OWNER TO postgres;
ALTER SEQUENCE public.user_id_seq OWNER TO postgres;
I ran those, and owner was changed.
NOTE: that effectively locked the application user out of DB completely.
So it was time to restore access. This is the query to gather information about all tables, views, sequences and functions:
SELECT 'GRANT ALL ON ' || table_schema || '.' || table_name ||' TO my_group;'
FROM information_schema.tables
WHERE
table_type = 'BASE TABLE' and
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON ' || table_schema || '.' || table_name ||' TO my_group;'
FROM information_schema.views
WHERE
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON SEQUENCE ' || sequence_schema || '.' || sequence_name ||' TO my_group;'
FROM information_schema.sequences
WHERE
sequence_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON FUNCTION ' || nspname || '.' || proname || '(' || pg_get_function_arguments(p.oid) || ') TO my_group;'
FROM pg_catalog.pg_proc p
INNER JOIN pg_catalog.pg_namespace n ON pronamespace = n.oid
WHERE
nspname = 'public'
It will output something like this:
GRANT ALL ON public.phones TO my_user;
GRANT ALL ON SEQUENCE public.user_id_seq TO my_user;
NOTE: you need to find/replace my_user to something that fits your needs.
Now the application was again running smoothly, but with reduced permission in effect. The problem with all this is that TRUNCATE-clause (or DELETE FROM -tablename-) are still working. To get the maximum out of enhanced security, some classification of data would be needed. But the client wasn't ready to do that (yet).
The third thing is to limit schema permissions so that only usage is allowed for the general public:
REVOKE ALL ON SCHEMA public FROM public;
GRANT USAGE ON SCHEMA public TO public;
Now only postgres can create new things there.
All there is to do at this point is to test the appliation. There should be errors for DB-access if something went wrong.
I was working with a legacy project with PostgreSQL 8 installation. A typical software developer simply does not care about DBA enough to think more than once about the permissions setup. The thinking is that for the purpose of writing lines of working code which executes really nice SQL-queries a user with lots of power in its sleeves is a good thing. This is something I bump into a lot. It would be a nice eye-opener if every coder would had to investigate a server which has been cracked into once or twice in the early programming career. I'm sure that would improve the quality of code and improve security thinking.
Anyway, the logic for ignoring security is ok for a development box, given the scenario that it is pretty much inaccessible outside the development team. When going to production things always get more complicated. I have witnessed production boxes which are running applications that have been configured to access DB with Admin-permissions. That happens in an environment where any decent programmer/DBA can spot out a number of other ignored things. Thinking about security is both far above the pay-grade and the skill envelope your regular coder possesses.
In an attempt to do things the-right-way(tm), it is a really good idea to create a specific user for accessing the DB. Even better idea is to limit the permissions so, that application user cannot run the classic "; DROP TABLE users; -- " because lacking the permission to drop tables. We still remember Exploits of a Mom, right?
Image courtesy of xkcd.com.
Back to reality... I was on a production PostgreSQL and evaluated the situation. Database has owner of postgres, schema public had owner of postgres, but all the tables, sequences and views where owned by the application user. So any exploit would allow the application user to drop all tables. Not cool, huh!
To solve this three things are needed: first, owner of the entire schema must be postgres. Second, the application user needs only to have enough permission for CRUD-operations, nothing more. And third, the schema must not allow users to create new items on it. As default everybody can create new tables and sequences, but if somebody really pops your box and can run anything on your DB, creating new items (besides temporary tables) is not a good thing.
On a PostgreSQL 8 something of a trickery is needed. Version 9.0 introduced us the "GRANT ... ALL TABLES IN SCHEMA", but I didn't have that at my disposal. To get around the entire thing I created two SQL-queries which were crafted to output SQL-queries. I could simply copy/paste the output and run it in pgAdmin III query-window. Nice!
The first query to gather all tables and sequences and change the owner to postgres:
SELECT 'ALTER TABLE ' || table_schema || '.' || table_name ||' OWNER TO postgres;'
FROM information_schema.tables
WHERE
table_type = 'BASE TABLE' and
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'ALTER SEQUENCE ' || sequence_schema || '.' || sequence_name ||' OWNER TO postgres;'
FROM information_schema.sequences
WHERE
sequence_schema NOT IN ('pg_catalog', 'information_schema')
It will output something like this:
ALTER TABLE public.phones OWNER TO postgres;
ALTER SEQUENCE public.user_id_seq OWNER TO postgres;
I ran those, and owner was changed.
NOTE: that effectively locked the application user out of DB completely.
So it was time to restore access. This is the query to gather information about all tables, views, sequences and functions:
SELECT 'GRANT ALL ON ' || table_schema || '.' || table_name ||' TO my_group;'
FROM information_schema.tables
WHERE
table_type = 'BASE TABLE' and
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON ' || table_schema || '.' || table_name ||' TO my_group;'
FROM information_schema.views
WHERE
table_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON SEQUENCE ' || sequence_schema || '.' || sequence_name ||' TO my_group;'
FROM information_schema.sequences
WHERE
sequence_schema NOT IN ('pg_catalog', 'information_schema')
UNION
SELECT 'GRANT ALL ON FUNCTION ' || nspname || '.' || proname || '(' || pg_get_function_arguments(p.oid) || ') TO my_group;'
FROM pg_catalog.pg_proc p
INNER JOIN pg_catalog.pg_namespace n ON pronamespace = n.oid
WHERE
nspname = 'public'
It will output something like this:
GRANT ALL ON public.phones TO my_user;
GRANT ALL ON SEQUENCE public.user_id_seq TO my_user;
NOTE: you need to find/replace my_user to something that fits your needs.
Now the application was again running smoothly, but with reduced permission in effect. The problem with all this is that TRUNCATE-clause (or DELETE FROM -tablename-) are still working. To get the maximum out of enhanced security, some classification of data would be needed. But the client wasn't ready to do that (yet).
The third thing is to limit schema permissions so that only usage is allowed for the general public:
REVOKE ALL ON SCHEMA public FROM public;
GRANT USAGE ON SCHEMA public TO public;
Now only postgres can create new things there.
All there is to do at this point is to test the appliation. There should be errors for DB-access if something went wrong.
Zend Framework 2: Touching headLink() twice on layout template
Friday, January 17. 2014
Zend Framework 2: preDispatch(), returning properly without executing action
Thursday, January 16. 2014
Getting ZF2 to do preDispatch() and postDispatch() like ZF1 had is widely known and documented. In your controller, add this:
protected function attachDefaultListeners()
{
parent::attachDefaultListeners();
$event_mgr = $this->getEventManager();
$event_mgr->attach('dispatch', array($this, 'preDispatch'), 100);
$event_mgr->attach('dispatch', array($this, 'postDispatch'), -100);
}
Two simple listeners are attached with proper priorities to trigger before and after the action.
However, to go somewhere else before the action is executed adds some complexity, as one can expect. In preDispatch() you can do one of two suggested things. A redirect:
// Do a HTTP/302 redirect
return $this->redirect()->toRoute(
'application', array('controller' => 'index', 'action' => 'index'
));
My issue here is, that it literally does a HTTP/302 redirect in your browser. Another problem is, that it still executes the action it was targeted to. It renders the view, runs all the listeners, does all the plugins and helpers as it started to do. It just redirects after all that. I don't want my user to do a redirect or to run all the bells and whistles including the action. Why cannot I simply return something else instead, like ZF1 could be programmed to do. On my top-1 list is to execute an action from another controller?
So, the another option to do is to simply call it quits right in the middle of preDispatch():
$url = $event->getRouter()
->assemble(
array('action' => 'index'),
array('name' => 'frontend')
);
$response = $event->getResponse();
$response->getHeaders()->addHeaderLine('Location', $url);
$response->setStatusCode(302);
$response->sendHeaders();
exit();
That's pretty much the same as previous, but uglier. exit()!! Really? In Zend Framework?! I'd rather keep the wheels rolling and machine turning like it normally would do until it has done all the dirty deeds it wants to do. Poking around The Net reveals, that nobody is really offering anything else. Apparently everybody simply are doing a copy/paste from the same sites I found.
This is what I offer. Discard the current operation, start a new one and return that! A lot better alternative.
Example 1, return JSON-data:
$event->stopPropagation(true);
// Skip executing the action requested. Return this instead.
$result = new JsonModel(array(
'success' => false,
'loginrequired' => true
));
$result->setTerminal(true);
$event->setResponse(new Response());
$event->setViewModel($result);
The key is in the setResponse()-call.
Example 2, call another action:
$event->stopPropagation(true);
// Skip executing the action requested.
// Execute anotherController::errorAction() instead.
$event->setResponse(new Response());
$result = $this->forward()->dispatch('Another', array(
'action' => 'error'
));
$result->setTerminal(true);
$event->setViewModel($result);
Hope this helps somebody else trying to do a ZF1 to ZF2 transition. In the end, there is only one thing similar between them. Their name has Zend Framwork in it.
Getting ZF2 to do preDispatch() and postDispatch() like ZF1 had is widely known and documented. In your controller, add this:
protected function attachDefaultListeners()
{
parent::attachDefaultListeners();
$event_mgr = $this->getEventManager();
$event_mgr->attach('dispatch', array($this, 'preDispatch'), 100);
$event_mgr->attach('dispatch', array($this, 'postDispatch'), -100);
}
Two simple listeners are attached with proper priorities to trigger before and after the action.
However, to go somewhere else before the action is executed adds some complexity, as one can expect. In preDispatch() you can do one of two suggested things. A redirect:
// Do a HTTP/302 redirect
return $this->redirect()->toRoute(
'application', array('controller' => 'index', 'action' => 'index'
));
My issue here is, that it literally does a HTTP/302 redirect in your browser. Another problem is, that it still executes the action it was targeted to. It renders the view, runs all the listeners, does all the plugins and helpers as it started to do. It just redirects after all that. I don't want my user to do a redirect or to run all the bells and whistles including the action. Why cannot I simply return something else instead, like ZF1 could be programmed to do. On my top-1 list is to execute an action from another controller?
So, the another option to do is to simply call it quits right in the middle of preDispatch():
$url = $event->getRouter()
->assemble(
array('action' => 'index'),
array('name' => 'frontend')
);
$response = $event->getResponse();
$response->getHeaders()->addHeaderLine('Location', $url);
$response->setStatusCode(302);
$response->sendHeaders();
exit();
That's pretty much the same as previous, but uglier. exit()!! Really? In Zend Framework?! I'd rather keep the wheels rolling and machine turning like it normally would do until it has done all the dirty deeds it wants to do. Poking around The Net reveals, that nobody is really offering anything else. Apparently everybody simply are doing a copy/paste from the same sites I found.
