VB2008从入门到精通(PDF格式英文版)-第75部分
按键盘上方向键 ← 或 → 可快速上下翻页，按键盘上的 Enter 键可回到本书目录页，按键盘上方向键 ↑ 可回到本页顶部！
————未阅读完？加入书签已便下次继续阅读！
actual object instances。 There could be a case for adding code that verifies if processor references
an actual object instance； but there is absolutely no point to adding code to verify if writer and
reader reference actual object instances。 Doing so would imply that your truth test block is
inplete and you have not thought through all the permutations that assign writer and reader。
Let’s look at the next truth test； which indicates that there is one mand…line argument。
When there is one mand…line argument； you need to verify which of the two mand
line variations it is：
TestProcessor。exe …help
or：
TextProcessor。exe lotto。txt
The first variation has an explicit mand…line parameter …help。 The second variation is
the identifier of a file that contains the input data。 Thus； the second If block contains another
If block to test another aspect of the truth test：
ElseIf args。Length = 1 Then
If args（0） = 〃…help〃 Then
Bootstrap。DisplayHelp（）
Return
End If
reader = File。OpenText（args（0））
writer = Console。Out
When testing for the …help argument； immediately after calling DisplayHelp（）； you must
use the Return keyword。 This is extremely important because； when the console application
calls DisplayHelp（）； it is saying； “I don’t care where the input or output data is ing from
since I am doing something else and thus must stop processing。” If you were to continue
processing； the writer and reader could reference invalid states and cause an exception to
be generated。
If all tests have been tried and failed； then the final Else calls the method DisplayHelp（） to
indicate that the mand line is incorrect； and the help shows the correct mand line。
Doing a plete test and debug of the console application is difficult because of the
streaming facilities。 When you stream data from one process to another； a process is started
…………………………………………………………Page 282……………………………………………………………
260 CH AP T E R 1 0 ■ L E A R N I N G A B OU T P E R S IS TE N CE
and stopped without an easy way for the debugger to jump in between。 Thus； to be able to
debug and verify if everything works； another strategy is needed。 The solution is to use prepro
cessor directives （discussed in Chapter 6）； like the following。
#If DEBUG_OUTPUT Then
Console。WriteLine（〃Argument count（〃 & args。Length & 〃）〃）
For Each argument as String in args
Console。WriteLine（〃Argument （〃 & argument & 〃）〃）
Next
#End If
The directive code is executed if DEBUG_OUTPUT is defined。 In the example； the argument
count and arguments are output to the console。 To activate the directive； the item DEBUG_OUTPUT
is defined at the project level or near the top of the source code file。
The shell is plete； and all that remains is the implementation of the text processor。
Implementing the TextProcessor Application
In implementing the text processor； the problem that confronts us is how to fix a data stream so
that it is consistent。 The main reasons for data inconsistencies are human error and carelessness。
Figure 10…3 shows what is displayed when you load the lottery number text file in Notepad。
It seems very obvious that something is not right with this file。
Figure 10…3。 The lottery text file loaded in Notepad
…………………………………………………………Page 283……………………………………………………………
CH A PT E R 1 0 ■ L E A R N I N G A B O U T P E R S IS T E N CE 261
However； what appears wrong in the Notepad display is not the real problem。 When the
text file is loaded using another text editor； such as Vim； the text is displayed as shown in
Figure 10…4。 As you can see； Vim has loaded the text file without any formatting errors。
■Note Vim is available from http：//vim。org。 It is a vi…derived clone that can be used on
Windows systems。
Figure 10…4。 Vim loads the text file in a nicely formatted display。
The real pressing problem lies in the structure of the data； which is illustrated in Figure 10…5。
Here； the data has new formatting； with extra columns； and the first column is not always in
the proper data format。 And to make matters worse； the badly formatted data has repeating
information。
The challenge of the application is to read the stream and fix all of the problems。 This requires
a thorough understanding of string processing and the different ways that text can be stored；
