Have an existing disk that has specialized partitions that arnd easily backed up using fsarchiver. For example, Windows 8 disk on a brand new computer.

Solution

1. Use fsarchiver probe to find the disk.
2. Be sure the disk is not mounted.
3. cd to a target directory with plenty of space. Note: the image will be compressed.
4. To create the image, run this command: dd if=/dev/sdc conv=sync,noerror bs=32M | pv | xz -c | split -b 23000m

Hints:

• The source disk for the above command is at /dev/sdc.
• Google the model of the hard disk shown on fsarchiver probe to find the buffer size of the disk (32MB in the above example)
• Be sure pv and xz are installed.
  
• The target media of the above is a blu-ray disk. (-b 23000m). Lower this number for DVD’s or CD’s.
• This process could take several days. Just let it run in the background. A 2T drive took me about 3 straight days.

1. New files will appear on the current directory as xaa, xab, xac, etc. When a new file appears, rename the old file to something helpful like: new_system_drive_01.dd.xz. DO NOT rename or burn a file that is still being created.
2. Burn these files to a blu-ray disk. I recommend x3b for Ubuntu.

References:

• How to use dd
• dd with xz
• dd with split

    <dependency>
	<groupId>org.apache.poi</groupId>
	<artifactId>poi</artifactId>
	<version>3.9</version>
    </dependency>
    <dependency>
	<groupId>org.apache.poi</groupId>
	<artifactId>poi-ooxml</artifactId>
	<version>3.9</version>
    </dependency>

java

        Workbook wb=new XSSFWorkbook(new FileInputStream("test.xlsx"));
        Sheet ws=wb.getSheetAt(wb.getActiveSheetIndex());
        int rowIndex=1;
        while (ws.getRow(rowIndex)!=null && rowIndex<100){
            Row row=ws.getRow(rowIndex);
            System.out.println(row.getCell(0).getStringCellValue());
            rowIndex++;
        }

LibreOffice and Microsoft Word are very similar products. However, the default settings are different for both products, especially for Heading Styles. Worse yet, the LibreOffice keyboard shortcut defaults are overridden by other Ubuntu keyboard shortcuts. This article helps redefine the settings and offers a template for better Heading Styles.

What is a Heading Style?

A Heading Style is a way to partition a large document into smaller chapters and sub-chapters. The Heading Styles take care of automatic numbering and can use numbers, letters, or roman numerals. Heading Styles are also used for building a Table Of Contents as well as cross references to other locations on the document.

Fixing the Keyboard Mapping

By default, Microsoft Word uses Alt-Shift-Left and Alt-Shift-Right to promote or demote a paragraph into a Heading Style. LibreOffice does not have this keyboard mapping. Instead, they use Ctrl-Alt-Up and Ctrl-Alt-Down. Unfortunately, these keys are also used to change the current desktop on Ubuntu.

On LibreOffice 3.5 (default version on Ubuntu 12.4), to set the keyboard mapping to Alt-Shift-Left and Alt-Shift-Right choose one of two methods:

Method 1: Load a configuration file

This method involves downloading a cfg file to the system and using it to replace the default keyboard mappings. This method will remove any other keyboard customizations you have made. Use at your own risk.

1. Download this file.
2. In LibreOffice, click Tools|Customize…
3. Click the Keyboard tab.
4. Click the Load button.
5. Select the file from step 1.
6. Click Ok.

Method 2: Manual Fix

This method is a bit more involved but you will retain any other keyboard mappings.

1. Click Tools|Customize…
2. Click the Keyboard tab.
3. Under Shortcut Keys, be sure the Writer radio button is selected.
4. Select the Shortcut Key: Alt+Shift+Left.
5. Under Functions, select the Numbering Category.
6. Select the Promote One Level With Subpoints Function.
7. Click Modify.
8. Select the Shortcut Key: Alt+Shift+Right.
9. Select the Demote One Level With Subpoints Function.
10. Click Modify.
11. Click Ok.

Outline Numbering Required

The above fixes will not work until Outline Numbering is activated on the current document. Also, the paragraph must be set to one of the Heading styles. Simply, press Ctrl-1 to make the current paragraph a Heading 1 style (this is a default keyboard shortcut for LibreOffice).

Using a Template

The default Heading styles are very ugly. There is no default numbering scheme and there not good visual queues to help the reader understand the depth of the next section break. Here’s the default template:

And here is an improved template:

To import the improved template:

1. Download this file.
2. Click File|Templates|Organize….
3. Right-click My Templates.
4. Click Import Template….
5. Find the file from step 1.
6. Click Close.

Now the template is available for use. To create a document based on this template:

1. Click New|Templates and Documents.
2. Click the template that was imported.

At this point, you should have a blank document ready to be used.

It’s often necessary to have several configurable parameters on a constructor for a class. Instantiating these classes can be a painful exercise in remembering the order of the parameters and the properly passing them down from parent classes.

Commons Configuration

Commons Configuration solves the problem of passing several primitive parameters to a class. Instead, of passing each primitive, simply pass a Configuration object and pull the primitives one-by-one.

Consider the following example:

public class DoSomething{
  public DoSomething(String user,String environment){
  }
}
</java>

In the above example, it’s necessary to change the signature when another parameter needs to be passed into the class. Thus, all client code using the constructor has to change as well.

Consider the additional code:

public class DoSomething{
  final public static String CONFIG_CURRENT_USER="Something.CurrentUser";
  final public static String CONFIG_ENVIRONMENT="Something.Environment";

  public DoSomething(Configuration config){
    this(config.getString(CONFIG_CURRENT_USER),config.getString(CONFIG_ENVIRONMENT));
  }
  public DoSomething(String user,String environment){
  }
}

Now, all the existing client code can remain intact, and a Configuration object can be passed to the new constructor.

Integration with Guice

Guice now becomes very useful. Generally speaking, injecting primitive types onto a constructor is not Guice’s strength. However, now that the constructor only needs to inject a Configuration object, leveraging Guice just means adding an Annotation like this:

public class DoSomething{
 ...
  @Inject
  public DoSomething(Configuration config){
    this(config.getString(CONFIG_CURRENT_USER),config.getString(CONFIG_ENVIRONMENT));
  }
 ...
}

Then, just create an injector in the client code:

Properties props=new Properties();
props.load(ClassLoader.getSystemResourceAsStream("app.properties"));
URL propsUrl=ClassLoader.getSystemResource("app.properties").toURI().toURL();

final Configuration config=new PropertiesConfiguration(propsUrl);

Injector injector=Guice.createInjector(new Module() {
    public void configure(Binder binder) {
        binder.bind(Configuration.class).toInstance(config);
    }
});

DoSomething something=injector.getInstance(DoSomething.class);

It is definitely a lot of client code, but really only needs to happen once for the entire project. JUnit tests can use the MapConfiguration to define the parameters directly in the code.

This a Java Swing-based application that I made to manage my ToDo list. There are several ToDo List Manager applications available. However, they only give a Name and a Done checkbox for each item. This project gives each ToDo item a hierarchy, workflow and relative priority. This makes it much easier for the user to assess which item is most important at any given time, but also keep track of items that are important for the future.

Workflow Status

Each ToDo item has a Workflow State. This state is determined explicitly by the user. Here is a list of available states:

• Open: The item needs to considered.
• Accept: The item is worth doing.
• Reject: The item is not worth doing.
• Close: The item is complete.
• Open Pending: The item needs to be considered only after another item changes state.
• Accept Pending: The item is worth doing only after another item changes state.

The item starts as Open then user determines if the item is really worth doing or not. This allows a brainstorming capability. It also retains the rejected items for later consideration. There is no need to delete an item. Many times, a Rejected item may be reconsidered. A Rejected item may have notes that explain why that item was not considered last time.

The two pending states, Open Pending and Accept Pending, delay the state on an item until other items change state. This is useful when more information will help a decision or if an item can’t be completed until another task is done.

Inherited Status

Each item in the system must be placed within a tree structure hierarchy. Each task is a child of one other task. This allows for each task’s state to be conditional on its parents. Consider the following example:

• The user wants to install Linux.
• The user must buy a Linux book.
• The user must burn CD’s for Linux.
• The user must download the iso images.
• The user must buy blank CDR’s before burning the CD’s

That’s quite a complicated ToDo list. At first glance, it’s not obvious what the user should do first. However, he could put this information in a tree to make it more straightforward. For example:

Figure 1: Hierarchy of Items

See in Figure 1 how the the “Install Linux” item is broken down into “Buy Linux Book” and “Burn CD’s”. Then the “Burn CD’s” item is broken down into “Download iso images” and “Buy Blank CDR’s”. This makes the dependencies more obvious. Now, if the user happened to find a book that included the Linux disks he needed, he could Reject “Burn CD’s”. Then, “Download iso images” and “Buy Blank CDR’s” would automatically become Inactive.

There are a number of possible inherited states. They are:

• Active: If all parents are Accepted.
• New: If a parent is Open.
• Waiting: If a parent has a Pending state.
• Inactive: If a parent is Closed or Rejected.

Filters

A Filter focuses the user’s attention only on the relevant ToDo items. Filters tie together Inherited and Workflow status to determine the overall condition of the items. There are six filter states:

• Assign: Shows all the non-Inactive Open items. Useful for seeing which items need to be Rejected or Accepted.
• Work: Shows all the Active Accepted items that have no Active Accepted parents. Useful for seeing what to do today.
• Events: Shows all the non-Inactive Pending nodes. Useful for seeing dependencies between unrelated nodes.
• Review: Shows all the Closed nodes. Useful for seeing progress over time.
• Recycle: Shows all the Rejected nodes. Useful for finding previously rejected items.
• No filter: Shows all nodes.

The user will likely spend most of their time in the Assign and Work filters. The Assign filter is great for sifting through tasks after a brainstorm session. The Work filter shows only what needs to be completed right now.

Priority

Each item has a priority once it is Accepted. The priority is always relative to other items. The program has a table-view that is used for setting and viewing priority. In the table, the user simply clicks-and-drags to move the items to match their real priority. Then, the user can click the Status column to Close an item when completed.

Figure 2: Table View

Pending

Some items will not be children, but will be dependent on others. For example, if the user wants to learn how to use GnuCash after Linux is installed. It’s not necessary to make GnuCash the parent of the “Install Linux” item. Instead the user could simply set a prerequisite of “Install Linux” for the “Learn GnuCash” item. Then set the Workflow State of “Learn GnuCash” to “Accepted Pending”. In that case, when “Install Linux” is Closed, “Learn GnuCash” will be Accepted.

