Collection

How to change your own password in SQL Server using ALTER LOGIN

2011-10-05T03:33:00.000-07:00

This isn't documented very well. If you are using ALTER LOGIN to change your own password, you need to supply your existing password as OLD_PASSWORD:

Alter Login [my_login] WITH PASSWORD = 'newpassword' OLD_PASSWORD = 'oldpassword'

Fixed-point combinator in Java

2011-09-08T04:04:00.000-07:00

Well it's not short enough to fit in a twitter post, ala Haskell, fix f = f (fix f). But I enjoyed this post on Stream of Unconsciousness.

Running out of syslog space on sybase

2010-12-23T03:54:00.000-08:00

dump tran db_name with no_log
go

Haskell telnet client to run a command on a remote system

2009-07-05T05:30:00.001-07:00

I've written a very simple little program using the Haskell Network.Socket library, that logs in to a remote system via telnet and runs a single command. It's the first Haskell that I've written in quite a long time, and it's nothing very impressive. In fact the hard part was getting the Socket library built on Windows Vista, it was all quite simple on Ubuntu.

Anyway, here's the current version of the code:

module Main where

import Control.Concurrent
import Data.Char
import Network
import Network.Socket
import System.IO
import System(getArgs)

telnetIAC :: Int
telnetIAC = 255

telnetWONT :: Int
telnetWONT = 252

newtype Credentials = Credential (String, String)

password :: Handle -> String -> IO ()
password h pwd = do
char <- hGetChar h      
case char of
  ':' -> do hPutStrLn h pwd
            hFlush h
  otherwise -> password h pwd


login :: Handle -> Credentials -> IO ()
login h c@(Credential (user, pwd)) = do
hPutStrLn h user
hFlush h
password h pwd
putStrLn "Logged in"


declineHandshakeOption :: Handle -> Credentials -> IO ()
declineHandshakeOption h c = do
ignoreChar <- hGetChar h      
optionChar <- hGetChar h
hPutChar h $ chr telnetIAC      
hPutChar h $ chr telnetWONT      
hPutChar h $ optionChar      
hFlush h      
handshakeAndLogin h c   

handleHandshakeInput :: Handle -> Char -> Credentials -> IO()
handleHandshakeInput h char credentials
| ord char == telnetIAC        = declineHandshakeOption h credentials
| char == ':'                  = login h credentials
| otherwise                    = handshakeAndLogin h credentials


handshakeAndLogin :: Handle -> Credentials -> IO ()
handshakeAndLogin h credentials = do
char <- hGetChar h     
handleHandshakeInput h char credentials    

main :: IO ()
main = withSocketsDo $ do       
 args <- getArgs      
 addrinfos <- getAddrInfo Nothing (Just (args !! 0)) (Just "23")       
 let serveraddr = head addrinfos       
 sock <- socket (addrFamily serveraddr) Stream defaultProtocol       
 connect sock (addrAddress serveraddr)       
 putStrLn "Connected"       
 h <- socketToHandle sock ReadWriteMode       
 hSetBuffering h (BlockBuffering Nothing)       
 handshakeAndLogin h $ Credential (args !! 1, args !! 2)       
 hPutStrLn h (args !! 3)       
 hFlush h       
 putStrLn "Command sent"       
 threadDelay 1000000

To compile:
ghc -package network -o TelnetClient TelnetClient.hs

To run:
./TelnetClient localhost username password "./command.sh"

Problems building Haskell packages on Windows

2009-07-05T02:49:00.000-07:00

I made a silly mistake today, and wasted quite a long time trying to build the Haskell network package using "cabal install" via cygwin. I eventually found out that in most cases you should use MingW and MSYS rather than cygwin (http://haskell.org/haskellwiki/Windows).

Where does Eclipse store its launch configurations?

2009-06-14T22:33:00.001-07:00

It's a simple question, and the answer is out there on the net, but it took me awhile to find it. Eclipse launch configurations at the workspace level are stored in .metadata\.plugins\org.eclipse.debug.core\.launches.

I had assumed they were in org.eclipse.jdt.core and completely bypassed the debug package while searching around for it on my filesystem. Obviously that was a mistake!

Note that this refers to launch configurations saved by Eclipse as you work. You can also save your configurations to an external file, using the "common" tab of the run dialog, and saving to a "shared file".