This is what I offer. Discard the current operation, start a new one and return that! A lot better alternative.
Example 1, return JSON-data:
$event->stopPropagation(true);
// Skip executing the action requested. Return this instead.
$result = new JsonModel(array(
'success' => false,
'loginrequired' => true
));
$result->setTerminal(true);
$event->setResponse(new Response());
$event->setViewModel($result);
The key is in the setResponse()-call.
Example 2, call another action:
$event->stopPropagation(true);
// Skip executing the action requested.
// Execute anotherController::errorAction() instead.
$event->setResponse(new Response());
$result = $this->forward()->dispatch('Another', array(
'action' => 'error'
));
$result->setTerminal(true);
$event->setViewModel($result);
Hope this helps somebody else trying to do a ZF1 to ZF2 transition. In the end, there is only one thing similar between them. Their name has Zend Framwork in it.
git and HTTPS (fatal: HTTP request failed)
Friday, January 10. 2014
Barcode weirdness: Exactly the same, but different
Tuesday, December 17. 2013
I was transferring an existing production web-application to a new server. There was a slight gap in the operating system version and I ended up refactoring lot of the code to compensate newer libraries in use. One of the issues I bumped into and spent a while fixing was barcode printing. It was a pretty standard Code 39 barcode for the shipping manifest.
The old code was using a commercial proprietary TrueType font file. I don't know when or where it was purchased by my client, but it was in production use in the old system. Looks like such a fonts sell for $99. Anyway, the new library had barcode printing in it, so I chose to use it.
Mistake. It does somewhat work. It prints a perfectly valid barcode. Example:
The issue is, that it does not read. I had a couple of readers to test with, and even the really expensive laser-one didn't register the code very well. So I had to re-write the printing code to simulate the old functionality. The idea is to read the TTF-font, convert it into some internal form and output a ready-to-go file into a data directory, so that conversion is needed only once. After that the library can read the font and produce nice looking graphics including barcodes. Here is an example of the TTF-style:
The difference is amazing! Any reader reads that easily. Both barcodes contain the same data in them, but they look so much different. Wow!
Perhaps one day a barcode wizard will explain me the difference between them.
I was transferring an existing production web-application to a new server. There was a slight gap in the operating system version and I ended up refactoring lot of the code to compensate newer libraries in use. One of the issues I bumped into and spent a while fixing was barcode printing. It was a pretty standard Code 39 barcode for the shipping manifest.
The old code was using a commercial proprietary TrueType font file. I don't know when or where it was purchased by my client, but it was in production use in the old system. Looks like such a fonts sell for $99. Anyway, the new library had barcode printing in it, so I chose to use it.
Mistake. It does somewhat work. It prints a perfectly valid barcode. Example:
The issue is, that it does not read. I had a couple of readers to test with, and even the really expensive laser-one didn't register the code very well. So I had to re-write the printing code to simulate the old functionality. The idea is to read the TTF-font, convert it into some internal form and output a ready-to-go file into a data directory, so that conversion is needed only once. After that the library can read the font and produce nice looking graphics including barcodes. Here is an example of the TTF-style:
The difference is amazing! Any reader reads that easily. Both barcodes contain the same data in them, but they look so much different. Wow!
Perhaps one day a barcode wizard will explain me the difference between them.
Why cloud platforms exist - Benchmarking Windows Azure
Tuesday, October 8. 2013
I got permission to publish a grayed out version of a project I was contracted to do this summer. Since the customer paid big bucks for it, you're not going to see all the details. I'm sorry to act as a greedy idiot, but you have to hire me to do something similar to see your results.
The subject of my project is something that personally is intriguing to me: how much better does a cloud-native application perform when compared to a traditional LAMP-setup. I chose the cloud platform to be Windows Azure, since I know that one best.
The Setup
There was a pretty regular web-application for performing couple of specific tasks. Exactly the same sample data was populated to Azure SQL for IaaS-test and Azure Table Storage for PaaS -test. People who complain about using Azure SQL can imagine a faster setup being used on a virtual machine and expect the thing to perform faster.
To simulate a real web application, memory cache was used. Memcache for IaaS and Azure Cache for PaaS. On both occasions using memory cache pushes the performance of the application further as there is no need to do so much expensive I/O.
Results
In the Excel-charts there are number of simulated users at the horizontal axis. There are two vertical axis used for different items.
Following items can be read from the Excel-charts:
- Absolute number of pages served for giving measurement point (right axis)
- Absolute number of pages returned, which returned erroneous output (right axis)
- Percentage of HTTP-errors: a status code which we interpret as an error was returned (left axis)
- Percentage of total errors: HTTP errors + requests which did not return a status code (left axis)
- Number successful pages returned per second (left axis)
Results: IaaS
I took a ready-made CentOS Linux-image bundled with Nginx/PHP-FPM -pair and lured it to work under Azure and connect to ready populated data from Azure SQL. Here are the test runs from two and three medium instances.
Adding a machine to the service absolutely helps. With two instances, the application chokes completely at the end of test load. Added machine also makes the application perform much faster, there is a clear improvement on page load speed.
Results: PaaS
Exactly the same functionality was implemented with .Net / C#.