User Interface

Now that we’ve discussed the pieces of the software. Take a look at the User Interface:

Figure 3: User Interface

The upper-left corner is the filter selection. The lower-left corner is the Tree. It has a tab to go between the Tree and Table. The right side is the details of the current item. In this case, it is the “Install Linux” item.

Class Model Diagram

The program consists of several classes, many of them subclasses of Swing classes. Here is the Class Model Diagram for the program:

Figure 4: Class Model Diagram

The Diligent class is the entry point of the program. It instantiates the DiligentFrame class. DiligentFrame is a container for all the Swing Components. It instantiates a table and tree model, DiligentTableModel and DiligentTreeModel respectively. DiligentTableModel gets its data from the DiligentTreeModel. DiligentTreeModel is a subclass of Swing’s DefaultTreeModel. DiligentTreeModel uses DiligentNode for all of its nodes.

DiligentNode is the main container for all the information of every ToDo item. It is a subclass of Swing’s DefaultMutableTreeNode. It has a static class member variable, filter, for filtering the entire tree. filter is of class DiligentFilter. DiligentFilter has parameters on its methods to filter for both the tree and the table views. Filtering for a tree allows the parents to not be filtered out if a child is not filtered out.

There are few classes that have a DiligentNode within them. DiligentDetails is responsible for showing the detailed information about each node. DiligentNodeSelector is a special dialog for choosing a single node. DiligentEvent analyzes the state of a node to trigger a change in state of another node.

There are four miscellaneous classes within the architecture. DateParse contains a static method to convert a String to a Date. The Period class contains static methods to convert interval dates to and from a number of milliseconds. For example, an interval date might be “2w1d” for two weeks and one day. The ResourceManager is responsible for loading the icons and can be optimized to read the icons only once from the disk. The DiligentNodeRenderer is used by the JTree to show the icons next to each ToDo item.

Future Upgrades

It’s easy to go use this architecture for many things. This program could become a full Project Management system. It currently saves all the information in a single file. It could be modified to save the information to a central database with several users. Each item could have an owning person or group assigned to it. For now, it works great for managing my own personal tasks.

Conclusion

This program works great for managing personal tasks. For a copy of this program download it here.

When creating a jar-with-dependencies, the resources in src/test/resources are automatically added to the jar. This folder often contains a .properties file only valid for the test environment and should not be included in the final jar.

Additional Note

Adding a resource tag to exclude the directory is ignored by Maven running in Eclipse. It’s not clear why.

Solution

Right-click the source folder “src/test/resources” and click “Build Path|Configure Exclusion/Inclusion Filters…”. Add an exclusion filter of **. Now the resources will not be included in the final jar file.

Testing from Eclipse

When running from eclipse, it is possible to include the content of src/test/resources in the classpath. Here’s how:

1. Right-click src/main/java
2. Click “Properties”
3. Click “Run/Debug Settings”
4. Select a Launch Configuration you want to change.
5. Click Edit.
6. Click the Classpath tab.
7. Click Advanced…
8. Click Add Folders and Click Ok
9. Select the src/test/resources folder on your project.

Now when running from eclipse, the test resources are in the classpath. It maybe necessary to make this change for several launch configurations.

By default Firefox opens PDF files using an Adobe plug-in which often fails and is trapped inside the browser.

Solution

1. Click Edit|Preferences|Applications.
2. Search for PDF
3. Change the Action to Use Adobe Reader (default).

This page describes how to configure Linux to properly route network traffic between two separate networks. This works the same for wireless and wired connections.

IP Address of Middle

Have a computer that has a NIC with an IP address on each network. I’m not sure if this is called a router or a bridge or something else. For this page, I call this computer the “middle”. Use ifconfig to verify that each network card is up. Try pinging a host on each network as a second check.

Routing of Middle

Check the routing for each of the cards in the middle. If pinging failed, it was probably because routing is broken. Use netstat -nr to quickly see the routes to each subnet. Then, just add routes to the appropriate NIC for each subnet. Adding a route will look like this:

route add -net 192.168.0.0 netmask 255.255.255.0 dev eth0
route add -net 192.168.1.0 netmask 255.255.255.0 dev eth1

NAT in Middle

Turn on NAT’ing on the middle:

echo 1 > /proc/sys/net/ipv4/ip_forward

Route the clients

For each client machine, simply adjust the routing to use the middle as a gateway. I’m assuming the middle has the IP address of .5 in each network (eg 192.168.0.5 and 192.168.1.5).

A computer on the 192.168.1.0 network would use the following command to NAT to 192.168.0.0:

route add -net 192.168.0.0 netmask 255.255.255.0 gateway 192.168.1.5 dev eth0

This tells the the client machine that all 192.168.0.x traffic goes thru eth0 to 192.168.1.5 for further routing. At that point it will use the routing tables in middle jump to the 192.168.0.x subnet and go to the target machine.

I want to have a POM that will create a JAR file with all dependencies, but exclude src/main/resources/config.properties.

Solution

1. Make a folder (not a source folder) in Eclipse called src/main/eclipse.
2. Put the config file in the new folder.
3. Add the folder to the class path by going into the Run Configuration.

Adding the folder to src/test/resources does not work unless src/test/resources is not a code folder.

I tried this, but somehow Maven always wants to move that file!

 <build>
 	<pluginManagement>
 		<plugins>
 			<plugin>
 				<groupId>org.maven.apache.plugins</groupId>
 				<artifactId>maven-resources-plugin</artifactId>
 				<version>2.4.3</version>
 			</plugin>
 		</plugins>
 	</pluginManagement>
  	<resources>
  		<resource>
  			<directory>src/main/resources</directory>
  			<excludes>
  				<exclude>**/*.properties</exclude>
  			</excludes>
  		</resource>
  	</resources>
  	<plugins>
  		<plugin>
  			<artifactId>maven-assembly-plugin</artifactId>
            <configuration>
                <archive>
            		<manifestFile>src/main/resources/MANIFEST.MF</manifestFile>
          		</archive>
                <descriptorRefs>
                    <descriptorRef>jar-with-dependencies</descriptorRef>
                </descriptorRefs>
        		<excludes>
        			<exclude>**/config.properties</exclude>
        		</excludes>
            </configuration>
  		</plugin>
  		<plugin>
        	<groupId>org.apache.maven.plugins</groupId>
        	<artifactId>maven-surefire-plugin</artifactId>
        	<version>2.7.2</version>
        	<configuration>
          		<skipTests>true</skipTests>
        	</configuration>
        </plugin>
        <plugin>
        	<groupId>org.apache.maven.plugins</groupId>
        	<artifactId>maven-resource-plugin</artifactId>
        	<version>2.4.3</version>
        </plugin>
  	</plugins>
  </build>

References

• http://stackoverflow.com/questions/2728993/how-to-exclude-poperties-file-from-jar-file
• http://stackoverflow.com/questions/4113697/in-maven-how-to-exclude-resources-from-the-generated-jar

rm ~/workspace/.metadata/.plugins/org.eclipse.datatools.sqltools.result/

When trying to delete a bunch of files using something like sudo rm *, I get this error:

sudo: unable to execute /bin/rm: Success

Solution

This is caused by a buffer overflow when expanding the *. Instead, use find. Here’s an example:

sudo find . -name '*' -exec rm {} \;

Note that the wildcard (*) is enclosed in single quotes which will (unlike double quotes) prevent it from expanding.

1. Install the log4cxx. sudo apt-get install liblog4cxx

2. Write code:

   # include "log4cxx/logger.h"
   # include "log4cxx/basicconfigurator.h"
   # include "log4cxx/helpers/exception.h"
   
   using namespace log4cxx;
   using namespace log4cxx::helpers;
   
   int main(){
   	BasicConfigurator::configure();
   	LoggerPtr logger(Logger::getLogger("main"));
   	LOG4CXX_DEBUG(logger, "Hello World");
   }
   

3. Compile with -llog4cxx.

4. Run it.

You should see something like:

1 [0x3f956f44b980] DEBUG main null - Hello World

I want to add some special parameter to my MANIFEST.MF, but I’m using the maven-assembly-plugin to build my project.

Solution

Change the POM to look something like this:

  		<plugin>
  			<artifactId>maven-assembly-plugin</artifactId>
            <configuration>
                <archive>
            		<manifestFile>src/main/resources/MANIFEST.MF</manifestFile>
          		</archive>
                <descriptorRefs>
                    <descriptorRef>jar-with-dependencies</descriptorRef>
                </descriptorRefs>
            </configuration>
  		</plugin>
  		<plugin>

I was able to compile and link a program using curlpp, but I get this error:

error while loading shared libraries: libcurlpp.so.0: cannot open shared object file: No such file or directory

Solution

This will may only work in Ubuntu:

1. sudo bash -c "echo /usr/local/lib/ > /etc/ld.so.conf.d/local.conf"
2. sudo /sbin/ldconfig

References

You want to always boot a certain kernel in Ubuntu. There is no /boot/grub/menu.lst.

Solution

1. Edit this file: /etc/default/grub/ and add:

   GRUB_DEFAULT=saved
   GRUB_SAVEDEFAULT=true
   

2. Run this utility: sudo update-grub

Now next time you reboot, choose your favorite kernel. This decision will be remembered.

References

• https://help.ubuntu.com/community/Grub2
• http://ubuntuforums.org/archive/index.php/t-1453448.html

This page is a tutorial on how to use LibSVM to a solve a simple, real world problem using a bash shell.

Process

1. Download LibSVM.
2. Untar and run make. ./configure not necessary.
3. Read Section A of this article. We are going to try the example in section A.1.
4. Download the training data here.
5. Scale the data: ./libsvm-3.0/svm-scale -l -1 -u 1 -s range1 svmguide1 > svmguide1.scale. This forces the dataset to be between -1 and 1. The name of the scale is range1 which will be used later.
6. Find the ideal parameter values: cd libsvm-3.0/tools/; python grid.py ../../svmguide1.scale; cd -. A plot will appear. this process takes significantly longer than the others. Just keep waiting. You have to cd to the tools directory for the python script to run properly. I got the result of 2.0 2.0 96.9893. I believe the first parameter is c and g is the second.
7. Using the above parameters, create the model: libsvm-3.0/svm-train -c 2 -g 2 svmguide1.scale. There will now be a svmguide1.scale.model file in your current directory.
8. Download the test data here.
9. Scale the test data: ./libsvm-3.0/svm-scale -r range1 svmguide1.t > svmguide1.t.scale
10. Run the test: ./libsvm-3.0/svm-predict svmguide1.t.scale svmguide1.scale.model svmguide1.t.predict
11. You should see an accuracy of “96.875%”.

Note: it would be easier to run tools/easy.py but this helps get an understanding of what it does underneath.

Data Files

The data files are a space-delimited array of tuples that start with the output classification. A sample line looks like this:

0 1:6.640930e-01 2:2.540829e+01 3:-1.155107e-01 4:4.739574e+01.

The above line shows a result for category “0”. The first parameter is 0.6640930, but is represented with the parameter number, a colon, and the number itself in scientific notation. The number does not have to be in scientific notation and is usually not in that form after it is scaled.

Note that both training and test data have the exact same format. An interesting test is to run the model on the training data. I tried it on the above and got 97.1188% accuracy. If the accuracy is bad on the training data, there are probably serious issues.

Scaled Data

The scaling information for the scaled files are not stored in the data file. Instead there is a special file to store this information. svm-scale uses the -s parameter to save this special file and the -r parameter to read the file and apply the same scaling as the training data. Here are the contents of the file produced by the above commands:

x
-1 1
1 0 297.05
2 -4.555206 581.0731
3 -0.7524385 0.7170606
4 8.157474000000001 180

Details

The article we used as a basis was clear on the point to be sure to scale and optimize the parameters. In this case, parameter optimization didn’t add much accuracy but other examples in that article show that it can be a huge difference.

Prediction

Training and prediction are basically the same command. The predicted results are stored in yet another file. This file just contains the predicted values for each line of input. I made my own set of test data here:

0 1:2.612900e+00 2:1.832880e+01 3:1.222401e-01 4:7.389243e+01
0 1:2.004300e+01 2:2.822780e+02 3:2.021259e-01 4:1.411875e+02

The first line was taken from a “0” and the second line was taken from a “1” in the sample data set. As expected, svm-predict gave an accuracy of 50% and the predict’s output file has a 0 and 1 for the expecting line. I get an error when trying a letter or a “?” mark in the first character of every line. If I put a space for the first character, I get a 100% accuracy result and the predict file shows all “1’s”..beware of this wrong answer.

The best strategy when doing prediction is to simply leave the first character of every line as a “0”.

After doing a sudo, I get this error:

X11 connection rejected because of wrong authentication.

Solution

1. Be sure the non-root user can get X working. Try running xterm.
2. Run xauth list as the non-root user. You should see something like:

machinename/unix:14 MIT-MAGIC-COOKIE-1 058aexd13d0436b3463abd79acd31234

1. Run echo $DISPLAY to see which number is used for this host. (eg localhost:14.0 means 14)
2. Login as root. sudo -i
3. Add the cookie. auth add machinename/unix:14 MIT-MAGIC-COOKIE-1 058aexd13d0436b3463abd79acd31234

References

• http://jianmingli.com/wp/?p=724

Maven doesn’t seem to connect to the Internet. I have the http_proxy environment variables set. I also configured Eclipse Preferences|General|Network Connections to connect to my proxy. However, when I start up eclipse, I get something like this:

1/19/11 3:25:34 PM CST: Updating index central|http://repo1.maven.org/maven2
1/19/11 3:25:35 PM CST: Unable to update index for central|http://repo1.maven.org/maven2

Solution

The m2 Maven plugin for Eclipse ignores the http_proxy Environment variable and the Network Connection configuration. Instead, it’s necessary to set the proxy using the settings.xml file. Settings.xml is found on your home directory. Here’s an example:

<settings xmlns="http://maven.apache.org/SETTINGS/1.0.0"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/SETTINGS/1.0.0
                      http://maven.apache.org/xsd/settings-1.0.0.xsd">
  <localRepository/>
  <interactiveMode/>
  <usePluginRegistry/>
  <offline/>
  <pluginGroups/>
  <servers/>
  <mirrors/>
  <proxies>
    <proxy>
      <id>myproxy</id>
      <active>true</active>
      <protocol>http</protocol>
      <host>proxy.somewhere.com</host>
      <port>8080</port>
      <username>proxyuser</username>
      <password>somepassword</password>
      <nonProxyHosts>*.google.com|ibiblio.org</nonProxyHosts>
    </proxy>
  </proxies>
  <profiles/>
  <activeProfiles/>
</settings>

There is a hyperlink to this file in Window|Preferences|Maven|User Settings in Eclipse. After making your changes, just click “Update Settings”.

References

• Maven Settings Reference

Waffles is an OpenSource machine learning library for C++ that also has a command-line interface. I am interested in using it for some projects, but I the documentation does not have full examples on how to use it, just a few one-liners here and there. The Wikipedia article on Naive Bayes has a good example walk-thru of Naive Bayes. I intend to implement that example with Waffles.

Data

The Wikipedia article claims the following input data can predict gender:

Then, giving it the sample data of: {6,130,8}, it should produce posterior numerator (male) = 6.1984e-09 and posterior numerator (female) = 5.3778e-04.

So, I created this CSV file.

Process

First, I converted the csv file to Waffles format, arff. Note that the CSV file must not have a header:

waffles_transform import gender.csv > gender.arff

Then, I tried to create a naivebayes model using the newly created arff file:

waffles_learn train gender.arff naivebayes

and got: GNaiveBayes does not support continuous attributes. You should discretize first to convert real values to nominals.

I’m not sure if discretizing the data will produce the desired result, but I tried it anyway:

waffles_learn train gender.arff discretize naivebayes > gender.twt

Then, I tried some cross validation:

waffles_learn crossvalidate -reps 50 -folds 2 gender.arff -labels 0 discretize naivebayes

I got: Attr: 3, Mean accuracy: 0.74, Deviation: 0.23815367248307. Not bad!

The -labels 0 tells the system to use the first column of data as the label (male,female). Running the same command several times gives different results.

Now, let’s convert the test data into an arff file.

waffles_transform import test.csv>test.arff

And let’s test it:

waffles_learn predict gender.twt test.arff

The raw numbers are definitely different from numbers given by Wikipedia:

I’m not 100% sure of the meaning of each number. The magnitude is very different and the male’s number is larger for waffles and smaller for Wikipedia.

Getting Help

During this process, I often sought help by just running the usage command for the command-line. For example, waffles_transform usage shows many options. Also, the waffles_wizard command is a GUI interface to help users build a command-line string.

Files

• test.arff
• test.csv
• gender.twt
• gender.arff
• gender.csv

I saw one post that claimed that changing the appearance settings in Linux will help ( System -> Preferences -> Appearance). This post claimed the utility is at /usr/bin/gnome-appearance-properties. In RedHat, I had some success with gnome-font-properties. I still had to change the font sizes in Eclipse at Window|Preferences|General|Appearance|Colors and Fonts.

Here are the settings that seemed to look nice (taken from this post):

Application/Document/Desktop font: Liberation Sans, size 8
Window title font: Liberation Sans Bold, size 10
Fixed with font: Liberation Mono, 10


Font rendering: Best contrast (the others left weird artifacts, but mostly up to taste here I think)

Under details:
Resolution: 99 dpi (to avoid the letters to get too close, need at least 1 px between them ;)
Smoothing: Grayscale

Hinting: Full

I also found this post which will give better colors. But it does not work well as sometimes things get very hard to read.

In Ubuntu

I also found this post, which recommends creating a file in your home directory to override the default settings in Linux. Then, all of the other windows can stay large. Here’s how:

• Create a file ~/.gtkrc-eclipse:

  style “eclipse” {
      font_name = “Sans Condensed 8”
  }
  class “GtkWidget” style “eclipse”
  

• Launch eclipse with the GTK2_RC_FILES environment variable set to the path to ~/.gktrc-eclipse. Like so: env GTK2_RC_FILES=$GTK2_RC_FILES:~/.gtkrc-eclipse eclipse
• It’s still necessary to change the font sizes in Window|Preferences|General|Appearance|Colors and Fonts.

RedHat

I was able to get a clean look in RedHat as well. I created a ~/.gtkrc-2.0 with the following:

style “gtkcompact”{
font_name=“Sans 8”
GtkButton::default_border={0,0,0,0}
GtkButton::default_outside_border={0,0,0,0}
GtkButtonBox::child_min_width=0
GtkButtonBox::child_min_heigth=0
GtkButtonBox::child_internal_pad_x=0
GtkButtonBox::child_internal_pad_y=0
GtkMenu::vertical-padding=1
GtkMenuBar::internal_padding=0
GtkMenuItem::horizontal_padding=4
GtkOptionMenu::indicator_size=0
GtkOptionMenu::indicator_spacing=0
GtkPaned::handle_size=4
GtkRange::trough_border=0
GtkRange::stepper_spacing=0
GtkScale::value_spacing=0
GtkScrolledWindow::scrollbar_spacing=0
GtkExpander::expander_size=10
GtkExpander::expander_spacing=0
GtkTreeView::vertical-separator=0
GtkTreeView::horizontal-separator=0
GtkTreeView::expander-size=8
GtkTreeView::fixed-height-mode=TRUE
GtkWidget::focus_padding=0
}
class “GtkWidget” style “gtkcompact”

It made it much cleaner. I did not have to set any environment variables or even restart X. It worked immediately after launching Eclipse again. This may work in other distros, but I haven’t tried.

References

• http://blog.xam.dk/?p=70

I installed jEdit on Windows 7 and it is trying to save backups of my untitled files to C:\Program Files\jEdit. I get an error that reads:

Cannot save: java.io.FileNotFoundException: C:\Program Files\jEdit\# Untitled-1#

Solution

It’s a known issue and a defect is filed.

Notes

Set JAVA_HOME is set.

I connect to a Windows machine from Linux using rdesktop. There are warnings or errors, but I can’t copy or paste between the two machines. For example, pasting a URL from Linux to a Firefox browser on a Windows machine will not paste anything and even locks up Firefox. Using the clipboard within the same machine works fine.

Solution

Add the following line to the .rdp file for that system:

clipboard:CLIPBOARD

Note, .rpd files are all in the ~/.tsclient directory.

Be sure that the script itself didn’t finish. If so, end the script with a wait command.

When coming out of sleep or during times of intensive processing and network activity, the following symptoms occur:

1. The WiFi connection stops working.
2. top shows that phy0 spikes the CPU.
3. dmesg shows several lines of ath5k phy0: gain calibration timeout and ath5k phy0: noise floor calibration timeout

Solution

Downgrade the kernel. This problem can be solved by using the 2.6.31 kernel. Here’s how to install it:

1. Do a sudo aptitiude search linux-image.
2. Install an older kernel. For Ubuntu 10.4, I selected the 2.6.31-11-rt kernel by doing a sudo apt-get install linux-image-2.6.31-11-rt.
3. Reboot.
4. When the grub menu appears select the old kernel.
5. Verify that the old kernel is fully functional for your system.