Firefox 3 crashing when using jQuery 1.3+ and ASP.Net

2009-06-03T16:58:00.000-07:00

Well, I had a very interesting problem today. I upgraded from jQuery 1.2.6 to 1.3.2, to fix some problems using jQuery in IE8. The 1.2.x series of jQuery does browser sniffing to determine how to do certain things, and since IE8 is now "more standards-compliant", some of these hacks do not work unless the "compatibility view" is used.

jQuery 1.3.x fixed all of this by using JavaScript object detection rather than browser sniffing, so that's great.

However.. due to the design of the buttons in our application, they are in fact a div that contain both text and an ASP.Net image button. The onclick handler for the div locates the image button using a jQuery selector and clicks it via javascript.

With jQuery 1.3.2 and Firefox 3, this causes a browser crash that can't be recovered from, you have to kill the browser in task manager. The interesting thing is, the postback does occur, my ASP.Net event handler runs, but the browser crashes completely, hogging the CPU until you have to kill it. Lovely!

What I found is that if you call .click() on the DOM element directly, the problem is solved -- I was calling .click() on the collection returned by the jQuery selector.

Changed from this: $('#saveButton').click()
To this: $('#saveButton')[0].click()

phew, problem solved!

SQL tuning training slides

2009-05-26T03:14:00.000-07:00

I'm writing this post so I can share the slides for a training session I gave to some colleagues last year about SQL statement and performance tuning.

Unfortunately the examples are missing, which includes the most useful part of the presentation -- particularly, how to generate and interpret MySQL explain plans.

However, somebody might still find this useful one day.

No protocol binding matches the given address

2009-04-05T17:38:00.000-07:00

If you see the following error when trying to call a WCF web service in the ASP.Net Development Web Server...

System.ServiceModel.ServiceActivationException: No protocol binding matches the given address 'http://0.0.0.0:2825/Services/Blah.svc/jsdebug'. Protocol bindings are configured at the Site level in IIS or WAS configuration. ---> System.InvalidOperationException: No protocol binding matches the given address 'http://0.0.0.0:2825/Services/Blah.svc/jsdebug'. Protocol bindings are configured at the Site level in IIS or WAS configuration.
--- End of inner exception stack trace ---
at System.ServiceModel.AsyncResult.End[TAsyncResult](IAsyncResult result)
at System.ServiceModel.Activation.HostedHttpRequestAsyncResult.End(IAsyncResult result)
at System.ServiceModel.Activation.HostedHttpRequestAsyncResult.ExecuteSynchronous(HttpApplication context, Boolean flowContext)
at System.ServiceModel.Activation.HttpHandler.ProcessRequest(HttpContext context)
at System.Web.HttpApplication.CallHandlerExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute()
at System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously)

Try rebooting your machine. Yes, seriously. I found no other information to help me solve this problem, and a simple reboot fixed it. A colleague of mine has also had the same problem before, with the same "fix".

I miss Java.

Maven Jetty Plugin - Embedded error: Bad temp directory

2008-08-31T15:56:00.000-07:00

For the record, this error is caused by running mvn jetty:run-war from the top-level directory of a modular project, rather than your web project directory.

http://markmail.org/message/ubjlpzebsgv5idxz

Making sure that GZIP encoding is turned on, using wget

2008-05-27T15:12:00.000-07:00

Something I've had to do many times, but I always seem to forget the syntax:

wget --header="accept-encoding: gzip" "http://"

If the saved response you get back is in binary gzip format instead of text/html, then you know that your gzip encoding is turned on and functioning correctly.

org.jboss.cache.CacheException: failed to get global transaction

2008-04-22T02:41:00.000-07:00

Today I found out the hard way that JBoss TreeCache as a second-level cache provider in Hibernate can only be used with the READ_ONLY and TRANSACTIONAL caching strategies. This is clearly listed in the Hibernate documentation, but when you are using the EJB3 Hibernate EntityManager on top of Hibernate it becomes easy to lose track of these kinds of things.

I was receiving the above exception in a BMT Stateful Session Bean when it tried to commit the UserTransaction. I attempted to fix it with a bit of worthwhile refactoring into CMT with a RequiresNew method, and was sure the problem would be solved -- no such luck.

I eventually stumbled on the relevant bit of the Hibernate docs, but it still didn't click until I remembered back months ago when I had studiously annotated many of our entities with @Cache(usage = CacheConcurrencyStrategy.NONSTRICT_READ_WRITE). Replacing this with the TRANSACTIONAL strategy had me smiling in no time.

w00t!

However, I really am starting to think that I should have stuck with ehcache. I'll see how things go for now, and how painful the clustering setup becomes when we decide to test it.