Here are the results:
Astonishing! Page load speed is so much higher on similar user loads, also no errors can be produced. I pushed the envelope with 40 times the users, but couldn't be sure if it was about test setup (which I definitely saturated) or Azure's capacity fluctuating under heavy load. The test with small role was also very satisfactory, it beats the crap out of running two medium instances on IaaS!
Conclusion
I have to state the obvious: PaaS-application performs much better. I just couldn't belive that it was impossible to get exact measurement from the point where the application chokes on PaaS.
I got permission to publish a grayed out version of a project I was contracted to do this summer. Since the customer paid big bucks for it, you're not going to see all the details. I'm sorry to act as a greedy idiot, but you have to hire me to do something similar to see your results.
The subject of my project is something that personally is intriguing to me: how much better does a cloud-native application perform when compared to a traditional LAMP-setup. I chose the cloud platform to be Windows Azure, since I know that one best.
The Setup
There was a pretty regular web-application for performing couple of specific tasks. Exactly the same sample data was populated to Azure SQL for IaaS-test and Azure Table Storage for PaaS -test. People who complain about using Azure SQL can imagine a faster setup being used on a virtual machine and expect the thing to perform faster.
To simulate a real web application, memory cache was used. Memcache for IaaS and Azure Cache for PaaS. On both occasions using memory cache pushes the performance of the application further as there is no need to do so much expensive I/O.
Results
In the Excel-charts there are number of simulated users at the horizontal axis. There are two vertical axis used for different items.
Following items can be read from the Excel-charts:
- Absolute number of pages served for giving measurement point (right axis)
- Absolute number of pages returned, which returned erroneous output (right axis)
- Percentage of HTTP-errors: a status code which we interpret as an error was returned (left axis)
- Percentage of total errors: HTTP errors + requests which did not return a status code (left axis)
- Number successful pages returned per second (left axis)
Results: IaaS
I took a ready-made CentOS Linux-image bundled with Nginx/PHP-FPM -pair and lured it to work under Azure and connect to ready populated data from Azure SQL. Here are the test runs from two and three medium instances.
Adding a machine to the service absolutely helps. With two instances, the application chokes completely at the end of test load. Added machine also makes the application perform much faster, there is a clear improvement on page load speed.
Results: PaaS
Exactly the same functionality was implemented with .Net / C#.
Here are the results:
Astonishing! Page load speed is so much higher on similar user loads, also no errors can be produced. I pushed the envelope with 40 times the users, but couldn't be sure if it was about test setup (which I definitely saturated) or Azure's capacity fluctuating under heavy load. The test with small role was also very satisfactory, it beats the crap out of running two medium instances on IaaS!
Conclusion
I have to state the obvious: PaaS-application performs much better. I just couldn't belive that it was impossible to get exact measurement from the point where the application chokes on PaaS.
Why Azure PaaS billing cannot be stopped? - revisit
Monday, October 7. 2013
In my earlier entry about Azure PaaS billing, I was complaining about how to stop the billing.
This time I managed to do it. The solution was simple: delete the deployments, but leave the cloud service intact. Then Azure stops reserving any (stopped) compute units for the cloud service. Like this:
Here is the proof:
Zero billing. Nice!
In my earlier entry about Azure PaaS billing, I was complaining about how to stop the billing.
This time I managed to do it. The solution was simple: delete the deployments, but leave the cloud service intact. Then Azure stops reserving any (stopped) compute units for the cloud service. Like this:
Here is the proof:
Zero billing. Nice!
Why Azure PaaS billing cannot be stopped?
Tuesday, October 1. 2013
In Windows Azure stopping an IaaS virtual machine stops the billing, there is no need to delete the stopped instance. When you stop a PaaS cloud service, following happens:
Based on billing:
This is really true. On 26th and 27th I had a cloud service running on Azure, but I stopped it. On 28th and 29th there is billing for a service that has been stopped, and for which I got the warning about. I don't know why on 30th there is one core missing from the billing. Discount, perhaps?
My bottom line is:
Why? What possible idea could be, that your PaaS cloud service needs to be deleted in order to stop billing? Come on Microsoft! Equal rules for both cloud services!
In Windows Azure stopping an IaaS virtual machine stops the billing, there is no need to delete the stopped instance. When you stop a PaaS cloud service, following happens:
Based on billing:
This is really true. On 26th and 27th I had a cloud service running on Azure, but I stopped it. On 28th and 29th there is billing for a service that has been stopped, and for which I got the warning about. I don't know why on 30th there is one core missing from the billing. Discount, perhaps?
My bottom line is:
Why? What possible idea could be, that your PaaS cloud service needs to be deleted in order to stop billing? Come on Microsoft! Equal rules for both cloud services!
Migrating data from SQL into Windows Azure Table Storage
Monday, September 16. 2013
The error messages when Azure Table Storage data insert fails are far from being descriptive.
This is the complete list of supported datatypes (or Property Types as they call them):
- Binary: An array of bytes up to 64 KB in size.
- Bool: A Boolean value.
- DateTime: A 64-bit value expressed as UTC time. The supported range of values is 1/1/1601 to 12/31/9999.
- Double: A 64-bit floating point value.
- GUID: A 128-bit globally unique identifier.
- Int: A 32-bit integer.
- Int64: A 64-bit integer.
