Fix broken sourcefile_attribute_parser

[emilycares]

Remove parsing of attribute_name_index and attribute_length from sourcefile_attribute_parser.
These fields don't need to be parsed here.

[Palmr]

Doesn't this differ from the spec?

https://docs.oracle.com/javase/specs/jvms/se8/html/jvms-4.html#jvms-4.7.10

Which says the SourceFile attribute struct is

SourceFile_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 sourcefile_index;
}

[emilycares]

It should not. This implementation works with a Java 8 compiler and also a Java 21 compiler.
I have recompiled the class with both compilers. And ran the tests.

Also, it seems to be that all other attribute parsers don't parse the first to fields to. For example:

classfile-parser/src/attribute_info/parser.rs

Lines 46 to 50 in fefca09

	pub fn code_attribute_parser(input: &[u8]) -> Result<(&[u8], CodeAttribute), Err<&[u8]>> {
	let (input, max_stack) = be_u16(input)?;
	let (input, max_locals) = be_u16(input)?;
	let (input, code_length) = be_u32(input)?;
	let (input, code) = take(code_length)(input)?;

Yes, it is weird that it is in the specification but does not need to be parsed here. At this point, I don't know why I don't have to parse these... And if I try, to parse then it errors.

u2 attribute_name_index;
u4 attribute_length;

[Palmr]

Name and length get parsed generically for all attributes here:

classfile-parser/src/attribute_info/parser.rs

Lines 16 to 19 in fefca09

	pub fn attribute_parser(input: &[u8]) -> Result<(&[u8], AttributeInfo), Err<&[u8]>> {
	let (input, attribute_name_index) = be_u16(input)?;
	let (input, attribute_length) = be_u32(input)?;
	let (input, info) = take(attribute_length)(input)?;

So the individual attribute parsers don't need to parse it themselves.

So your change is correct, good spot 👍