Temporary Solutions

1. Do a pm-suspend-hybrid several times until the wifi works again.
2. Restart the machine several times until the wifi works again.
3. Do a modprobe -r ath5k and use a different network card.
4. Using a 2.6.35 kernel reduces the frequency this problem occurs.

Details

uname -r

2.6.32-26-generic

lspci|grep Wireless shows:

14:00.0 Ethernet controller: Atheros Communications Inc. AR5001 Wireless Network Adapter (rev 01)

dmesg |grep "ath5.*chip"

ath5k phy0: Atheros AR2425 chip found (MAC: 0xe2, PHY: 0x70)

modinfo ath5k shows:

filename:       /lib/modules/2.6.32-26-generic/kernel/drivers/net/wireless/ath/ath5k/ath5k.ko
version:        0.6.0 (EXPERIMENTAL)
license:        Dual BSD/GPL
description:    Support for 5xxx series of Atheros 802.11 wireless LAN cards.
author:         Nick Kossifidis
author:         Jiri Slaby
srcversion:     5A81CAB958F60B02DE47E18
alias:          pci:v0000168Cd0000001Dsv*sd*bc*sci
alias:          pci:v0000168Cd0000001Csv*sd*bc*sci
alias:          pci:v0000168Cd0000001Bsv*sd*bc*sci
alias:          pci:v0000168Cd0000001Asv*sd*bc*sci
alias:          pci:v0000168Cd00000019sv*sd*bc*sci
alias:          pci:v0000168Cd00000018sv*sd*bc*sci
alias:          pci:v0000168Cd00000017sv*sd*bc*sci
alias:          pci:v0000168Cd00000016sv*sd*bc*sci
alias:          pci:v0000168Cd00000015sv*sd*bc*sci
alias:          pci:v0000168Cd00000014sv*sd*bc*sci
alias:          pci:v0000168Cd00001014sv*sd*bc*sci
alias:          pci:v000010B7d00000013sv*sd*bc*sci
alias:          pci:v0000A727d00000013sv*sd*bc*sci
alias:          pci:v0000168Cd00000013sv*sd*bc*sci
alias:          pci:v0000168Cd00000012sv*sd*bc*sci
alias:          pci:v0000168Cd00000011sv*sd*bc*sci
alias:          pci:v0000168Cd00000007sv*sd*bc*sci
alias:          pci:v0000168Cd00000207sv*sd*bc*sci
depends:        mac80211,led-class,cfg80211,ath
vermagic:       2.6.32-26-generic SMP mod_unload modversions
parm:           nohwcrypt:Disable hardware encryption. (bool)
parm:           all_channels:Expose all channels the device can use. (bool)

Can't call method "addr" on an undefined value at /usr/bin/wakeonlan line 117.

The port and ip address can’t be combined (eg myip.com:0). Instead, use -p for port number. It should instead say:

wakeonlan -i myip.com -p 9 6c:f0:49:50:36:fe and no error will happen.

Use tab-completion!

lsmod

Get a mapping of which hardware is using which driver:

lpci -v

also

lshw

Get the version and other info of a module:

modinfo <module>

Load a module (and its dependencies):

modprobe -v <module>

Unload a module (and its dependencies):

modprobe -r <module>

Then, run sudo /usr/sbin/update-grub.

Stop recording: C-x )

Apply to region: M-x apply-macro-to-region-lines

Edit postgresql.conf (usually in /etc/postgresql/8.4/main/) to contain these lines:

log_statement=‘all’
log_destination=‘csvlog’
logging_collector=on

Now, the log should appear at /etc/postgresql/8.4/main/pg_log/*.csv

Then, when I use xxd to get a hexdump of the file I see something like this:

0000000: efbb bf2d 2d20 2449 6424 0a2d 2d20 2448 ...-- $Id$.-- $H

Notice how it starts with a efbb bf. These three bytes tell most programs that the file is unicode. It’s called a Byte Order Mark or BOM. Many editors, like emacs, will automatically write these annoying characters and you can’t always see them in your file..unless you do this…

To make emacs stop writing these characters, it’s necessary to change the coding system of emacs. Here’s how:

Do a C-x RET r and select raw-text. The raw-text encoding in emacs allows you to see the hidden characters. However, there are times when you may not want to see these characters..in such a case use utf-8-with-signature.

dd if=/dev/zero of=filename bs=1 count=1 seek=$((100*1024*1024*1024))

This creates what is called a “sparse file”. However, df will not recognize the reduction in size. From what I can tell, there is no way to get df to recognize sparse files. du and df will always disagree.

root      3513  0.0  0.0   7196  1068 ?        Ss   Nov30   0:00 /usr/sbin/sshd
root      5201  0.0  0.0  10092  2968 ?        Ss   Nov30   0:00  _ sshd: user [priv]
user      5207  0.0  0.0  11944  3620 ?        S    Nov30   0:11  |   _ sshd: user@pts/1
user      5208  0.0  0.0   5724  2584 pts/1    Ss   Nov30   0:02  |       _ -bash
user     14745  0.0  0.0   5724  1524 pts/1    S    16:42   0:00  |           _ -bash
user     14746  0.0  0.0   3800   500 pts/1    S    16:42   0:00  |           |   _ sleep 100
user     15370  0.0  0.0   5724  1524 pts/1    S    16:52   0:00  |           _ -bash
user     15371  0.0  0.0   3800   472 pts/1    S    16:52   0:00  |           |   _ sleep 100
user     15596  0.0  0.0   4604  1088 pts/1    R+   16:56   0:00  |           _ ps auxf
root      5597  0.0  0.0  10092  2976 ?        Ss   Nov30   0:00  _ sshd: user [priv]
user      5600  0.0  0.0  10244  1844 ?        S    Nov30   0:00  |   _ sshd: user@pts/2
user      5601  0.0  0.0   5720  2536 pts/2    Ss+  Nov30   0:00  |       _ -bash
root      8833  0.0  0.0  10092  2976 ?        Ss   15:17   0:00  _ sshd: user [priv]
user      8835  0.0  0.0  10092  1824 ?        S    15:17   0:00      _ sshd: user@pts/3
root      8836  0.0  0.0   4632  1368 pts/3    Ss+  15:17   0:00          _ /bin/sh

You can see three pseudo terminals created by sshd and a few bash scripts that are running.

You have a remote machine that has an X11-based program that you want to open in Windows.

Execution

Follow these steps to get it to work:

# A bunch of status will print to stdout.
XWin -multiwindow &

# This is case sensitive.
export DISPLAY=:0.0

# You should see something like "debug1: Requesting X11 forwarding with authentication spoofing."
ssh -Yv myhost

Try a simple program to test it. I like to use xterm.

SET GLOBAL general_log = 'ON';

Then run tail:

tail -f /var/lib/mysql/hostname.log

where hostname is the hostname of the database server.

Of course, you can turn it off by:

SET GLOBAL general_log = 'OFF';

References

• http://dev.mysql.com/doc/refman/5.1/en/query-log.html

This will create a 5GB file called myfile. Notice the count=0 and seek=... commands. Just seek the size you need. ls -l should show the correct size. I hear that sometimes Linux can’t fully recognize the size until the file is used. If that is the case try file myfile.

sudo /etc/init.d/network-manager restart

Just do a search here. Look down a ways on the page and there is a search engine form that is very helpful. Here is an example of pdflatex.

netstat -tlp

lsof -iTCP -s TCP:LISTEN

exec 3< <(echo hi)
cat <&3
exec 3>&-

Read and write from a pipe using fd=4:

exec 4> >(cat)
echo hi >&4
exec 4>&-

Check the current file descriptors for current shell:

lsof -p $$

Keep the mysql connection open using fd=3 for writing:

exec 3> >(mysql)
echo "SELECT 1;" >&3
echo "SELECT 2;" >&3
exec 3>&-

Keep the mysql connection open using fd=3 for reading:

exec 3< <(echo "SELECT 1;"|mysql|sed '1d')
while read <&3
do
  echo $REPLY
done

Note: sed '1d' removes the header.

Split input:

exec 3> >(awk '{print 3,$1}')
exec 4> >(awk '{print 4,$1}')
echo hi|tee >(cat >&3) >(cat >&4)|cat >/dev/null

produces:

3 hi
4 hi

MySQL’s timezone is stored in time_zone. Use the following to discover it:

"SHOW VARIABLES LIKE 'time_zone'

It is possible to change the timezone for a session using the following syntax:

SET time_zone = '+0:00'

Even further, it’s possible to set the timezone globally using this syntax:

SET GLOBAL time_zone = '+0:00'

References

• http://dev.mysql.com/doc/refman/5.1/en/time-zone-support.html

#!/bin/bash
#$URL$
#$Id$

# Enable debugging
# set -v
# set -x

# Exit if any undefined variable is used.
set -u

# Exit this script if it any subprocess exits non-zero.
set -e

# If any process in a pipeline fails, the return value is a failure.
set -o pipefail

Note that $URL$ and $Id$ are meant for subversion.

I have found a subset of people love to have their taskbar on the left vertical edge of the screen. If you have several windows open at any given time, this will help you a lot. Here’s an example:

This behavior is supported natively by Windows (at least since XP or even earlier). Linux, at least Ubuntu, does not easily support this feature. You are welcome to try it, there are all types of bugs related to it.

The Solution

I tried many solutions and even other distros, and found a very easy solution, Avant Window Navigator. This little utility has surpassed them all. Here’s an example.

References:

• http://brainstorm.ubuntu.com/item/1906/

To enable these for an individual file, type this:

svn propset svn:keywords "Id URL" myfile

That’s a lot of typing for large projects and a lot of upkeep. To enable these for certain file extensions that you svn add, type the following in ~/.subversion/config

[miscellany]
enable-auto-props = yes

[auto-props]
*.cpp = svn:keywords=Id URL
*.sh = svn:keywords=Id URL;svn:executable
...

This file likely exists with a bunch of commented out stuff. BE SURE THAT THE svn:keywords IN [auto-props] DOES NOT HAVE QUOTES.

1. Start | All Programs | Accessories
2. Right click Command Prompt.
3. Run as Administrator
4. Type sfc /scannow
5. Reboot.

The scan will take about ten minutes or so. After doing this, the Windows core files should work again. It will give a link to a log that might be helpful if sfc didn’t work as expected.