- String: A UTF-16-encoded value. String values can be up to 64 KB in size.
Really. Nothing more. You just have to get along with that one!
The list is taken from Windows Azure Table Storage and Windows Azure SQL Database - Compared and Contrasted.
Things you fail to notice:
- .Net DateTime Structure as range of 00:00:00 (midnight), January 1, 0001 Anno Domini (Common Era) through 11:59:59 P.M., December 31, 9999 A.D. (C.E.) in the Gregorian calendar. Not from January 1, 1601 AD.
- That shouldn't be an issue. My app had problems and it had recorded dates into year 201. This was a really nice way of finding that out.
- In intergers, there are no unsigned versions.
- In decimal numbers, there is no decimal, a 128-bit floating point number. You have to settle with Double, a IEC 60559:1989 (IEEE 754) compliant version.
- There is no reasonable way of storing money-type data which needs an exact number, no floating point conversions.
- The string really is UTF-16, a two byte -version. It stores up to 32768 characters.
- Which is Not much when compared to TEXT or varchar(max) which range from 2 GiB to anything you have
Hopefully this list helps somebody. I spent a nice while finding all these out.
The error messages when Azure Table Storage data insert fails are far from being descriptive.
This is the complete list of supported datatypes (or Property Types as they call them):
- Binary: An array of bytes up to 64 KB in size.
- Bool: A Boolean value.
- DateTime: A 64-bit value expressed as UTC time. The supported range of values is 1/1/1601 to 12/31/9999.
- Double: A 64-bit floating point value.
- GUID: A 128-bit globally unique identifier.
- Int: A 32-bit integer.
- Int64: A 64-bit integer.
- String: A UTF-16-encoded value. String values can be up to 64 KB in size.
Really. Nothing more. You just have to get along with that one!
The list is taken from Windows Azure Table Storage and Windows Azure SQL Database - Compared and Contrasted.
Things you fail to notice:
- .Net DateTime Structure as range of 00:00:00 (midnight), January 1, 0001 Anno Domini (Common Era) through 11:59:59 P.M., December 31, 9999 A.D. (C.E.) in the Gregorian calendar. Not from January 1, 1601 AD.
- That shouldn't be an issue. My app had problems and it had recorded dates into year 201. This was a really nice way of finding that out.
- In intergers, there are no unsigned versions.
- In decimal numbers, there is no decimal, a 128-bit floating point number. You have to settle with Double, a IEC 60559:1989 (IEEE 754) compliant version.
- There is no reasonable way of storing money-type data which needs an exact number, no floating point conversions.
- The string really is UTF-16, a two byte -version. It stores up to 32768 characters.
- Which is Not much when compared to TEXT or varchar(max) which range from 2 GiB to anything you have
Hopefully this list helps somebody. I spent a nice while finding all these out.
Using PHP, Zend Framework, PDO and FreeTDS in Windows Azure
Wednesday, September 4. 2013
Earlier I wrote about IPv6-connectivity with MS SQL server into Linux / PHP with FreeTDS.
This time my quest with FreeTDS continued, I put together the minimal possible CentOS 6.4 Linux with enough parts to produce a Nginx / PHP-FPM / Windows Azure SQL Database -based web application. The acronym could be not LAMP, but NPFWASD. No idea how to pronounce "npf-wasd", though.
I packaged a Hyper-V -based Linux .vhd into Azure virtual machine IaaS-image and created couple of load-balanced HTTP-ports into it. The problem was to lure PHP's PDO to connect into Azure SQL via FreeTDS dblib. I spent a good while banging my head and kicking it, before it stopped resisting and started to obey my commands.
Everything would have gone much better, if only I had the proper version of FreeTDS installed into the Linux. When I realized that the TDS-protocol version is hyper-important in Azure SQL, I realised that my FreeTDS-version was not the one it was supposed to be. My own-package would have been the correct one (see the earlier post). My tsql -C says:
Compile-time settings (established with the "configure" script)
Version: freetds v0.92.dev.20130721
freetds.conf directory: /etc
MS db-lib source compatibility: yes
Sybase binary compatibility: yes
Thread safety: yes
iconv library: yes
TDS version: 7.1
iODBC: no
unixodbc: yes
SSPI "trusted" logins: no
Kerberos: yes
The default TDS version of 7.1 is really, really important there. With that I can do:
tsql -H -my-designated-instance-in-Azure-.database.windows.net \
-p 1433 \
-U -the-application-SQL-user-without-admin-rights- \
-D -my-own-database-in-the-SQL-box-
It simply works, displays the prompt and everything works as it should be. In my Zend Framework application configuration I say:
resources.db.adapter = "Pdo_Mssql"
resources.db.params.host = "-my-designated-instance-in-Azure-.database.windows.net"
resources.db.params.dbname = "-my-own-database-in-the-SQL-box-"
resources.db.params.username = "-the-application-SQL-user-without-admin-rights-"
resources.db.params.password = "-oh-the-top-secret-passwrod-"
resources.db.params.version = "7.1"
resources.db.params.charset = "utf8"
resources.db.params.pdoType = "dblib"
No issues there. Everything works.
I received couple of comments from other people when I announced that I would try such a feat. It appeared that most people are running their own SQL-instances of various kinds because of performance reasons. The Azure SQL -service is definitely not the fastest there is. But what if you're not in a hurry. The service is there, easily available, cheap and functional, even from Linux/PHP.
