I think that's meant to be a text chunk:

Lines 650 to 657 in 12399c0

    
           // write auth chunk 
        
           write_chunk(Footer.auth, OutputFile); 
        
           // write text chunk 
        
           write_chunk(Footer.text, OutputFile); 
        
           // write mixe chunk 
        
           write_chunk(Footer.mixe, OutputFile);

That code was added in commit: b8425af ("Refactor NSFe metadata generation"), part of PR #262.

Apparently the refactored code only initializes the text chunk conditionally, but writes it out unconditionally. It should probably do this the other way around, considering that the condition is whether the text in question is empty.

So the chunk identifier here is uninitialized data; fortunately, the size is not, but unfortunately, the uninitialized data could turn out to be a defined chunk type (or an unknown mandatory chunk type), which could cause the file to be incorrectly processed (or rejected). (Theoretically, it's worse than that; most things you can do with uninitialized memory are undefined behavior, though I wouldn't swear that CFile::Write is one of those things. But this is bad enough as it is.)

(I sure hope the text is already stored in UTF-8, but didn't see any clear evidence of this.)

	// write auth chunk
	write_chunk(Footer.auth, OutputFile);

	// write text chunk
	write_chunk(Footer.text, OutputFile);

	// write mixe chunk
	write_chunk(Footer.mixe, OutputFile);

NSF2 export with corrupted NSFe chunk ident? #335

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions