Oracle: create index for column that allows NULL

CREATE INDEX table1_nullable_col1_idx
       ON table1 (nullable_col1, 1);

The ", 1" makes Oracle add into the index the record whose nullable_col1 field has a NULL value.

Oracle: update one table with joined condition from another table

UPDATE table1 
       SET t1_col1 = 'something'
       WHERE id IN (SELECT t1.id FROM table1 t1, table2 t2
                                 WHERE t1.id = t2.id
                                 AND t2.t2_col1 = 'some condition');

Adding a CA to the trust store for Eclipse

If a Eclipse plugin uses a self-signed updating web site, you may encounter the authentication failure error when updating the plugin.

The solution is to add the CA of the self-signed certificate to the trust store that Eclipse uses.

By default, Eclipse uses the Java trust store at $JAVA_HOME/lib/security/cacerts.

To list the certificates in it, go to $JAVA_HOME/lib/security/ and run command:
$ $JAVA_HOME/bin/keytool -list -keystore ./cacerts

You may not have the permission to add a new certificate into the default trust store. We can make a copy of the default trust store and add the new certificate in the new copy.
$ cd /path/to/my/trust/store/location
$ cp $JAVA_HOME/lib/security/cacerts mytruststore

The password of Java's default cacerts is "changeit". You will need to input it when adding a new certificate into mytruststore.

Run the following command to import the new certificate (e.g. cloudServicesRootCA.cer, this post shows how to download the certificate of a CA) into mytruststore:
$ $JAVA_HOME/bin/keytool -alias cloudServicesRootCA -import -file cloudServicesRootCA.cer -keystore mytruststore

Exit Eclipse if you are running it and let Eclipse know the new trust store by adding it to Eclipse's configuration file eclipse.ini which locates in the root directory of the Eclipse installation. Use an editor to open eclipse.ini and add/modify these two parameters:
-Djavax.net.ssl.trustStore=/path/to/my/trust/store/location/mytruststore
-Djavax.net.ssl.trustStorePassword=changeit

Windows 10: use MMC (Microsoft Management Console) to add a trusted CA

1. Right click on the Windows icon. Select menu Run.

2. Input command: mmc

3. In MMC, click on menu File->Add/Remove Snap-in...

4. Select Certificates in the snap-ins list; Click on the Add button. "Certificates - Current User" will be added in the "Selected snap-ins" list. Click on the OK button.

5. In the left panel, expand "Certificates - Current User"; Right click on "Trusted Root Certification Authorities"; Select "All Tasks->Import...".

6. Use the Import Wizard to browse and select the certificate to import.

Note: Only add the CA if you absolutely trust it.

Come curl command options

To access a web page:
$ curl http://www.google.com

If the web server does not provide a trusted certificate, or it is using a self-signed certificate, we can accept the it by using the -k option:
$ curl -k http://www.google.com

To turn on verbose mode:
$ curl -v -k http://www.google.com

To use HTTP 1.0:
$ curl -v -k http://www.google.com --http1.0

To remove the "Host:..." header:
$ curl -v -k http://www.google.com --http1.0 -H 'Host:'

To remove more headers:
$ curl -v -k http://www.google.com --http1.0 -H 'Host:' -H 'User-Agent:' -H 'Accept:'

IIS web application to reject HTTP requests without a Host header

Step 1:

Download and install the URL Rewrite module (https://www.iis.net/downloads/microsoft/url-rewrite).

Step 2:

Add the rewrite rule in the web.config for the web application:

<system.webServer> 
......
    <rewrite>
            <rules>
              <rule name="RequestBlockingRule1" patternSyntax="Wildcard" stopProcessing="true">
                <match url="*" />
                    <conditions>
                        <add input="{HTTP_HOST}" pattern="*.*" negate="true" />
                    </conditions>
                    <action type="AbortRequest" />
                </rule>
            </rules>
    </rewrite>
</system.webServer>

Now, any HTTP requests without a "Host:..." header will get a connection reset.

CVS: change file type from text to binary

$ cvs admin -kb my-files
$ cvs update -A my-files
$ cvs commit -m 'make them binary' my-files

To prevent IIS from showing IP in the Location header of HTTP response

Change web.config of the application to add a hostname to be shown in the Location header:

<system.webServer>
......
    <serverRuntime alternateHostName="myserver" />
......
</system.webserver>

This change may trigger an authentication error complaining about the locked section in the configuration. To fix it, run command:
> %windir%\system32\inetsrv\appcmd unlock config -section:system.webServer/serverRuntime

It unlocks section system.webServer/serverRuntime at configuration path "MACHINE/WEBROOT/APPHOST". What it actually does is changing the following file
        C:\Windows\System32\inetsrv\config\applicationHost.config
by adding
        <serverRuntime />
to the "Allow" sections:
<location path="" overrideMode="Allow">
    <system.webServer>
    ......
        <serverRuntime />
    </system.webServer>
</location>

However, "myserver" is only returned in the Location header when the request does not contain the Host header. If the request headers include the Host, the value of the Host header will be returned in the Location header.

Add a new Event Source and create Event Logs on Windows

Open a PowerShell terminal.

The New-EventLog command creates an Event Log MyLog and registers an Event Source "MyLogSource" to it:
> New-EventLog -source MyLogSource -LogName MyLog

Using the Event Viewer, you can see the Event Log is generated under the category "Applications and Services Logs".

You may use the Write-Event command to write an event to the Event Log:
> Write-EventLog -source MyLogSource -LogName MyLog -EventId 1 -Message "Hello World"

In C#, you can write an event to the above Event Log like this:
EventLog.WriteEntry("MyLogSource", "Hello World", EventLogEntryType.Information);

If you decide to delete the Event Log together with all the events in it, you can run command Remove-EventLog:
> Remove-EventLog -LogName "MyLog"

gcc: using snprintf to append strings does not work

Considering this piece of code:

#include <stdio.h>

int main()
{
  char buf[1000];

  snprintf(buf, sizeof(buf), "%s", "Hello World");
  printf("%s\n", buf);

  // trying to append 1024 to the string.
  snprintf(buf, sizeof(buf), "%s%d", buf, 1024);
  printf("%s\n", buf);
}

One may expect the second printf will output a string of "Hello World1024". However, for the latest versions of the GCC, you may get the following output from the code:

Hello World
1024

The second printf only prints out "1024".