Earlier I wrote about IPv6-connectivity with MS SQL server into Linux / PHP with FreeTDS.
This time my quest with FreeTDS continued, I put together the minimal possible CentOS 6.4 Linux with enough parts to produce a Nginx / PHP-FPM / Windows Azure SQL Database -based web application. The acronym could be not LAMP, but NPFWASD. No idea how to pronounce "npf-wasd", though.
I packaged a Hyper-V -based Linux .vhd into Azure virtual machine IaaS-image and created couple of load-balanced HTTP-ports into it. The problem was to lure PHP's PDO to connect into Azure SQL via FreeTDS dblib. I spent a good while banging my head and kicking it, before it stopped resisting and started to obey my commands.
Everything would have gone much better, if only I had the proper version of FreeTDS installed into the Linux. When I realized that the TDS-protocol version is hyper-important in Azure SQL, I realised that my FreeTDS-version was not the one it was supposed to be. My own-package would have been the correct one (see the earlier post). My tsql -C says:
Compile-time settings (established with the "configure" script)
Version: freetds v0.92.dev.20130721
freetds.conf directory: /etc
MS db-lib source compatibility: yes
Sybase binary compatibility: yes
Thread safety: yes
iconv library: yes
TDS version: 7.1
iODBC: no
unixodbc: yes
SSPI "trusted" logins: no
Kerberos: yes
The default TDS version of 7.1 is really, really important there. With that I can do:
tsql -H -my-designated-instance-in-Azure-.database.windows.net \
-p 1433 \
-U -the-application-SQL-user-without-admin-rights- \
-D -my-own-database-in-the-SQL-box-
It simply works, displays the prompt and everything works as it should be. In my Zend Framework application configuration I say:
resources.db.adapter = "Pdo_Mssql"
resources.db.params.host = "-my-designated-instance-in-Azure-.database.windows.net"
resources.db.params.dbname = "-my-own-database-in-the-SQL-box-"
resources.db.params.username = "-the-application-SQL-user-without-admin-rights-"
resources.db.params.password = "-oh-the-top-secret-passwrod-"
resources.db.params.version = "7.1"
resources.db.params.charset = "utf8"
resources.db.params.pdoType = "dblib"
No issues there. Everything works.
I received couple of comments from other people when I announced that I would try such a feat. It appeared that most people are running their own SQL-instances of various kinds because of performance reasons. The Azure SQL -service is definitely not the fastest there is. But what if you're not in a hurry. The service is there, easily available, cheap and functional, even from Linux/PHP.
What programming languages to learn?
Monday, August 26. 2013
This is a classic question which I get to answer a lot. N00bs know the answer, but somebody outside the IT-business might ask something like that. This is also quite a popular question among young people trying to figure out if programming would be for them.
Anyway, here 5 Programming Languages Everyone Should Know from two people who actually have created some of the most popular languages currently used.
Nobody should call themselves a professional if they knew only one language.
- Bjarne Stroustrup
Larry Wall
See his interview: http://youtu.be/LR8fQiskYII
His list:
- JavaScript
- Java
- Haskell
- C
- Perl
Perl is not a surprise in his list. He created it in the 80s.
Bjarne Stroustrup
See his interview: http://youtu.be/NvWTnIoQZj4
His list:
- C++
- Java
- Python
- JavaScript
- C
- C#
Again, seeing C++ in his list is not a big surprise, he was one of the authors of the language in the 80s. The funny thing is that he mentions 6 languages.
Linus Torvalds
This two year old interview keeps popping up. In this video http://youtu.be/Aa55RKWZxxI mr. Linux mentions one programming language not to use.
The again, this person is well known from his more than colorful opinions about various issues. But anyway his work on Linux kernel and Git version management are well known, he is a fan of C.
me
Being a blog-author I have to express an opinion of my own. To solely copy/paste opinions of three very skilled persons is too much of a cheap thing. So, here goes:
- C
- Pretty much all languages created after 1970 owe something to C, it is imperative to know this.
- JavaScript
- When doing any kind of web-stuff, this is the only language being used in 100% of the cases. All browsers run this and it is the de-facto client-side language today.
- C#
- Very versatile compiled language by Microsoft, has lot of influence from C, C++, Java, PHP, Perl, etc. the list goes on. It is mainly used with .Net to create server-side stuff.
- PHP
- IMHO the most important web-server language there is. This is wildly popular and shares similarity with C, JavaScript, Perl, Visual Basic, etc.
In addition to learning programming languages, I encourage everybody to learn also following widely popular frameworks:
- Microsoft .Net
- Zend Framework
My reasoning between this is that if you understand how they work, you're pretty well covered and also going to Python/Django or Ruby on Rails is much easier task. I know that these are web-frameworks and people program a lot other stuff than web, but sticking to the topic of what languages to learn, these are the first ones to try. There are so many other frameworks, especially in PHP-land, but they don't have such an essential position as the framework made by people who created the PHP-language. In Microsoftland there are no other significiant frameworks to learn. Anyway, both are properly documented and lot of information can be found of them.
This is a classic question which I get to answer a lot. N00bs know the answer, but somebody outside the IT-business might ask something like that. This is also quite a popular question among young people trying to figure out if programming would be for them.
Anyway, here 5 Programming Languages Everyone Should Know from two people who actually have created some of the most popular languages currently used.