References

• http://support.microsoft.com/kb/929833

pg_dump -hlocalhost -Upostgres -a -t table database|bzip2 > tabledata.bz2

To restore:

bzip2 -c -d tabledata.bz2 | psql -hlocalhost -Upostgres database

First, set $x

x='-e h\ti'

Now look at these results

Explanation

I haven’t found an easier way, but several harder ways to do this. The harder ways often include some sort of loop or creating a new array with the same name and appending the new value to the old values.

References:

• http://www.poundbangwhack.com/2010/02/04/how-to-append-values-to-an-array-in-bash/

Install mldonkey:

apt-get install mldonkey-server

I saw this:

*** WARNING: ucf was run from a maintainer script that uses debconf, but
             the script did not pass –debconf-ok to ucf. The maintainer
             script should be fixed to not stop debconf before calling ucf,
             and pass it this parameter. For now, ucf will revert to using
             old-style, non-debconf prompting. Ugh!


         Please inform the package maintainer about this problem. 

It didn’t seem critical, so I continued.

Initial Rejection

The install went okay, but when I try to access it through telnet, I get this:

$ telnet localhost 4000
Trying 127.0.0.1…
Connected to localhost.
Escape character is ‘^]’.
Telnet connection from 127.0.0.1 rejected (see allowed_ips setting)
Connection closed by foreign host.

This was fixed by a simple restart:

sudo /etc/init.d/mldonkey-server restart

Then, I got this:

$ telnet localhost 4000
Trying 127.0.0.1…
Connected to localhost.
Escape character is ‘^]’.
Welcome to MLDonkey 2.9.5
Welcome on mldonkey command-line


Use ? for help

MLdonkey command-line:
>

Note About Configuration Files

After some research, I found that the allowed_ips is a parameter in a file called download.ini. There was no /etc/mldonkey. Instead, I did a lot of digging and found /etc/default/mldonkey-server, which has a value called MLDONKEY_DIR which was set to /var/lib/mldonkey. So, the file I needed was at /var/lib/mldonkey/downloads.ini. However, do not touch this file. The system is designed to be configured through telnet or the web interface. When I changed these files directly, the file would revert to the old values.

Changing allowed_ips

So I just went to the telnet prompt and changed the allowed_ips value directly:

$ telnet localhost 4000
Trying 127.0.0.1…
Connected to localhost.
Escape character is ‘^]’.
Welcome to MLDonkey 2.9.5
Welcome on mldonkey command-line


Use ? for help

MLdonkey command-line:
> set allowed_ips 192.168.0.0/24
option allowed_ips value changed

Then, the web interface became available to the entire network at http://my-machine:4080

Here’s how to wake a system up remotely.

Inside the Machine

First, be sure the BIOS is properly configured for WoL. After that, run the following command to see if it’s supported by your card:

sudo ethtool eth0

The result should look something like this:

Settings for eth0:
        Supported ports: [ TP ]
        Supported link modes:   10baseT/Half 10baseT/Full
                                100baseT/Half 100baseT/Full
        Supports auto-negotiation: Yes
        Advertised link modes:  Not reported
        Advertised pause frame use: No
        Advertised auto-negotiation: Yes
        Link partner advertised link modes:  Not reported
        Link partner advertised pause frame use: No
        Link partner advertised auto-negotiation: No
        Speed: 100Mb/s
        Duplex: Full
        Port: Twisted Pair
        PHYAD: 0
        Transceiver: internal
        Auto-negotiation: on
        MDI-X: Unknown
        Supports Wake-on: pg
        Wake-on: d
        Current message level: 0x0000003f (63)
        Link detected: yes

Notice the Supports Wake-on and Wake-on properties. Here’s how to decode them:

              p  Wake on phy activity
              u  Wake on unicast messages
              m  Wake on multicast messages
              b  Wake on broadcast messages
              a  Wake on ARP
              g  Wake on MagicPacket™
              s  Enable SecureOn™ password for MagicPacket™
              d  Disable (wake on nothing).  This option clears all previous options.

For this tutorial, I just use the g option. By default, it’s usually set to d. Just type this to change it to g:

sudo ethtool -s eth0 wol g

Run sudo ethtool eth0 again and check Wake-on to see if it worked.

Inside the LAN

After your BIOS is properly configured, turn off the machine and try sending the wake message from inside the LAN. Use the wakeonlan utility. It usually does not come by default but is often available by apt-get install wakeonlan.

You’ll need the system’s MAC ID, by typing ifconfig and you should see something like:

eth0      Link encap:Ethernet  HWaddr 00:12:34:56:77:88

          inet addr:192.168.0.206  Bcast:192.168.0.255  Mask:255.255.255.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:121930 errors:0 dropped:0 overruns:0 frame:0
          TX packets:100023 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3836347 (3.8 MB)  TX bytes:1836 (1.8 KB)
          Interrupt:27 Base address:0xc000


lo        Link encap:Local Loopback

          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1⁄128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:710806 errors:0 dropped:0 overruns:0 frame:0
          TX packets:710806 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:66446307 (66.4 MB)  TX bytes:66446307 (66.4 MB)

wlan0     Link encap:Ethernet  HWaddr 00:11:22:33:45:66

          inet addr:192.168.0.207  Bcast:192.168.0.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1238388 errors:0 dropped:0 overruns:0 frame:0
          TX packets:924491 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:943593318 (943.5 MB)  TX bytes:124590574 (124.5 MB)

The following line gives the MAC ID:

eth0 Link encap:Ethernet HWaddr 00:12:34:56:77:88

HWaddr is what you want. eth0 is the wired connection and wlan0 is wireless. Likely, wireless will NOT work.

Now write down the HWaddr and halt the target machine. Then, type this:

wakeonlan 00:12:34:56:77:88

The machine should turn itself on in a matter of minutes.

WoL over the Internet

Now, try to monitor for the packet with this command:

sudo tcpdump -i eth0 udp port 9

Run wakeonlan locally and you should see something like this:

22:42:22.017251 IP yourmachine.local.52793 > 255.255.255.255.discard: UDP, length 102

It should happen every time you execute the wakeonlan command. Keep this running. It will tell us if the packet is actually getting through without having to constantly power down the target machine.

Now find your external IP address. I like using http://whatsmyipaddress.com. Now try to send a packet to your target IP address (eg 1.2.3.4 shown below) from your external Internet machine:

wakeonlan -i 1.2.3.4 00:12:34:56:77:88

If you see the packet in tcpdump, great. If not, go to the next section.

Getting WoL through a Firewall

Keep the tcpdump mentioned above running. Then, go into the router configuration and set udp port 9 traffic to be sent to the IP address 255.255.255.255. This is the “broadcast” address. This is great for home use and testing, but not a good idea in a big organization as every IP on your LAN will see the packet.

Now, try the wakeonlan again. Be sure tcpdump caught it. If not, check your router configuration again.

References:

• http://ubuntuforums.org/showthread.php?t=1362739

Warning: No xauth data; using fake authentication data for X11 forwarding.

To fix it, simply type this on the local machine:

xauth add :0 . `mcookie`

Then, type the following to verify: xauth list

You should see something like:

user/unix:0  MIT-MAGIC-COOKIE-1  abcdef1234567890abcdef1234567890a

ssh-agent bash

then

ssh-add

or maybe

ssh-add /home/user/.ssh/id_dsa

Long Version

Before trying anything, be sure that the communication between to the two hosts is using keys. Type this:

ssh target

you should see this:

Enter passphrase for key '/home/user/.ssh/id_rsa':

not

user@target's password:

If you get prompted for a passphrase instead of a password, go here, but be sure to add a passphrase. It is extremely dangerous to use a private key that doesn’t have a passphrase. If anyone gets access to that private key, they can use it. However, adding a passphrase initially brings you back to the problem of asking for the passphrase several times. ssh-agent can fix this problem. To get started, simply type:

ssh-agent bash

This creates a new bash process that allows you to add private keys. When adding a new private key you will be prompted for the passphrase once and only once. Do that by typing:

ssh-add

Then, the key at ~/.ssh/id_dsa will be added and you should not get prompted for a passphrase. Then, type exit to have the OS forget your passphrase.

To verify that your key has been added, type:

ssh-add -l

It should show you the fingerprints and filenames of all keys in the agent session.

Using an ssh-agent in a script

Use the following bash code to reuse and ssh-agent in a script:

#!/bin/bash

tempfile=/tmp/ssh-agent.test

# Check for an existing ssh-agent
if [ -e $tempfile ]
then
    echo "Examining old ssh-agent"
    . $tempfile
fi

# See if the agent is still working
ssh-add -l > /dev/null

# If it's not working yet, just start a new one.
if [ $? != 0 ]
then
    echo "Old ssh-agent is dead..creating new agent."

    # Create a new ssh-agent if needed
    ssh-agent -s > $tempfile
    . $tempfile

    # Add the key
    ssh-add
fi    

# Show the user which keys are being used.
ssh-add -l

ssh-agent -s creates an ssh-agent and prints out three lines that basically set the appropriate environment variables for ssh-add to function properly. This script saves the output of ssh-agent -s to a known file location that can be reused each time the script is run. It also detects if the ssh-agent is no longer working. If so, it simply launches a new agent and adds the key.

The nice part about this script is that it will work if a script is sudo’d. However, the file will get a permission’s error if the script is run with sudo and then run without sudo.

/sudo:myserver:…

Debugger entered–Lisp error: (file-error “Host myserver' looks like a remote host,sudo’ can only use the local host”)

Well, there is a way to tell Emacs to get to that server and open the file with a sudo. Add this to the .emacs file:

(add-to-list ‘tramp-default-proxies-alist ‘(“.*” “`root\‘” “/ssh:%h:“))

Restart Emacs and /sudo will start working on remote sites.

This only seems to work in Cygwin Emacs. I tried in Ubuntu and got:

tramp-compute-multi-hops: Host solomonson.com' looks like a remote host,sudo’ can only use the local host

After much experimentation, I found that this works on Ubuntu:

(set-default ‘tramp-default-proxies-alist (quote ((“.*” “\`root\’” “/ssh:%h:“))))

See also this file in your Linux install: /etc/protocols

grep flags /proc/cpuinfo

It might looks something like:

flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr sse4_1 lahf_lm
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr sse4_1 lahf_lm
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr sse4_1 lahf_lm
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr sse4_1 lahf_lm

The lm attribute indicates the processor is 64-bit.

If it’s a windows machine, use SystemRescueCD to boot it into Linux.

