bodoni / truetype Goto Github PK

Hi, so @vvuk has PR'd Windows support for webfonts in Servo servo/servo#15096 (comment), for what it uses your crate.

I was just wondering how well does this crate handle invalid input, because we may need to reconsider it if it can't handle it properly (given Servo would be feeding it random network data).

I've taken a quick look at the code and it seems reasonable, but I'm totally not an expert in fonts, so I wonder if you could help me figure out whether all the usages of unsafe are sound here.

Thanks for the answers!

Hello fellow Rustacean,
we (Rust group @sslab-gatech) found a memory-safety/soundness issue in this crate while scanning Rust code on crates.io for potential vulnerabilities.

Issue Description

tape::Tape::take_bytes() method creates an uninitialized buffer and passes it to user-provided Read implementation. This is unsound, because it allows safe Rust code to exhibit an undefined behavior (read from uninitialized memory).

This part from the Read trait documentation explains the issue:

It is your responsibility to make sure that buf is initialized before calling read. Calling read with an uninitialized buf (of the kind one obtains via MaybeUninit<T>) is not safe, and can lead to undefined behavior.

How to fix the issue?

The Naive & safe way to fix the issue is to always zero-initialize a buffer before lending it to a user-provided Read implementation. Note that this approach will add runtime performance overhead of zero-initializing the buffer.

As of Feb 2021, there is not yet an ideal fix that works with no performance overhead. Below are links to relevant discussions & suggestions for the fix.

• Rust RFC 2930
• Discussion in Rust Internals Forum

Hi. In the first font I tried, accessing CharMapping failed with:

thread 'main' panicked at 'not yet implemented', /home/simon/.cargo/registry/src/github.com-1ecc6299db9ec823/truetype-0.26.2/src/char_mapping.rs:109:21

It looks like this is because CharMapping::read tries does parse every Record but only supports a couple formats. If the font contains any in an unsupported format, unimplemented!() is called before parsing finishes. Maybe this font has some cmap records in a format that’s supported, and accessing those might be enough.

I’m not sure what a better API would be though. What do you think?

Here, this crate causes UB by "Producing an invalid value". According to this crater failure, this zero-initializes an instance of truetype::char_mapping::Encoding4, for which 0 is not a valid value. The type seems defined by a macro so I could not track this down further.

Maybe a good replacement would be MaybeUninit, which tracks the possibility of uninitialized values at the type level to make sure the compiler does not make any false assumptions.

When trying to parse a name table extracted from a font file, the parsing always fails due to underflow while reading. In normal usage, the entire file is provided to NamingTable::read, but in this case, just the bytes of the name table itself are given. I believe the issue is here: https://github.com/bodoni/truetype/blob/master/src/naming_table.rs#L124 -- that + 1 is incorrect; the length of the needed data buffer is just the max of offset + length.

     Finished dev [unoptimized + debuginfo] target(s) in 3.26 secs
     Running `target/debug/examples/render`
thread 'main' panicked at 'not yet implemented', /Users/ioover/.cargo/registry/src/github.com-1ecc6299db9ec823/truetype-0.22.0/src/char_mapping.rs:109
note: Run with `RUST_BACKTRACE=1` for a backtrace.

https://github.com/bodoni/truetype/blob/master/src/char_mapping.rs#L109