Nobody should call themselves a professional if they knew only one language.
- Bjarne Stroustrup
Larry Wall
See his interview: http://youtu.be/LR8fQiskYII
His list:
- JavaScript
- Java
- Haskell
- C
- Perl
Perl is not a surprise in his list. He created it in the 80s.
Bjarne Stroustrup
See his interview: http://youtu.be/NvWTnIoQZj4
His list:
- C++
- Java
- Python
- JavaScript
- C
- C#
Again, seeing C++ in his list is not a big surprise, he was one of the authors of the language in the 80s. The funny thing is that he mentions 6 languages.
Linus Torvalds
This two year old interview keeps popping up. In this video http://youtu.be/Aa55RKWZxxI mr. Linux mentions one programming language not to use.
The again, this person is well known from his more than colorful opinions about various issues. But anyway his work on Linux kernel and Git version management are well known, he is a fan of C.
me
Being a blog-author I have to express an opinion of my own. To solely copy/paste opinions of three very skilled persons is too much of a cheap thing. So, here goes:
- C
- Pretty much all languages created after 1970 owe something to C, it is imperative to know this.
- JavaScript
- When doing any kind of web-stuff, this is the only language being used in 100% of the cases. All browsers run this and it is the de-facto client-side language today.
- C#
- Very versatile compiled language by Microsoft, has lot of influence from C, C++, Java, PHP, Perl, etc. the list goes on. It is mainly used with .Net to create server-side stuff.
- PHP
- IMHO the most important web-server language there is. This is wildly popular and shares similarity with C, JavaScript, Perl, Visual Basic, etc.
In addition to learning programming languages, I encourage everybody to learn also following widely popular frameworks:
- Microsoft .Net
- Zend Framework
My reasoning between this is that if you understand how they work, you're pretty well covered and also going to Python/Django or Ruby on Rails is much easier task. I know that these are web-frameworks and people program a lot other stuff than web, but sticking to the topic of what languages to learn, these are the first ones to try. There are so many other frameworks, especially in PHP-land, but they don't have such an essential position as the framework made by people who created the PHP-language. In Microsoftland there are no other significiant frameworks to learn. Anyway, both are properly documented and lot of information can be found of them.
Exploring Dijit.Editor of Dojo toolkit
Sunday, August 25. 2013
My favorite JavaScript-library Dojo has a very nice HTML-editor. During a project I even enhanced it a bit. Anyway the Dojo/Dijit-documentation is not the best in the world, so I'll demonstrate the three operating modes that are available. All of them have the same functionality, but how they appear visually to the person doing the HTML-editing differs.
Classic fixed size
This is the vanilla operating mode. From the beginning of time, a HTML <TEXTAREA> has been like this (without any formatting, of course). A fixed block-container of multi-line text editor which will scroll on overflow.
Example:
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="150px"
data-dojo-props="extraPlugins:['insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
There is really nothing special, I'm just using two extra plugins insertanchor and viewsource to produce two new icons to the editor toolbar to add functionality for the end user. I found out that the plugins really need to be all lower-case for them to load properly. The class-names and filenames are in CamelCase, but putting them like that makes loading fail.
The obvious thing is that the editor is 150 pixels high. I didn't set the width, but since the editor is a simple div, any size can be set.
Auto-expanding
This is a plugin expansion for the previous. The only real difference is that this type of editor does not overflow. Ever. It keeps auto-expanding to any possible size to be able to display the entire text at once. During testing I found out that the auto-resize -code does not work on all browsers. There seems to be a discrepancy of exactly one line on for example Chrome. The bug manifests itself when you try to edit the last line, pretty much nothing is visible there. I didn't fix the bug, as I concluded that I won't be using this mode at all.
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="" minHeight="100px"
data-dojo-props="extraPlugins:['alwaysshowtoolbar','insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
This has three things to be taken into notice:
- The auto-expansion is achieved by plugin called alwaysshowtoolbar. It does not work in current Dojo version of 1.9.1, I had to fix it. See the patch at the end of this post.
- It is absolutely critical to set the height="". Forget that and the alwaysshowtoolbar-plugin does not work.
- It is advisable to set a minimum height, in my example I'm using 100 pixels. The editor will be really slim, if there is no text. This sets the size into something visually appealing.
Manually resizable
This is how a <TEXTAREA> behaves on many modern browsers. When using plugin statusbar you'll get a handle to resize the block. During testing I found out that it is a bad idea to allow the user to be able to make the editor wider. I enhanced the class with additional parameter which gets passed to plugin's constructor to limit the ResizeHandle functionality.
Example:
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="200px"
data-dojo-props="extraPlugins:[{name:'statusbar',resizeAxis:'y'},'insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
Note that specifying the resizeAxis won't work in your code. If you really want the code I can e-mail it to you, it is longish enough for me not to post it here. The height is 200 pixels initially, but user can make the editor any size.
Hope this helps to clarify the complexity of this fine editor. It would be also nice if a filed bug report would be processed in Dojo-project. Also their discussion in Google Groups is closed, the group is set to moderate any posts, but there is nobody doing that. So, effectively new people cannot post. There is nothing left, but to complain in a blog.
