LAMP & Oracle Technical Information

strategy statement from team

Hi all,

Over the last weeks there has been a lot of disussion, both internally and externally, about the strategies which have been proposed. However, we also missed a lot of voices from our community. We take responsibility for leaving many of you behind by focusing on a very corporate-management solution to the initial question which prompted this process. A question we think still is relevant: The identity of openSUSE both as a Community and as a Project.

Initially our goal was to answer: ”Who is openSUSE and what does it (want to) do?” prompted by the discussion about the default desktop at the openSUSE conference last year. In five years the openSUSE project has evolved from a fully company-driven project to a communty project where everybody can contribute. This has brought uncertainty and a lack of direction. The current lack of a clear ‘story behind it all’ is hampering our ability to establish a common identity and sense of security. From a marketing point of view, it becomes an uphill battle…

Throughout the process, we consulted some people and the discussion about a strategy started with the goal to solve this issue. However, many feel that ‘strategy’ and the approach to find one is not fitting our community. We lost most of you in the second paragraph of the strategy pages on the wiki – too much talk.

We would like to go back to the start and focus on describing who we are, as a community, instead of finding new ways to go. The input you all have given us by mail, forums, IRC and in person was valuable and we will use that. So that is what we will do:

• Highlight the story behind openSUSE

• Identify what users we target and illustrate what we offer to them,

• Connect it with the issues that matter most to our community

And then we will document this story, image, direction, strategy – or however your call it .

From you all – we will continue to seek your input on it once we post it. By mail, forum, IRC or in person – again. Without your help it won’t be much, so please think about that!

Greetings,

Your strategy team

As a result of the openSUSE Boosters’ ‘HackMeck‘ two weeks ago at FrOSCoN we are proud to present you with a new beta of openSUSE Connect.

Connect is supposed to become the central user database for the openSUSE project. Sounds bland, don”t it? But you know the Boosters, everything we do comes with a grain of spice and Connect is no different. The spice here are a lot of nifty social network features like user profiles, friending, groups, an event calendar and possibly more. Thats possible because on top of the user database we use a Free Software social network framework called Elgg. Elgg will help us to go a step further in one of the most important areas of the openSUSE project: Connecting our community. We do a very good job connecting code at the moment but there is no central place for openSUSE users to mingle, form relationships and meet collaborators.

Try it!

Did we whet your appetite? Want to try it? No problem, just head over to our beta instance http://connect.opensuse.org and login as user geeko with the password opensuse to try it out. Make some friends, create a group or run a poll. This instance is regularly deployed with the newest code from our git repository so you will always get the latest and greatest. But please don’t forget that this is a beta If you encounter any problems, guess what, make a bugreport in our bugzilla!

Help out!

Or how about you get your hands dirty? So far our experience with Elgg is wonderful. It’s a tidy, extensible and well designed piece of software. The community is very helpful and there is a lot (if not to say a butt-load) of functionality available. And if something is not there already we have found that we can easily add it. You could too you know? Elgg runs on a combination of Apache, MySQL and the PHP scripting language and as this is the most popular web server environment in the world we hope we can attract more people to help to fit Elgg to openSUSE’s needs. And on top of that it’s really easy to hack on it! The changes we did so far at the HackMeck and the last couple of weeks are self-contained in plugins that extend the basic functionality. The powerful data model and view system of Elgg make it possible to change it to openSUSE’s needs without ever touching the core functions. So if you are interested in helping, get to know Elgg and then get in contact with the openSUSE Boosters.

We hope you will enjoy this new openSUSE tool. And remember: Have a lot of fun…

- Added exclude pattern support to ruleset.xml file so you can specify ignore patterns in a standard (request #17683)
-- Use new exclude-pattern tags to include the ignore rules into your ruleset.xml file
-- See CodeSniffer/Standards/PHPCS/ruleset.xml for an example
- Added new --encoding command line argument to specify the encoding of the files being checked
-- When set to utf-8, stops the XML-based reports from double-encoding
-- When set to something else, helps the XML-based reports encode to utf-8
-- Default value is iso-8859-1 but can be changed with --config-set encoding [value]
- The report is no longer printed to screen when using the --report-file command line option (request #17467)
-- If you want to print it to screen as well, use the -v command line argument
- The SVN and GIT blame reports now also show percentage of reported errors per author (request #17606)
-- Thanks to Ben Selby for the patch
- Updated the SVN pre-commit hook to work with the new severity levels feature
- Generic SubversionPropertiesSniff now allows properties to have NULL values (request #17682)
-- A null value indicates that the property should exist but the value should not be checked
- Generic UpperCaseConstantName Sniff now longer complains about the PHPUnit_MAIN_METHOD constant (request #17798)
- Squiz FileComment sniff now checks JS files as well as PHP files
- Squiz FunctionCommentSniff now supports namespaces in type hints
- Fixed a problem in Squiz OutputBufferingIndentSniff where block comments were reported as not indented
- Fixed bug #17092 : Problems with utf8_encode and htmlspecialchars with non-ascii chars
-- Use the new --encoding=utf-8 command line argument if your files are utf-8 encoded
- Fixed bug #17629 : PHP_CodeSniffer_Tokens::$booleanOperators missing T_LOGICAL_XOR
-- Thanks to Matthew Turland for the patch
- Fixed bug #17699 : Fatal error generating code coverage with PHPUnit 5.3.0RC1
- Fixed bug #17718 : Namespace 'use' statement: used global class name is recognized as constant
- Fixed bug #17734 : Generic SubversionPropertiesSniff complains on non SVN files
- Fixed bug #17742 : EmbeddedPhpSniff reacts negatively to file without closing php tag
- Fixed bug #17823 : Notice: Please no longer include PHPUnit/Framework.php