as discussed in Chapter 3。 When you are processing data streams； you need to be aware of the
format of the data stream。 In this example； we are processing ASCII text； and thus will be manipu
lating bits according to the rules of the ASCII lookup table。
Whitespace characters are special characters in the text lookup table。 They are associated
with numbers； but their representation is in the form of an action that the user can see。 For
example； the character between single quotation marks （' '） is a space； the character t is a tab；
and the character n is a newline。 The reason Notepad does not format the lottery text file nicely
（Figure 10…3） is because of the whitespace characters used to indicate a newline。 In Figure 10…6； the
highlighted buffer entry 0A is the hexadecimal character that indicates a linefeed； or newline； in
the lottery text file。
…………………………………………………………Page 284……………………………………………………………
262 CH AP T E R 1 0 ■ L E A R N I N G A B OU T P E R S IS TE N CE
Figure 10…5。 Structural problems of this data stream
Figure 10…6。 Newline character used in lotto。txt
Figure 10…7 is a file created by Notepad。 Notepad expects not a single whitespace character；
but two whitespace characters to indicate a newline： 0D and 0A。
…………………………………………………………Page 285……………………………………………………………
CH A PT E R 1 0 ■ L E A R N I N G A B O U T P E R S IS T E N CE 263
Figure 10…7。 Newline characters used by Notepad
Deciphering the Format
The echo has served its purpose of providing a way to develop an application in a top…down
manner。 The next step is to remove the echo code and start writing the code that will fix the
data stream。
Fixing the data stream is not a trivial undertaking； because you are yet again faced with a
state problem。 You don’t want to fix one part of the stream； only to end up with a problem in
another part of the stream。 Thus； you need to incrementally fix the stream and make sure at
each step that there are no ramifications。
The first step is to break the data stream into individual fields （each value in a column is a
field in this case）。 In Figure 10…5； the data stream had two parts； where the upper part seemed
to have a single space between the numbers and the lower part had the amount of space neces
sary to align the numbers。 The difference between the upper and lower parts is the whitespace
characters used。 So； the first step will be to clean up the whitespace。
The following is the code that reads the buffer； splits it up； and reassembles the content
into a new buffer。 The code is intermediate code that adds special bracket markers to indicate
what the text contains。
Imports System。IO
Imports System。Text
' TODO： Fix up this class
Public Class LottoTicketProcessor ： Implements IProcessor
Public Function Process（ByVal input As String） As String
Implements IProcessor。Process
Dim reader As TextReader = New StringReader（input）
Dim retval As New StringBuilder（）
…………………………………………………………Page 286……………………………………………………………
264 CH AP T E R 1 0 ■ L E A R N I N G A B OU T P E R S IS TE N CE
Do While reader。Peek（） …1
Dim splitUpText As String（） = _
reader。ReadLine。Split（New Char（） ｛〃 〃c； ControlChars。Tab｝）
Dim c1 As Integer
For c1 = 0 To splitUpText。Length 1
retval。Append（（〃（〃 & splitUpText（c1） & 〃）〃））
Next
retval。Append（ControlChars。NewLine）
Loop
Return retval。ToString（）
End Function
End Class
In the implementation of Process（）； the text will be parsed line by line。 Then each line
is split into the individual fields。 You could write the parsing routines yourself； but to parse a
buffer line by line； it is more efficient to use StringReader。 StringReader accepts the string to
parse and is then assigned to a TextReader interface instance。
As each line of text is parsed； the most efficient approach to building a buffer is to use
StringBuilder。 You could keep appending data to the string； but if you do that too often the
application’s performance will suffer。
The String type is an immutable type； which means once an object is initialized； you
cannot change the state of the object。 The advantage of immutable types is that they increase
the speed of your application； because code can assume once an object has been assigned； it
will never change。 The downside is that once an object is assigned； to modify the object state
even slightly； you must instantiate a new object； which would be the c
