xml_parser_c

One .h file for xml parsing for C

It's very easy to use it

const char* xml_string = "<Error><Code>AccessDenied</Code><Message>Access Denied</Message></Error>";

const char* result = xml_parse(xml_string, -1, "Error.Code", 1, 1);

printf("$s", result);

Output:

AccessDenied

Function and parameters:

Signature

const char* xml_parse(const char* xml, int length, const char* pattern, int access, char is_copy)

Params

1. xml - input string in xml format
2. length - Total characters for parsing (no feof checks). Or -1 - until the end (for zero terminated string)
3. pattern - string with .(dot) separator for tags. You can use array [] notation to access equal tags on one level. Also there is @ to access attributes. Example of complicated valid pattern: "root.element[13].item[21].name@lang" to access lang attribute on third level of deepness name tag.
4. access - There is 4 values: 0,1,2,3 to access different part of target tag:

0.....1...........2......3
<code>AccessDenied</code>

5. is_copy - Flag. Values are: 0 - return pointer from input string, 1 - create a new string.

Understanding access and is_copy parameters

For this kind of input string:

<Error><Code>AccessDenied</Code><Message>Access Denied</Message></Error>

And this kind of pattern:

Error.Code

This table shows how these two parameters access and is_copy change the target result

access	is_copy	result
0	1	<code>AccessDenied</code>
1	1	AccessDenied
2	1	</code>
3	1	empty string
0	0	<code>AccessDenied</code><Mess...
1	0	AccessDenied</code><Mess...
2	0	</code><Mess...
3	0	<Mess...

What about single tags?

For example input string is as follows:

<Error><Code>AccessDenied</Code><single name="any"/></Error>

Pattern like so

Error.single

access	is_copy	result
0	1	<single name="any"/>
1	1	empty string
2	1	empty string
3	1	empty string
0	0	<single name="any"/></Erro...
1	0	</Erro...
2	0	</Erro...
3	0	</Erro...

Comments?

For example we have this gipotethic xml:

<root>
    <element1 /> 
    <!--This is the comment--> 
    <code>38389338</code> 
    <!--This is the comment # 2-->
    <element2 color="black"/>
</root>

Pattern to get a comment 1:

root.!--

pattern	access	is_copy	result
root.!--	0	1	<!--This is the comment-->
root.!--	1	1	This is the comment
root.!--	2	1	-->
root.!--	3	1	empty string
root.!--	0	0	<!--This is the comment--> <code>3...
root.!--	1	0	This is the comment--> <code>3...
root.!--	2	0	--> <code>3...
root.!--	3	0	<code>3...
root.!--[1]	0	1	<!--This is the comment # 2-->
root.!--[1]	1	1	This is the comment # 2
root.!--[1]	2	1	-->
root.!--[1]	3	1	empty string
root.!--[1]	0	0	<!--This is the comment # 2--> <eleme...
root.!--[1]	1	0	This is the comment # 2--> <eleme...
root.!--[1]	2	0	--> <eleme...
root.!--[1]	3	0	<eleme...

Note: For comment N use: root.!--[N-1] pattern root.!--[0] same as root.!-- For undefined comment use: root.!--[%d] pattern and vararg param

CDATA sections

For example we have xml:

<root>
    <element1 /> 
    <!--This is the comment--> 
    <code><![CDATA[9879879879]]></code> 
    <!--This is the comment # 2-->
    <element2 color="black"/>
</root>

Pattern to get a CDATA section:

root.code.![CDATA[

access	is_copy	result
0	1	<![CDATA[9879879879]]>
1	1	9879879879
2	1	]]>
3	1	empty string
0	0	<![CDATA[9879879879]]></code...
1	0	9879879879]]></code...
2	0	]]></code...
3	0	</code...

License

MIT