Simple XML tag validator

[kapa]

Hi

The program is suppose to search tags from xml file and determine does
every tag have a pair, meaning if there is a <startthere must be
</start>, and make count how many there is that particular type. Or in
case <foo/just take notice and count.

Ok, I have managed to make the program return the tags like below, and
return the value how many tags there is:

things
basket
fruit
/fruit
bug/
/basket
foo/
/things
Now my problem is that I really don't now how to compare is there a
match, and if found how to make sure the found pair is no longer compared.

Now I have put the tags into a two-dimensional array. But I'm wondering
should I use stack or what, and how? My head is empty at the moment, so
any help would be appreciate.

bool getTag( ifstream &in, char tag[], int taglen ) ;
void cleanTag( char tag[], int taglen, int &count ) ;
void printTags( char tags[][30], int count ) ;

int main()
{
int count = 0 ;
char tag[150] ;
char tags[50][30];

ifstream in ;
in.open( "Test1.xml" ) ;

if( !in.is_open() )
cout << "Can't open the file!\n" ;
else
{
while( getTag( in, tag, 150 ) )
{
cleanTag( tag, 30, count ) ;
if( strlen( tag ) != 0 )
strcpy( tags[count-1], tag ) ;
}
}
printTags( tags, count ) ;
return 0 ;
}

void printTags( char tags[][30], int count )
{
for( int i=0; i<count; i++ )
{
for( int j=0; tags[i][j] != '\0'; j++ )
{
cout << tags[i][j] ;
}
cout << endl ;
}
}

[Zeppe]

kapa wrote:

Hi

hi!

The program is suppose to search tags from xml file and determine does
every tag have a pair, meaning if there is a <startthere must be
</start>, and make count how many there is that particular type. Or in
case <foo/just take notice and count.

Ok, I have managed to make the program return the tags like below, and
return the value how many tags there is:

things
basket
fruit
/fruit
bug/
/basket
foo/
/things
Now my problem is that I really don't now how to compare is there a
match, and if found how to make sure the found pair is no longer compared.

If you just want to check the pairs, a simple method would be this:
* you have threee types of tags: opening (<foo>), closing (</foo>),
and alone (<foo />).
* you create a stack of tag that are opened and that are waiting for
the closing tag.
* for each tag, you read it:
- if the tag is alone, ignore it and go ahead;
- if the tag is opening, you put it in the top of the stack and go ahead;
- if the tag is closing, it has to match to the opening tag at the
top of the stack. If it does, everything is ok and you pop the top of
the stack and go ahead. Otherwise (empty stack or unmatching elements)
there is an error and the xml is badly formatted.

Example:

things
basket
fruit
/fruit
bug/
/basket
foo/
/things

step1: insert things

things <-

step2: insert basket

basket <-
things

step3: insert fruit

fruit <-
basket
things

step 4: /fruit matches with the top, remove fruit

basket <-
things

step5: /bug doesn't match with the top: there is an error

Of course, now everything is bad and you should end your analysis. If
you want to implement some intelligence and some reasoning on the error
cause, it becomes more complicated.

Now I have put the tags into a two-dimensional array. But I'm wondering
should I use stack or what, and how? My head is empty at the moment, so
any help would be appreciate.

The stack was a good intuition! Use that one defined by the stl
(#include <stack>).

Another good idea would be to define a simple class for the problem

class Tag
{
public:
enum Type {OPENING, CLOSING, ALONE};
public:
Tag(Type t, std::string name);

std::string GetName(); // returns the pure name

Type GetType(); // returns the type of the tag

bool IsMatching(const Tag& t); // returns true if t is closing and
this is opening with the same name and viceversa
};

Good work!

Bye,

Zeppe