Minor Bugfix release.

Bug Fixes
#4739 - regexp parsing of header values does not balance quoting correctly
- Fix sponsored by http://webyog.com
#17325 - empty body messages are valid messages
#17276 - remove &new usage which throws errors now

Metalink multichannel download, so package candy melts your screen, not your internet connection.

openSUSE 11.4 Milestone 1 is available today, Thursday, September 2 for developers, testers and community members to test and participate in the development of openSUSE 11.4. M1 starts off openSUSE 11.4 development at a cracking pace with performance improvements in the package management network layer and version updates to major components.

This milestone contains libzypp version 8.1, which has a new backend for http and ftp package downloads. MultiCurl replaces the old MediaAria backend, and brings support for zsync transfers and better Metalink download support. These will improve both repository refresh and package install and update performance. Metalink allows the multi-channel download of packages by downloading the individual blocks of a package in parallel from multiple servers. ZSync reduces the amount of data to download by only fetching the changed parts of a file instead of the whole file. This speeds up repository refreshes, since due to the way the repository data is structured, it is easy to locate the parts of the metadata that changed since the last update. The new Curl-based zypp backend also gives libzypp and therefore zypper and YaST better support for network proxies, by using the same proxy configuration as the rest of YaST instead of its own, and adds support for HTTP BASIC password-protected repositories. And as an added bonus, MultiCurl should eliminate slow and hanging package installations that occurred due to bugs in the old MediaAria backend.

Zsync efficiently downloads only the changed metadata. Sweet!

Other major components that have received updates from upstream projects for Milestone 1 include XOrg 1.9, KDE 4.5 and GNOME 2.32.0 Beta 1. Automated testing and brave openSUSE Factory testers have been validating early builds to make sure that Milestone 1 is suitable for others to test, so please download Milestone 1 and report bugs – the earlier a bug is reported in the development cycle, the more likely it is that it will be fixed on release day, March 10, 2011.

The next milestone is scheduled for September 30.

We're always looking for ways to make AWS an even better value for our customers. If you've been reading this blog for an extended period of time you know that we reduce prices on our services from time to time.

Effective September 1, 2010, we've reduced the On-Demand and Reserved Instance prices on the m2.2xlarge (High-Memory Double Extra Large) and the m2.4xlarge (High-Memory Quadruple Extra Large) by up to 19%.  If you have existing Reserved Instances your hourly usage rate will automatically be lowered to the new usage rate and your estimated bill will reflect these changes later this month.  As an example, the hourly cost for an m2.4xlarge instance running Linux/Unix in the us-east Region from $2.40 to $2.00. This price reduction means you can now run database, memcached, and other memory-intensive workloads at substantial savings. Here's the full EC2 price list.

As a reminder, there are many different ways to optimize your costs. When compared to On-Demand instances, Reserved Instances enable you to reduce your overall instance costs by up to 56%.  You pay a low, one-time fee to reserve an instance for a one or three year period. You can then run that instance whenever you want, at a greatly reduced hourly rate.

For background processing and other jobs where you have flexibility in when they run, you can also use Spot Instances by placing a bid for unused capacity. You job will run as long as your bid is higher than the current spot price.

-- Jeff;

 

 

Mike from Cirrhus9 sent me some information about their recent implementation of FindTheBest. With his permission, I am sharing some of the details with you so that you can get a better idea of how a sophisticated developer brings together a wide variety of technologies in order to construct a complete web application.

What is FindTheBest?

FindTheBest is an objective comparison search engine that allows people to choose a topic, compare options and decide what's best for them. It makes for faster and more informed decisions by allowing for easy comparison between all of the available options.

FindTheBest is organized into nine broad categories including: Arts and Entertainment, Business and Economy, Education, Health, Reference, Science, Society, Sports and Recreation and Technology. Each category includes dozens of Apps from Adventure Travel Vacations to Job Websites. Each App consists of a variety of related listings — from Vail and Whistler ski resorts under the Ski Resorts App to Barack Obama and Angelina Jolie under the Celebrities App — and each listing can be sorted by numerous key filters.

