BowerStudios.com

I wanted to schedule a job such that it would start only after it had finished rather than x minutes after it finished. I was surprised that quartz doesn't seem to have an option to do this. My guess is that they expect you to fire the job off asynchronously instead of blocking.

However there is a way to do this, though it is not as elegant as it could be:

1. Run the job every x minutes
2. Make sure the Quartz / Spring Batch Jobs are blocking
3. Set the Quartz Jobs to be non-concurrent - When a new one tries to start, it will only do so if a previous one is not running.

The Grails Spring Batch Plugin by default starts jobs asynchronously so that you can start jobs from the ui and not tie it up ( they are non-blocking). I opted to add a synchronous job launcher to my resources.groovy, so that I could use one for the ui and one for quartz, and block it if I liked:

syncJobLauncher(SimpleJobLauncher){
jobRepository = ref("jobRepository")
taskExecutor = { org.springframework.core.task.SyncTaskExecutor executor -> }
}

I use a service to interact with Spring Batch, choosing whether I want to run the job synchronously or not

class EtlService {
//Transactions are handled within Spring Batch
static transactional = false
//
//Asynchronous Job launcher - default for Spring Batch Grails Plugin - used for controllers
JobLauncher jobLauncher
//Synchronous Job launcher - used for quartz jobs
JobLauncher syncJobLauncher
//
JobRepository jobRepository
//
/**
* The names of jobs created via the grails-app/batch/*BatchConfig.groovy
* definitions should be injected here
*/
def myFirstJob
def mySecondJob
//
String launchJob(String jobName, boolean async = true) {
Job job = null
switch(jobName){
case 'myFirstJob':
job = myFirstJob
break
case 'mySecondJob':
job = mySecondJob
break
default:
return "Did not find Batch Job: $jobName"
}
JobLauncher selectedLauncher = async ? jobLauncher : syncJobLauncher
selectedLauncher.run(job, new JobParameters([date: new JobParameter(new Date())]))
return "Batch Job($jobName) launched from EtlService"
}
}

The Job:
class MyFirstJob {
private static final Logger LOGGER = LoggerFactory.getLogger(MyFirstJob)
EtlService etlService
def concurrent = false
static triggers = {
simple name: 'myFirstJobTrigger', repeatInterval: 1000l * 1 * 60 // execute job every 1 minutes
}
def execute() {
LOGGER.info("Beginning myFirstJob...")
etlService.launchJob("myFirstJob", false)
LOGGER.info("Completing myFirstJob...")
}
}

• Terra Terminal
• Terminator

This snippet will allow you to process a collection in "batches". In other words, grab the next n records and pass them to a function.

def vals = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k']
//
batches(vals, 3){ batch, batchNumber ->
println "batch $batchNumber: $batch"
}
//
public void batches(Collection vals, int batchSize, Closure action){
Collection batch
int batchNumber = 0
for(int i = 0; i < vals.size(); i = i+batchSize){
batch = []
(i..(i+batchSize-1)).each{
if(it< vals.size()){
batch << vals[it]
}
}
action.call(batch, batchNumber++)
}
}

While we want to make sure we push i18n to the views (gsp) in all cases, sometimes a service is the endpoint.

To access your message, inject the MessageSource into your service:
def messageSource

And then use it:
messageSource.getMessage(error, Locale.default)

You'll need to provide the Locale. In the above we are just selecting the default locale for the server

I was using the messageSource with the grails/spring validation framework to record what errors were occuring in a spring batch job. Hence, I was able to just pass the ObjectError to the getMessage method instead of having to specify the message code and params.

def messageSource
String validationExceptionForLocale(ValidationException){
StringBuilder body = new StringBuilder()
v.errors.allErrors.eachWithIndex {ObjectError error, Integer i ->
body.append("Error ${i+1}: ")
body.append(messageSource.getMessage(error, Locale.default))
body.append("\n")
return body.toString()
}

}

Pretty simple to do:
String underscoreToCamelCase(String underscore){
if(!underscore || underscore.isAllWhitespace()){
return ''
}
return underscore.replaceAll(/_\w/){ it[1].toUpperCase() }
}

To test it:
def vals = ['', ' ', 'id', 'transaction_id' ,'created_date']
vals.each{ String colName ->
println "converted: " + underscoreToCamelCase(colName)
}

References:

• How to convert camelcase string to underscored string
• Utility methods for camelcase -conversion

There are several good options. The gnome shell extension is rather nifty. The command line offers a lot more power though. Here are some nice references: linux-mag.com -- formums.virtualbox.org

I often keep a copy of documentation cached locally for offline reading / faster searching. Most of the time, projects provide this, but some do not. I have recently come across a project called HTTrack that allows finer grained control than wget --mirror

In my case, it was because I was using the wrong constructor:

I was using:
public HttpHost(final String hostname, int port)

instead of:
public HttpHost(final String hostname, int port, final String scheme)

Looking at the plug, hot is on the right, neutral is on the left, ground is on the bottom.

Neutral is the longer slot on some plugs.

If you are looking at an uninstalled outlet, hot is the brass colored terminal, while neutral is the silver one.

From wiki.answers.com

• Fixing a non-responsive screen
• Getting to the Parent Settings on the device