I found several entries on the EmacsWiki TRAMP page talking about how version control gets in the way. But this file and directory had no source control. Then, I found this:

http://www.fifi.org/doc/tramp/tramp-emacs.html

It mentions that ssh actually uses a perl script to do its dirty work. My guess is that the perl script has a few issues. So instead, I used scp and it was more than a 1000 times faster. I have yet to see any disadvantage.

I also found this interesting article:

http://stackoverflow.com/questions/148578/using-emacs-tramp-vs-rsync-for-remote-development

Haven’t tried these.

Using Festival Text-to-Speech in Ubuntu doesn’t work after the install. Here are some steps I took to fix it. Also, some changes to make it useful in everyday work.

Getting Festival to Work

Festival is the free text-to-speech engine that is extremely popular.

Here’s how to get it:

sudo apt-get install festival

Here’s how to test it:

echo "hello world"|festival --tts

You may see this error:

Linux: can't open /dev/dsp

If so, add the following lines to your .festivalrc file:

(Parameter.set 'Audio_Command "aplay -q -c 1 -t raw -f s16 -r $SR $FILE")
(Parameter.set 'Audio_Method 'Audio_Command)

Getting it to read the clipboard

Now, if you want it to read info from your clipboard, install this:

sudo apt-get install xclip

And type this:

xclip -o|festival --tts

Now, you can go a step further and create a shortcut key for reading text. Here’s a good one:

#!/bin/bash

# This script reads the information from the clipboard outloud.

# Look for festival being run.
running=$(pgrep festival)

if [ -z $running ]
then
    # read it
    xclip -o|festival --tts
else
    # kill it
    killall festival;killall aplay;sleep .1;killall aplay
fi

I call it talk.sh. Be sure to do a chmod +x talk.sh to it.

Assigning a Shortcut

Now, to assign to a shortcut key. I’m using Ubuntu which uses GNOME. if you use something else..you’re on your own. Otherwise, click System->Keyboard Shortcuts. Then add the path to the script and assign a shortcut.

I assigned it to the Windows-A keystroke. You can click it once to start and again to stop. Unfortunately, the script assumes you only have one instance of festival.

Adjusting the Playback Speed

If you want it to read faster, change the .festivalrc file:

(Parameter.set 'Audio_Command "aplay -q -c 1 -t raw -f s16 -r $(($SR*140/100)) $FILE")

The ¹⁴⁰⁄₁₀₀ means 140% of original speed which seems about right to me for most texts.

Improving Voice (Ubuntu 12)

The default voices in Festival do not sound great. Try downloading cmu_us_slt_arctichts.tar.gz. This is tar.gz file. Here’s how to install it:

1. tar xvzf cmu_us_slt_arctichts.tar.gz
2. sudo mv cmu_us_slt_arctic_hts /usr/share/festival/voices/english/

3. Update /usr/share/festival/voices.scm to have cmu_us_slt_arctic_hts at the top (Hint: look for the work “kal” in the file)

   (defvar default-voice-priority-list
   '(cmu_us_slt_arctic_hts
   kal_diphone
   cmu_us_bdl_arctic_hts
   cmu_us_jmk_arctic_hts
   ...

Improving Voices (Ubuntu 10)

The above instructions may not work in Ubuntu 10..instead, try this:

Here’s a bash script to add new voices. These are the best I could find anywhere:

# Setup
cd
dir=nitech_us
mkdir $dir
cd $dir

# Download the voices
for voice in awb bdl clb rms slt jmk
do
  wget "http://hts.sp.nitech.ac.jp/archives/2.0.1/festvox_nitech_us_"$voice"_arctic_hts-2.0.1.tar.bz2"
done

# Unpack
tar xvf *.bz2

# Install
sudo mkdir -p /usr/share/festival/voices/us
sudo mv lib/voices/us/* /usr/share/festival/voices/us/
sudo mv lib/hts.scm /usr/share/festival/hts.scm

Setting a Default Voice

The default voice in Festival is configurable, but it doesn’t seem to work. It was necessary to change /usr/share/festival/voices.scm directly. Simply update the default-voice-priority-list. It should like something like this:

(defvar default-voice-priority-list
'(nitech_us_slt_arctic_hts
kal_diphone
cmu_us_bdl_arctic_hts
cmu_us_jmk_arctic_hts
cmu_us_slt_arctic_hts
cmu_us_awb_arctic_hts
; cstr_rpx_nina_multisyn ; restricted license (lexicon)
; cstr_rpx_jon_multisyn ; restricted license (lexicon)
; cstr_edi_awb_arctic_multisyn ; restricted license (lexicon)
; cstr_us_awb_arctic_multisyn
ked_diphone
don_diphone
rab_diphone
en1_mbrola
us1_mbrola
us2_mbrola
us3_mbrola
gsw_diphone ;; not publically distributed
el_diphone
)
"default-voice-priority-list
List of voice names. The first of them available becomes the default voice.")

Notice how I put nitech_us_slt_arctic_hts at the top. This is my favorite voice.

References

• http://ubuntuforums.org/showthread.php?t=677277
• http://ubuntuforums.org/showthread.php?t=751169
• http://ubuntuforums.org/showpost.php?p=9516900&postcount=3
• http://ubuntuforums.org/showthread.php?t=751169&page=15

$ echo -n "encode me" | openssl enc -base64

Note: -n removes the newline character added by “echo”.

Decode a base64 to a string:

$ echo "ZW5jb2RlIG1lCg==" | openssl enc -base64 -d

libapache2-mod-php5: Depends: libc6 (>= 2.7-1) but 2.3.6.ds1-13etch10+b1 is to be installed
Depends: libdb4.6 but it is not installable
Depends: libkrb53 (>= 1.6.dfsg.2) but 1.4.4-7etch6 is to be installed
Depends: libssl0.9.8 (>= 0.9.8f-5) but 0.9.8c-4etch4 is to be installed
Depends: libxml2 (>= 2.6.28) but 2.6.27.dfsg-6 is to be installed
Depends: php5-common (= 5.2.13-0.dotdeb.1) but 5.2.0+dfsg-8+etch16 is to be installed.

See also

It looks like it was time to upgrade from Etch to Lenny. Here’s how:

1. Edit /etc/apt/sources.list and replace all occurrences of etch with lenny.
2. sudo apt-get update
3. sudo apt-get dist-upgrade

But, I also had some issues with certain keys not being allowed. This command fixed it: sudo gpg --keyserver wwwkeys.eu.pgp.net --recv-keys 9AA38DCD55BE302B

I bought a new WiFi router that claimed to have special enhancements to boost the signal. I wanted to see how much of a boost was created. Here’s how it was accomplished:

Setup

I found that iwconfig prints out a lot of data. Here’s an example:

lo        no wireless extensions.


eth0      no wireless extensions.

wlan0     IEEE 802.11bg  ESSID:“…”

          Mode:Managed  Frequency:2.437 GHz  Access Point: 00:11:22:33:44:55

          Bit Rate=24 Mb/s   Tx-Power=20 dBm

          Retry  long limit:7   RTS thr:off   Fragment thr:off
          Power Management:off
          Link Quality=29⁄70  Signal level=-81 dBm

          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:0  Invalid misc:0   Missed beacon:0

I’m not sure what it all means but Link Quality and Signal level look very interesting. So, I wrote the following script:

#/bin/bash

while [ 1 ]
do
    iwconfig 2>&1|grep Link
    sleep 1
done

Now, I get something interesting every second. It looked something like this:

          Link Quality=31⁄70  Signal level=-79 dBm

          Link Quality=31⁄70  Signal level=-79 dBm

          Link Quality=30⁄70  Signal level=-80 dBm

          Link Quality=29⁄70  Signal level=-81 dBm

          Link Quality=31⁄70  Signal level=-79 dBm

          Link Quality=30⁄70  Signal level=-80 dBm

          Link Quality=31⁄70  Signal level=-79 dBm

Notice that the signal is in dBm and is always negative. Also notice that Link Quality went up as signal became less negative.

The Test

I walked around the house while watching the Link Quality number change. I noticed a huge difference when a wall was in the way. It was even possible to determine the outer boundaries of the signal. There may be other ways to view this information, but this was very easy for me.

Cygwin does not have useradd or adduser. Instead, you should first create a user in Windows and run some commands in Cygwin to read the user list.

Create a Windows User

Create the user in Windows (Usually Start|Control Panel|Manage User Accounts|Add…)

Link to the Windows Users

Type mkpasswd -l > /etc/passwd and mkgroup -l > /etc/group to get all of the local users and group synced up. Otherwise, replace the -l with a -d to get all users and groups from the domain.

[ERROR] /usr/sbin/mysqld: Can't open file: './database/table.frm' (errno: 24)

errno: 24 simply means that too many files are open for the given process. There is a read-only mysql variable called open_files_limit that will show how many open files are allowed by the mysqld:

SHOW VARIABLES LIKE 'open%'

A lot systems set this to something very low, like 1024. Unfortunately, the following will NOT work:

SET open_files_limit=100000

MySQL will respond with:

ERROR 1238 (HY000): Variable 'open_files_limit' is a read only variable

However, it is possible to make a change to /etc/my.cnf. This file may not exist, if not, just create it. Be sure it has the following contents:

[mysqld]
open_files_limit = 100000

Then, be sure to restart mysql:

sudo /etc/init.d/mysql restart

Now, SHOW VARIABLES LIKE 'open%' should show 100000. The number you use may be different.

See Also

• mysqldump error 29
• How to change number of open files
• Oracle tuning: Some of these hints work for Linux

sudo mysql_install_db

However, if mysql does not start up again, check the log (usually at /var/lib/mysql/$HOSTNAME.err). Maybe it shows something like:

/usr/sbin/mysqld: Can’t find file: ‘./mysql/plugin.frm’ (errno: 13)
[ERROR] Can’t open the mysql.plugin table. Please run mysql_upgrade to create it.

The (errno: 13) is a permissions thing. So do this:

sudo chown -R mysql:mysql /var/lib/mysql

Now try to start mysql.

mplayer stereo.mp3 -af pan=1:1:0

or try

mplayer stereo.mp3 -af pan=1:0:1

to hear only the right channel. Then, to convert the file permanently:

mencoder stereo.mp3 -af pan=1:1:0 -oac copy -o mono.mp3

which will create a new file, mono.mp3, that will have only the left channel. If the file has video, try this:

mencoder stereo.wmv -af pan=1:1:0 -oac mp3lame -ovc copy -o mono.wmv

Notice the -oac mp3lame option. The -af options will be ignored with -oac copy.

