Long escaped tag URI sequences throw BufferOverflowException

[GoogleCodeExporter]

What steps will reproduce the problem?
Attempt to load a YAML document containing a long sequence of tag URI escape 
characters (%HH), at least 257 characters long. A test case:


/**
 * Try loading a tag with a very long escaped URI section (over 256 bytes' worth).
 * @throws IOException
 */
public void testLongURIEscape () throws IOException {
    YamlLoader loader = new YamlLoader();

    // Create a long escaped string by exponential growth...
    String longEscURI = "%41";  // capital A...
    for ( int i = 0; i < 10; ++i ) {
        longEscURI = longEscURI + longEscURI;
    }

    String yaml = "foo: !<"+longEscURI+">";

    try {
        loader.load ( yaml );
    } catch ( Exception e ) {
        assertTrue (
                "Exception must be due to bad tag, not due to buffer overflow but was "+e.getClass(),
                e instanceof ConstructorException
        );
    }
}


(The ConstructorException is expected as there is no class with a name 
consisting of 1024 upper-case "A" characters.)


Instead of blindly allocating 256 bytes of buffer and hoping that works, detect 
the exact length we will need by looking ahead.

A patch file that fixes the issue is attached.

Original issue reported on code.google.com by JordanAn...@gmail.com on 2 Mar 2011 at 8:00

Attachments:

• snakeyaml_rev1293.patch

[GoogleCodeExporter]

I have changed the test to be successful.
The patch is slightly changed to start with 1.

Original comment by py4fun@gmail.com on 3 Mar 2011 at 1:15

• Changed state: Started
• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Agreed.

Original comment by JordanAn...@gmail.com on 3 Mar 2011 at 1:21

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Original comment by py4fun@gmail.com on 3 Mar 2011 at 4:17

• Changed state: Fixed
• Added labels: ****
• Removed labels: ****