FindTheBest's structured search allows the user to quickly sort through factors and filters important to them, ultimately helping them make more objective and more informed decisions to important questions and answers.

The user interface blends sorting, filtering, and comparison to produce fast, data-driven pages. Here's a page with information on over 9000 libraries:

How Was it Implemented?

FindTheBest runs on a 64-bit Ubuntu server. The server boots from an EBS volume and has an XFS RAID-0 two volume EBS storage configuration, with daily snapshot backups via cron-driven calls to Eric Hammond's ec2-consistent-snapshot script. The server runs a classic LAMP stack with the addition of a lightweight email server and a full-text search engine, all monitored by a custom-tuned Nagios/Groundwork setup.

-- Jeff;

 

Under normal conditions, an Amazon S3 object in a bucket that is part of a CloudFront distribution can be cached at a CloudFront edge location per the object's TTL (Time to Live). In many situations it is possible to come up with a reasonable value for the TTL ahead of time. In other cases you may want the benefits of CloudFront's caching but you may also need to make changes to the S3 object at unpredictable times.

We've just added a new invalidation function to the CloudFront API. You can now POST a list of one or more objects to a CloudFront distribution and the objects will be removed from all of the edge locations within minutes. The invalidation happens in an asynchronous fashion and you can have several invalidation requests pending at the same time.

You can use this new feature in many different ways. Here are some ideas:

1. Update a CSS style sheet or some JavaScript that changes very infrequently.
2. Remove a video that was not properly encoded.
3. Remove information (e.g. a news story) that is inaccurate or no longer relevant.
4. Remove information that is the subject of a DMCA takedown notice.

There are no charges for the first 1000 invalidations per month. After that, each one will cost you $0.005 (one half of one cent).

You can still use the TTL feature and you can also use versioned URLs. Both techniques are preferred when you have the ability to control or predict the proper hold time for an object. There's no additional cost for either one, and there's no need to wait for the invalidation to take effect (typically 10 to 15 minutes). Invalidation is appropriate when the hold time is unpredictable. 

TTLs and versioned URLs are great when you have tight control over the object's lifetime, with new objects replacing the old on a regular cycle or as part of a planned release. Invalidation is appropriate when objects can change with little or no notice.

The following third-party products already include support for this new feature:

• CloudBuddy Personal.
• CloudBerry Explorer.

Let me know if your product supports it, and I'll amend this blog post to include it. Leave a comment or email me at awseditor@amazon.com.

The AWS Simple Monthly Calculator now supports CloudFront Invalidations and RDS Reserved DB Instances.

-- Jeff;

As our community grows, it is imperative that we preserve the things that got us here; namely, keeping Drupal a fun, welcoming, challenging, and fair place to play. The new Drupal Code of Conduct (DCOC) states our shared ideals with respect to conduct. Think of this as coding standards for people. It is an expression of our ideals, not a rulebook. It is a way to communicate our existing values to the entire community.

Our friends at Ubuntu have blazed a brilliant trail in this area. They use Drupal as their CMS, and in turn we have embraced their Code of Conduct. This code of conduct is essentially identical to that used by Ubuntu, except that the name of the project has been changed, and the conflict resolution process has been removed since we don't have one.

The DCOC has been under discussion for several months on groups.drupal.org and discussed further at Drupalcon Conpenhagen. Folks who are interested in talking more about the DCOC should do so in the Drupal.org Policies group.

The short version:

1. Be considerate
2. Be respectful
3. When we disagree, we consult others.
4. When we are unsure, we ask for help.
5. Step down considerately.

read more

On September 18, it is international Software Freedom Day. Software Freedom Day aims to celebrate Free Software and the people behind it. It wants to spread the word about Free Software and help people find each other.

We, the openSUSE community, of course need to be part of this! So this is a call to you all: mark september 18 in your agenda. Check the Software Freedom Day website and find the meeting in your neighboorhood.

If there is none, how about setting one up? Go to the Start Guide and get to work. You can organize your own event!

Of course, openSUSE will help you. If you plan to go to an event or organize one, get in contact with us at marketing@opensuse.org ! If you mail us your information we can send you a package with openSUSE Promo DVD’s, a poster and some stickers so you can be an ambassador for openSUSE at your local event.

Help us spread the word!

Greetings,

The openSUSE marketing team

First release.