Also, during the conversion, you may see a bunch of these messages:

Pos: 522.3s  14486f (60%) 2110.43fps Trem:   0min  32mb  A-V:-0.002 [117:127]
71 duplicate frame(s)!
Pos: 522.3s  14487f (60%) 2110.27fps Trem:   0min  32mb  A-V:0.001 [117:127]
30 duplicate frame(s)!
Pos: 522.4s  14488f (60%) 2110.42fps Trem:   0min  32mb  A-V:-0.000 [117:127]
31 duplicate frame(s)!
Pos: 522.4s  14489f (60%) 2110.56fps Trem:   0min  32mb  A-V:-0.001 [117:127]

These messages will not prevent the conversion from completion. Just let the file finish and it should play.

Note

Sometimes a : is used instead of an =. For example pan:1:1:0 instead of pan=1:1:0.

mencoder original.wmv -oac copy -ovc copy -audiofile dub.mp3 -o new.wmv

where: |file|description| |—|—| |original.wmv|The original video to dub.| |dub.mp3|An audio file with the new audio. This file should be the same length as the video.| |new.wmv|The final video that is created.|

There will likely be a lot of messages that look like this:

Pos: 522.3s  14486f (60%) 2110.43fps Trem:   0min  32mb  A-V:-0.002 [117:127]
71 duplicate frame(s)!
Pos: 522.3s  14487f (60%) 2110.27fps Trem:   0min  32mb  A-V:0.001 [117:127]
30 duplicate frame(s)!
Pos: 522.4s  14488f (60%) 2110.42fps Trem:   0min  32mb  A-V:-0.000 [117:127]
31 duplicate frame(s)!
Pos: 522.4s  14489f (60%) 2110.56fps Trem:   0min  32mb  A-V:-0.001 [117:127]

These are just warnings, about something that will not prevent the conversion from completion. There may be thousands of these lines, but the conversion will eventually complete. Be patient.

usermod -a -G GROUP USER

or

gpasswd -a USER GROUP

Where USER is the user you want to add and GROUP is the group the user will now be a member. If you are currently logged in as USER, Login again for the new changes to take effect. Then, verify with:

id

It should look something like:

uid=10777(user) gid=16875(group1) groups=127(group2),16839(group3) context=…

It must be MySQL 5.1 or later. Then, to see if partitioning is available:

SHOW VARIABLES LIKE '%partition%';

or

SHOW PLUGINS;

If neither of these queries mention partitions, be sure to recompile with the --with-partition option.

Setting up a table

There are several types of partitions, but LIST seems most compatible with partitioning by date. Let’s setup a core table:

CREATE TABLE core(
  core_id INT NOT NULL AUTO_INCREMENT,
  stamp DATETIME,
  data VARCHAR(80),
  PRIMARY KEY(core_id,stamp)
)

PARTITION BY LIST(TO_DAYS(stamp))(
  PARTITION core_20100715 VALUES IN (TO_DAYS('2010-07-15')),
  PARTITION core_20100716 VALUES IN (TO_DAYS('2010-07-16'))
);

Notice that the primary key must include stamp or MySQL will give an error.

Testing the partition

Now a simple insert:

INSERT INTO core (stamp,data) VALUES ('20100715','a')

This should work.

Now a simple insert outside of the existing partitions:

INSERT INTO core (stamp,data) VALUES ('20100717','x')

Gives this error:

SQL Error (1526): Table has no partition for value

Read the information_schema

It’s possible to see the existing partitions in the information_schema database:

SELECT *
FROM information_schema.PARTITIONS
WHERE TABLE_NAME='core';

Which shows us: | TABLE_CATALOG | TABLE_SCHEMA | TABLE_NAME | PARTITION_NAME | SUBPARTITION_NAME | PARTITION_ORDINAL_POSITION | SUBPARTITION_ORDINAL_POSITION | PARTITION_METHOD | SUBPARTITION_METHOD | PARTITION_EXPRESSION | SUBPARTITION_EXPRESSION | PARTITION_DESCRIPTION | TABLE_ROWS | AVG_ROW_LENGTH | DATA_LENGTH | MAX_DATA_LENGTH | INDEX_LENGTH | DATA_FREE | CREATE_TIME | UPDATE_TIME | CHECK_TIME | CHECKSUM | PARTITION_COMMENT | NODEGROUP | TABLESPACE_NAME | |—————|————–|————|—————-|——————-|—————————-|——————————-|——————|———————|———————-|————————-|———————–|————|—————-|————-|—————–|————–|———–|————-|————-|————|———-|——————-|———–|—————–| | NULL | test | core | core_20100715 | NULL | 1 | NULL | LIST | NULL | TO_DAYS(stamp) | NULL | 734333 | 0 | 0 | 16384 | NULL | 0 | 4194304 | NULL | NULL | NULL | NULL | | default | NULL | | NULL | test | core | core_20100716 | NULL | 2 | NULL | LIST | NULL | TO_DAYS(stamp) | NULL | 734334 | 0 | 0 | 16384 | NULL | 0 | 4194304 | NULL | NULL | NULL | NULL | | default | NULL |

Adding and Removing Individual Partitions

Here’s how to add additional partitions:

ALTER TABLE core
ADD PARTITION (PARTITION core_20100717 VALUES IN (TO_DAYS('2010-07-17')));

Here’s how to remove a partition:

ALTER TABLE core
DROP PARTITION core_20100717;

Adding and Removing Partitioning on Existing Tables

If the table was not originally partitioned, it is possible to add it:

ALTER TABLE core
PARTITION BY LIST(TO_DAYS(stamp))(
  PARTITION core_20100715 VALUES IN (TO_DAYS('2010-07-15')),
  PARTITION core_20100716 VALUES IN (TO_DAYS('2010-07-16'))
);

And remove it:

ALTER TABLE core
REMOVE PARTITIONING;

Adding and removing partitions doesn’t seem to change existing data on the table, but new partitions must include all existing data.

Storing a Partition on a Separate Directory

Each partition may be stored on a separate data and index directory:

ALTER TABLE core
PARTITION BY LIST(TO_DAYS(stamp))(
  PARTITION p20100715 VALUES IN (TO_DAYS('2010-07-15'))
  DATA DIRECTORY '/data/2010-07-15'
  INDEX DIRECTORY '/data/2010-07-15',
  PARTITION p20100716 VALUES IN (TO_DAYS('2010-07-16'))
  DATA DIRECTORY '/data/2010-07-16'
  INDEX DIRECTORY '/data/2010-07-16'
);

Note, DATA DIRECTORY and INDEX DIRECTORY are ignored on Windows after MySQL 5.1.24.

You may get this error: SQL Error (1210): Incorrect arguments to DATA DIRECTORY

If so, it’s likely because the DATA DIRECTORY is already being used by another part of MySQL. The partition data directories can’t be subdirectories of MySQL’s current data directory. The best thing to do is make a new directory maybe mysql_part that is easily accessible.

Partition Naming between Tables

Different tables can use the same names for their partition. Consider the following example:

CREATE TABLE core(
  core_id INT NOT NULL AUTO_INCREMENT,
  stamp DATETIME,
  data VARCHAR(80),
  PRIMARY KEY(core_id,stamp)
)
PARTITION BY LIST(TO_DAYS(stamp))(
  PARTITION p20100715 VALUES IN (TO_DAYS('2010-07-15')),
  PARTITION p20100716 VALUES IN (TO_DAYS('2010-07-16'))
);

CREATE TABLE core2(
  core_id INT NOT NULL AUTO_INCREMENT,
  stamp DATETIME,
  data VARCHAR(80),
  PRIMARY KEY(core_id,stamp)
)
PARTITION BY LIST(TO_DAYS(stamp))(
  PARTITION p20100715 VALUES IN (TO_DAYS('2010-07-15')),
  PARTITION p20100716 VALUES IN (TO_DAYS('2010-07-16'))
);

Both tables use the exact same partition names: p20100715 and p20100716. No error occurs.

See also

• MySQL 5.1 Documentation: Partitioning
• MySQL 5.1 Documentation: CREATE TABLE
• MySQL 5.1 Documentation: ALTER TABLE

Use -h127.0.0.1 instead of -hlocalhost when invoking mysql.

Details

When typing mysql on the command-line, by default it connects to a socket. This also happens when typing mysql -hlocalhost. It even happens when specifying a TCP port number mysql -hlocalhost -P3306. Simply type \s to see what server mysql is connected to. Here’s an example:

$ mysql -hlocalhost -P3306
Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 1
Server version: 5.0.77 Source distribution


Type ‘help;’ or ‘\h’ for help. Type ‘\c’ to clear the buffer.

mysql> \s

mysql  Ver 14.12 Distrib 5.0.77, for redhat-linux-gnu (x86_64) using readline 5.1

Connection id:          1
Current database:
Current user:           root@localhost
SSL:                    Not in use
Current pager:          stdout
Using outfile:          “
Using delimiter:        ;
Server version:         5.0.77 Source distribution
Protocol version:       10
Connection:             Localhost via UNIX socket
Server characterset:    latin1
Db     characterset:    latin1
Client characterset:    latin1
Conn.  characterset:    latin1
UNIX socket:            /var/lib/mysql/mysql.sock
Uptime:                 55 days 2 hours 20 min 48 sec

Threads: 2  Questions: 188848126  Slow queries: 1  Opens: 247  Flush tables: 1  Open tables: 64  Queries per second avg: 39.670

The Connection field shows that we are indeed connected to a UNIX socket. However, if we type the dotted-decimal, it is connected to a TCP port:

$ mysql -h127.0.0.1
Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 12
Server version: 5.0.77 Source distribution


Type ‘help;’ or ‘\h’ for help. Type ‘\c’ to clear the buffer.

mysql> \s

mysql  Ver 14.12 Distrib 5.0.77, for redhat-linux-gnu (x86_64) using readline 5.1

Connection id:          12
Current database:
Current user:           root@localhost
SSL:                    Not in use
Current pager:          stdout
Using outfile:          “
Using delimiter:        ;
Server version:         5.0.77 Source distribution
Protocol version:       10
Connection:             127.0.0.1 via TCP/IP
Server characterset:    latin1
Db     characterset:    latin1
Client characterset:    latin1
Conn.  characterset:    latin1
TCP port:               3306
Uptime:                 55 days 2 hours 24 min 45 sec

Threads: 2  Questions: 188848139  Slow queries: 1  Opens: 247  Flush tables: 1  Open tables: 64  Queries per second avg: 39.668

The .my.cnf file tells the mysql command-line utility how to connect to a MySQL database. It is typically stored in the home directory of the current user. Here’s a sample file:

[mysql]
# Set a larger maximum packet size for the client
max_allowed_packet=512M
user=root
password=my_secret!p@sswordIsVerylong#

Notice the plaintext password is in the file. It’s a good idea to set read-write privileges only for the user (ie chmod 600). Now the user should be able to simply type mysql at a command-line and the MySQL monitor should automatically connect to the database.

