Condor_dbq defines the following packages (indented lines show the functions in
the package):

CondorUserLog
	(constructor)
	FileExists
	Open
	Close
	SetPosition
	(destructor)
	ReadFileChunk
	GetNextRecord

CondorWork
	(constructor)
	<regular expressions to handle userlog entries>
	Print
	RestoreJobs
	MakeCondorId
	IsTerminalState
	SetPosition
	NumIncompleteJobs
	ParseRecordToEvent
	GetNextEvent
	GetNewEvents
	AddCondorIdsFromLog
	PrintEvent
	RemoveAlreadyTerminatedJobEvents
	ApplyJobUpdates
	ProcessNewEvents

CondorActiveWork
	(constructor)
	AddWork
	NumWork
	NumIncompleteJobs
	ProcessWork
	CheckIfWorkSubmitted
	RecoverChosenWork
	RestoreStateFromDatabase
	Recover

In main::, there are SQL statements to set up tables: InitializeSqlStmts sets
up a hash table with the following entries:
		createWorkTable
		createJobsTable
		createWorkSeq
		createWorkFunction
		createWorkTrigger
		createJobsFunction
		createJobsTrigger
		createSubmitUser

	Permissions?
		grantSubmitUserWorkTable
		grantSubmitUserJobsTable
		grantSubmitUserWorkSeq
		createExecuteUser
		grantExecuteUserWorkTable
		grantExecuteUserJobsTable
		revokePublicSchemaCreate
		createLangPlpgsql

	Updates?
		updateWorkToChosen
		selectChosenWork
		updateWorkToInBatch
		updateWorkToTmpFailed
		updateWorkToFailed
		insertJob
		updateJob
		updateWorkJobs
		updateWorkComplete
		updateWorkJobsError
		getChosenWork
		updateWorkToInitial
		updateAllChosenWorkToInitial
		getInBatchWork
		getInBatchJobs

	GetDbConnectionInfo
	ConnectToDb
	ExecuteDbOperation
	GetPreparedStmt
	DoCreateLangPlpgsql
	DoRevokePublicSchemaCreate
	DoInitDb
	DoCreateSubmitUser
	DoCreateWorkUser
	Log
	LogAndExit
	Usage
		Prints out command line options
		
	ProcessOptions
		Calls on GetOptions to read the command line
		On error, calls on Usage

	Initialize
		First thing that runs.  Sets up everything with ProcessOptions and
		InitializeSqlStmts

	CheckCondorSubmitRunning
	GetFileContents
	ShellQuoteArg
	ExecuteCommand
	SubmitToCondor
	DoAdminCmds
	DoSubmit
	DoGrabWork
	DoSubmitWork
	DoProcessQueue
	Main 
		Program runs in one of three modes:
			Admin
		Submitting
		Processing the Queue

Submitting is easy: Jobs are inserted into the 'work' table.  condor_dbq
then checks that rows were entered.

The Processing is a bit more complicated.  Condor_dbq gets a list of jobs it is
to look at.  It runs a query over the jobs it needs to look at.  Then it
iterates over the jobs to find jobs it has successfully submitted and failed to
submit. Jobs that are successfully submitted are moved to the 'jobs' table.
When the job completes it sets the state field of the job in 'work' to
'complete' or 'failed'. To determine the status of jobs, it reads the userlogs.