- fixed bug #16275: split() is deprecated in PHP 5.3

note:
open todo items:
- fix pattern escaping using GLOB instead of LIKE or create and register own implementation of LIKE

- fixed bug #16281: getTableFieldDefinition() does not recognize NULL defaults
with datatype [Holger Schletz]
- fixed bug #16384: alterTable() does not remove NOT NULL constraints [Holger Schletz]
- fixed bug #16405: Compatibility issues with escaped strings [hschletz]

open todo items:
- enable pg_execute() once issues with bytea column are resolved
- use pg_result_error_field() to handle localized error messages (Request #7059)
- add option to use unnamed prepared statements
(see http://www.postgresql.org/docs/current/static/protocol-flow.html "Extended Query")

First release

note:
- please use the latest ext/oci8 version from pecl.php.net/oci8
(binaries are available from snaps.php.net and pecl4win.php.net)
- by default this driver emulates the database concept other RDBMS have by
using the "database" option instead of "username" in the DSN as the username name.
This behaviour can be disabled by setting the "emulate_database" option to false.
- the multi_query test failes because this is not supported by ext/oci8
- the null LOB test failes because this is not supported by Oracle

open todo items:
- enable use of read() for LOBs to read a LOB in chunks
- buffer LOBs when doing buffered queries

- fixed bug #16003: incorrect check for error after mysqli_store_result
- fixed bug #16147: first prepared statement is emulated when using factory with mysql
- fixed bug #17037: 'on update' not mentioned in tableInfo()
- fixed bug #17065: There is no NEW row in on DELETE trigger (fix error in FK constraint triggers)
- fixed bug #17650: lastInsertId can not handle bigint, forces cast to integer [alexpw]
- return ON UPDATE|DELETE action in getTableConstraintDefinition()
- result->free() now works on multiple result sets

open todo items:
- use a trigger to emulate setting default now()

- fixed bug #15650: mysqli function used in setCharset()
- fixed bug #16003: incorrect check for error after mysql_store_result
- fixed bug #16147: first prepared statement is emulated when using factory with mysql
- fixed bug #16669: hostspec is ignored when protocol is unix
- fixed bug #17037: 'on update' not mentioned in tableInfo()
- fixed bug #17065: There is no NEW row in on DELETE trigger (fix error in FK constraint triggers)
- fixed bug #17650: lastInsertId can not handle bigint, forces cast to integer [alexpw]
- return ON UPDATE|DELETE action in getTableConstraintDefinition()

note:
- the multi_query test failes because this is not supported by ext/mysql

open todo items:
- use a trigger to emulate setting default now()

- Fixed bug #16612: Added the timestamp database attribute [genericbob]
- fixed bug #16118: escape doesn't take into account trailing backslashes [urkle]
- request #16903: Add ability to use ODBTP extension [hedroom]
- fixed numRows() with setLimit()

open todo items:
- explore fast limit/offset emulation (Request #4544)

- fixed regression in lastInserID()
- fixed bug #15971: typo in standaloneQuery() [elpaso]
- fixed bug #16350: typo in setTransactionIsolation() [nilya]
- fixed bug #16741: Manager::listTableIndexes and listTableContraints should check NULL or 0 [nilya]
- fixed bug #17676: use connection resource when fetching LOBs [renskiy]

- fixed bug #15912: MDB2::loadClass can return MDB2_OK without loading class
- fixed bug #16020: unable to set params via dsn string in parseDSN()
- fixed bug #16508: reusing var name + return by ref [rix0r]
- fixed bug #16727: quote(url, 'clob')) fetches URL even if lob_allow_url_include option is false
- fixed unreported bug in parseDSN() and oracle's Easy Connect syntax
- fixed bug #16973 and #17039: Wrong _skipDelimitedStrings behavior with escaped quotes
- fixed bug #16994: incompatible declaration of setTransactionIsolation() in some drivers
- dropped PHP 4 support
- PHP 5.3 compatibility fixes
- performance tweaks (reduced number of internal function calls)
- added new sqlsrv driver (native SQL Server driver) [Chris Pucci, Mike Ketcham]
- added new ODBC experimental driver [pschellhaas]
- Switched test suite to PHPUnit 3.5

open todo items:
- handle autoincrement fields in alterTable()
- add length handling to LOB reverse engineering
- add EXPLAIN abstraction
- add cursor support along the lines of PDO (Request #3660 etc.)
- add PDO based drivers, especially a driver to support SQLite 3 (Request #6907)
- add support to export/import in CSV format
- add more functions to the Function module (MD5(), IFNULL(), etc.)
- add support for database/table/row LOCKs
- add support for CHECK (ENUM as possible mysql fallback) constraints
- generate STATUS file from test suite results and allow users to submit test results
- add support for full text index creation and querying
- add tests to check if the RDBMS specific handling with portability options
disabled behaves as expected
- handle implicit commits (like for DDL) in any affected driver (mysql, sqlite..)
- add a getTableFieldsDefinitions() method to be used in tableInfo()
- drop ILIKE from matchPattern() and instead add a second parameter to
handle case sensitivity with arbitrary operators
- handle LOBs in buffered result sets (Request #8793)