If it does not work

The following command may not work: sudo mysql

Why? Well, try this: sudo strace mysql 2>&1|grep open

And you will probably get something like:

open("/etc/ld.so.cache", O_RDONLY)      = 3
open("/usr/lib64/libncursesw.so.5", O_RDONLY) = 3
open("/lib64/libpthread.so.0", O_RDONLY) = 3
open("/lib64/libcrypt.so.1", O_RDONLY)  = 3
open("/lib64/libnsl.so.1", O_RDONLY)    = 3
open("/lib64/libm.so.6", O_RDONLY)      = 3
open("/lib64/libc.so.6", O_RDONLY)      = 3
open("/lib64/libdl.so.2", O_RDONLY)     = 3
open("/home/user/.my.cnf", O_RDONLY)  = -1 EACCES (Permission denied)
open("/etc/nsswitch.conf", O_RDONLY)    = 3
open("/etc/ld.so.cache", O_RDONLY)      = 3
open("/lib64/libnss_files.so.2", O_RDONLY) = 3
open("/etc/services", O_RDONLY)         = 3
open("/usr/share/mysql/charsets/Index.xml", O_RDONLY) = 4

Notice the EACCES line? That means that user’s .my.cnf file could not be read by root. This can happen if you secured the system properly. So, how do we get sudo to look at /root/.my.cnf?

sudo -H strace mysql 2>&1|grep open

Instead of an EACCES line, it should have something like:

open("/root/.my.cnf", O_RDONLY) = 3

Of course, the grep hides some important details and strace gives us details we no longer need, so just type this:

sudo -H mysql

Consider the following program:

$x=12345678901.234567000;
$y=($x-int($x))*1000000000;
printf("%f->%f\n",$x,$y);

Here’s what it prints:

12345678901.234568->234567642.211914

I was expecting:

12345678901.234567->234567000

This appears to be some sort of rounding issue in Perl. Adding a new package seems to help:

use bignum;
$x=12345678901.234567000;
$y=($x-int($x))*1000000000;
printf("%f->%f\n",$x,$y);
print $x;

prints:

12345678901.234568->234567000
12345678901.234567

Notice how the second line is correct. The variable is fine, but printf is having some issues.

Note that if the variable is being converted from a string there are still issues. Consider the following:

use bignum;
$x="12345678901->234567000";
$y=($x-int($x))*1000000000;
printf("%f:%f\n",$x,$y);

prints:

12345678901.234568->234567642.211914

So use +0 to cast to a number:

$x="12345678901.234567000";
$x+=0;# Cast to a number
$y=($x-int($x))*1000000000;
printf("%f->%f\n",$x,$y);

prints:

12345678901.234568->234567000.000000

MySQL

The same rounding issue also appears in MySQL. Consider the following:

SELECT ('12345678901.234567'-FLOOR('12345678901.234567'))*1000000000;

prints:

234567642.211914

Take off the single quotes and all is well:

SELECT (12345678901.234567-FLOOR(12345678901.234567))*1000000000;

prints:

234567000.000000

It appears that the string to number conversion is broken in MySQL, so you have to define it explicitly. Unfortunately, you can’t just cast to a double:

SELECT (CAST('12345678901.234567' as double)-FLOOR(CAST('12345678901.234567' as double)))*1000000000;

prints:

SQL Error (1064): You have an error in your SQL syntax;

So, it’s necessary to use a decimal, but that fails too:

SELECT (CAST('12345678901.234567' as decimal)-FLOOR(CAST('12345678901.234567' as decimal)))*1000000000;

prints:

SQL Error (1064): You have an error in your SQL syntax;

Nice. Well, I guess I need to specify the width explicitly.

SELECT (CAST('12345678901.234567' as decimal(65,30))-FLOOR(CAST('12345678901.234567' as decimal(65,30))))*1000000000;

prints:

234567000.000000000000000000000000000000

PostgreSQL

The same thing happens in PostgreSQL:

SELECT ('12345678901.234567'-FLOOR('12345678901.234567'))*1000000000;

prints:

234567642.211914

And with float casts:

SELECT ('12345678901.234567'::float-FLOOR('12345678901.234567'::float))*1000000000;

prints:

234567642.211914

But, decimal works:

SELECT ('12345678901.234567'::decimal-FLOOR('12345678901.234567'::decimal))*1000000000;

234567000

Connecting to MySQL in plaintext is dangerous. Also, leaving port 3306 open your server is not very safe. If the MySQL server has an SSH port open, it’s possible to use it for MySQL queries. Simply open a bash prompt on Cygwin and type:

ssh -N -L 5000:localhost:3306 your-server &

Of course, replace your-server with the machine name of the MySQL server. Now, it’s possible to connect to the MySQL database using 127.0.0.1:5000. Sounds strange, but port 5000 on your machine is now MySQL on the remote machine. It’s called a tunnel and it’s a great way to get around needing to make holes in your firewall. It also has the benefit of being encrypted. This should work in Linux and Cygwin (windows).

Reverse Tunnels

It’s also possible to send traffic on your remote webserver to your local machine. It looks something like this:

ssh -N -R 5000:localhost:3306 your-server &

After this command, anything going to your-server’s port 5000 will be sent to 3306 on the local machine.

Verfication and Destruction

To verify the port is open, type lsof -i in Linux or netstat -a|grep LISTENING|grep 5000 in Cygwin. These commands will give you a process id number (pid). Just type kill <pid> to end the connection. For example kill 1234 stops pid 1234.

Everything in the directory is added. Then, trying to remove before my next check-in fails: svn remove myfile.exe

However, this worked: svn revert myfile.exe

• Use System.Net.Sockets
• TcpListener listener=new TcpListener(System.Net.IPAddress.Parse("0.0.0.0"),2000);

• Create the following callback:

  protected void DoAcceptCallback(IAsyncResult ar){
  TcpListener tl = (TcpListener)ar.AsyncState;
  Socket s = tl.EndAcceptSocket(ar);
   
  //Log the traffic
  DataRow dr=traffic.NewRow();
  dr["ID"] = 0;
  dr["RemoteIP"] = s.RemoteEndPoint.ToString();
  traffic.Rows.Add(dr);
   
  s.Close();
  }
  

By default, teasers are enabled and there is not a simple WebUI way to remove these. There are two things you can do: 1. Make sure the preview of each node is blank, or 2. Change the template.

To change the template, open a file at: themes/bluemarine/node.tpl.php and find the following line:

<div class="content"><?php print $content?></div>

And change it to:

`<div class="content"><?php if (!$teaser) print $content?></div>`

The wmic NTEVENT command can be used to dump out the contents of EventViewer logs to the command-line. Here’s an example: wmic NTEVENT WHERE "LogFile='security' AND TimeGenerated > '20100709173000.000000-300'"

The TimeGenerated is NOT epoch, but is a string in the form of YYYYMMDDHHmmSS.uuuuuu-ZZZ where:

It’s also possible to add a GET to the query like so: wmic NTEVENT WHERE "LogFile='security' AND TimeGenerated > '20100709173000.000000-300'" GET TimeGenerated

And here’s how to output as a CSV file: wmic NTEVENT WHERE "LogFile='security' AND TimeGenerated > '20100709173000.000000-300'" GET TimeGenerated,User /format:csv

See also:

• http://technet.microsoft.com/en-us/library/cc784189%28WS.10%29.aspx
• http://xinn.org/misc-scripts/wmic.txt

WMITools.exe can do a lot of this stuff, but practically locks up my system.

Using Microsoft Log Parser

Microsoft has a tool called Log Parser that can can also parse the event logs.

# Use C-t instead of C-a since C-a is used in bash for "go to beginning of line."
# first letter defines the screen command key.
# second letter defines the literal insert key.
escape ^Tt

# Give 3000 lines of scrollback
scrollback 3000

# Don't annoy me with the startup message for screen.
startup_message off

# Use 256 colors
term screen-256color

# Map some handy keys
bindkey "^[[1;5D" prev  # change window with ctrl-left
bindkey "^[[1;5C" next  # change window with ctrl-right

# Setup the statusbar
hardstatus alwayslastline
hardstatus string '%{= kG}%-Lw%{= kW}%50> %n*%f %t%{= kG}%+Lw%< %{= kG}%-=%D %m/%d/%y | %C:%s %A | %1`%{-}'

Also an additional hint, to reload the .screenrc file do this: C-t : source $HOME/.screenrc

This process will copy an entire partition to a file. It will not copy an entire disk to a file. The advantage is that you can change the size of the disk and it only copies the data, not the empty bits on the disk.

Prepare

1. Download SystemRescueCd and burn it to a CD.
2. Have a ssh server available on your network.
3. If networking is not an option, consider connecting the disk to a system that can do the backup.
4. On Windows machines, be sure to run JKDefrag.
5. Especially for Windows, it seems to work best if the system shutdown was graceful. Do not try this from a hibernated system or immediately after a crash unless it’s the last resort.
6. BE CAREFUL! Probably best not to use these tools for systems with critical data. In such cases, just buy something more advanced.

Make the image:

1. Boot the source system with SystemRescueCd.
2. Enable the network. (ifconfig eth0 x.y.z.1)
3. Create a /mnt directory for the destination. (mkdir /mnt/ssh)
4. Mount ssh. I used sshfs login@ssh.server.org:/path/to/dir /mnt/ssh
5. Identify the partitions fsarchiver probe
6. Use fsarchiver. I used fsarchiver savefs -s 680 /mnt/ssh/test.fsa /dev/sda1 for CD. I used a -s 4470 for DVD’s. Note the .fsa extension. Note: process may max out for a long time but only write 243 bytes to the output file, but it will work eventually.
7. See also http://www.fsarchiver.org/QuickStart

Restore the image:

1. Boot the target system with SystemRescueCd.
2. Enable the network. ifconfig eth0 x.y.z.1
3. Create a /mnt directory for the source. Maybe mkdir /mnt/ssh.
4. Run startx to get into a UI. Then use gparted to edit partitions.
5. Mount ssh I used sshfs login@ssh.server.org:/path/to/dir /mnt/ssh
6. Identify the partitions fsarchiver probe.
7. Use fsarchiver I used fsarchiver restfs /mnt/ssh/disk.fsa id=0,dest=/dev/sda1,mkfs=reiserfs Note: this will automatically span files.
8. See also http://www.fsarchiver.org/QuickStart