Appendix 1:AlwaysShowToolbar.js patch to fix the plugin-loading
--- dojo/dijit/_editor/plugins/AlwaysShowToolbar.js.orig 2013-07-19 13:21:17.000000000 +0300
+++ dojo/dijit/_editor/plugins/AlwaysShowToolbar.js 2013-08-02 17:31:44.384216198 +0300
@@ -13,7 +13,7 @@
// module:
// dijit/_editor/plugins/AlwaysShowToolbar
- return declare("dijit._editor.plugins.AlwaysShowToolbar", _Plugin, {
+ var AlwaysShowToolbar = declare("dijit._editor.plugins.AlwaysShowToolbar", _Plugin, {
// summary:
// This plugin is required for Editors in auto-expand mode.
// It handles the auto-expansion as the user adds/deletes text,
@@ -198,4 +198,11 @@
}
});
+ // Register this plugin.
+ // For back-compat accept "alwaysshowtoolbar" (all lowercase) too, remove in 2.0
+ _Plugin.registry["alwaysshowtoolbar"] = _Plugin.registry["alwaysshowtoolbar"] = function(args){
+ return new AlwaysShowToolbar();
+ };
+
+ return AlwaysShowToolbar;
});
My favorite JavaScript-library Dojo has a very nice HTML-editor. During a project I even enhanced it a bit. Anyway the Dojo/Dijit-documentation is not the best in the world, so I'll demonstrate the three operating modes that are available. All of them have the same functionality, but how they appear visually to the person doing the HTML-editing differs.
Classic fixed size
This is the vanilla operating mode. From the beginning of time, a HTML <TEXTAREA> has been like this (without any formatting, of course). A fixed block-container of multi-line text editor which will scroll on overflow.
Example:
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="150px"
data-dojo-props="extraPlugins:['insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
There is really nothing special, I'm just using two extra plugins insertanchor and viewsource to produce two new icons to the editor toolbar to add functionality for the end user. I found out that the plugins really need to be all lower-case for them to load properly. The class-names and filenames are in CamelCase, but putting them like that makes loading fail.
The obvious thing is that the editor is 150 pixels high. I didn't set the width, but since the editor is a simple div, any size can be set.
Auto-expanding
This is a plugin expansion for the previous. The only real difference is that this type of editor does not overflow. Ever. It keeps auto-expanding to any possible size to be able to display the entire text at once. During testing I found out that the auto-resize -code does not work on all browsers. There seems to be a discrepancy of exactly one line on for example Chrome. The bug manifests itself when you try to edit the last line, pretty much nothing is visible there. I didn't fix the bug, as I concluded that I won't be using this mode at all.
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="" minHeight="100px"
data-dojo-props="extraPlugins:['alwaysshowtoolbar','insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
This has three things to be taken into notice:
- The auto-expansion is achieved by plugin called alwaysshowtoolbar. It does not work in current Dojo version of 1.9.1, I had to fix it. See the patch at the end of this post.
- It is absolutely critical to set the height="". Forget that and the alwaysshowtoolbar-plugin does not work.
- It is advisable to set a minimum height, in my example I'm using 100 pixels. The editor will be really slim, if there is no text. This sets the size into something visually appealing.
Manually resizable
This is how a <TEXTAREA> behaves on many modern browsers. When using plugin statusbar you'll get a handle to resize the block. During testing I found out that it is a bad idea to allow the user to be able to make the editor wider. I enhanced the class with additional parameter which gets passed to plugin's constructor to limit the ResizeHandle functionality.
Example:
HTML for declarative instantiation:
<div id="terms-Editor" data-dojo-type="dijit.Editor"
height="200px"
data-dojo-props="extraPlugins:[{name:'statusbar',resizeAxis:'y'},'insertanchor','viewsource']">
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
</div>
Note that specifying the resizeAxis won't work in your code. If you really want the code I can e-mail it to you, it is longish enough for me not to post it here. The height is 200 pixels initially, but user can make the editor any size.
Hope this helps to clarify the complexity of this fine editor. It would be also nice if a filed bug report would be processed in Dojo-project. Also their discussion in Google Groups is closed, the group is set to moderate any posts, but there is nobody doing that. So, effectively new people cannot post. There is nothing left, but to complain in a blog.
Appendix 1:AlwaysShowToolbar.js patch to fix the plugin-loading
--- dojo/dijit/_editor/plugins/AlwaysShowToolbar.js.orig 2013-07-19 13:21:17.000000000 +0300
+++ dojo/dijit/_editor/plugins/AlwaysShowToolbar.js 2013-08-02 17:31:44.384216198 +0300
@@ -13,7 +13,7 @@
// module:
// dijit/_editor/plugins/AlwaysShowToolbar
- return declare("dijit._editor.plugins.AlwaysShowToolbar", _Plugin, {
+ var AlwaysShowToolbar = declare("dijit._editor.plugins.AlwaysShowToolbar", _Plugin, {
// summary:
// This plugin is required for Editors in auto-expand mode.
// It handles the auto-expansion as the user adds/deletes text,
@@ -198,4 +198,11 @@
}
});
+ // Register this plugin.
+ // For back-compat accept "alwaysshowtoolbar" (all lowercase) too, remove in 2.0
+ _Plugin.registry["alwaysshowtoolbar"] = _Plugin.registry["alwaysshowtoolbar"] = function(args){
+ return new AlwaysShowToolbar();
+ };
+
+ return AlwaysShowToolbar;
});
