Fix GIF parser

Previously, the GIF parser would assume a GIF is animated if it had the
Netscape Application Extension. This can lead to false positives where a
GIF can have all the indicators of an animated GIF but only one Image
Descriptor.

Now, the code checks for the absence of the Graphic Control Extension
before an Image Descriptor to mean non-animated and otherwise checks if
there are multiple Image Descriptors. This unfortunately requires
parsing much further into the data stream.

The example cli now also supports specifiying the buffer size.

Cargo.toml

@@ -10,8 +10,13 @@ crate-type = ["cdylib", "lib"]
 [dependencies]
 byteorder = "1.5.0"
 pyo3 = "0.22.1"
+clap = { version = "4.5.16", features = ["derive"] }
 
 [dev-dependencies]
+clap = { version = "4.5.16", features = ["derive"] }
 paste = "1.0.15"
 serde = { version = "1.0.195", features = ["derive"] }
 serde_json = "1.0.111"
+
+[[example]]
+name = "cli"

README.md

@@ -30,7 +30,8 @@ else:
     size.is_animated
 ```
 
-You should not pass the entire image data, the first kilobyte or so should suffice.
+You should not pass the entire image data, the first kilobyte or so should suffice for most formats, other than GIF
+where a larger amount of data may be required to determine whether the image is animated or not.
 
 ### API
 
@@ -60,6 +61,11 @@ that does actual image parsing to determine if the data is actually an image.
 and does not necessarily support all features or variants of those formats, as a result, there might be
 false positives and false negatives.
 
+## Example CLI
+
+You can use `cargo example --cli` for a simple command line tool to try out this library. See
+`cargo example --cli -- --help` for details.
+
 ## Building
 
 Use [maturin](https://www.maturin.rs/) to build: `maturin build`

examples/cli.rs

@@ -1,20 +1,32 @@
 extern crate imgsize;
+use clap::Parser;
 use std::fs::File;
+use std::io;
 use std::io::Read;
-use std::{env, io};
+use std::path::PathBuf;
+
+#[derive(Parser)]
+struct Arguments {
+    #[arg(short, long, default_value_t = 1024)]
+    buf_size: usize,
+    paths: Vec<PathBuf>,
+}
 
 pub fn main() -> io::Result<()> {
-    let mut buffer = [0u8; 1024];
-    for path in env::args().skip(1) {
+    let arguments = Arguments::parse();
+    let mut buffer = vec![0u8; arguments.buf_size];
+    for path in arguments.paths.iter() {
+        let name = path.to_str().unwrap();
         let mut file = File::open(&path)?;
-        file.read(&mut buffer)?;
-        match imgsize::get_size(&buffer) {
+        let read = file.read(&mut buffer)?;
+        match imgsize::get_size(&buffer[..read]) {
             Some(size) => println!(
                 "{}: {}x{}, {}, animated={}",
-                path, size.width, size.height, size.mime_type, size.is_animated
+                name, size.width, size.height, size.mime_type, size.is_animated
             ),
-            None => println!("{}: unsupported format", path),
+            None => println!("{}: unsupported format", name),
         }
+        buffer.fill(0);
     }
     Ok(())
 }

python-tests/test_sample_files.py

@@ -1,9 +1,7 @@
 import json
-
 import pytest
-from conftest import ROOT
 
-BYTES_TO_READ = 1024
+from conftest import ROOT
 
 
 def find_examples():
@@ -12,7 +10,7 @@ def find_examples():
         output_path = input_path.with_suffix('.output')
         if not output_path.exists(): continue
         with input_path.open('rb') as fobj:
-            data = fobj.read(BYTES_TO_READ)
+            data = fobj.read()
         with output_path.open('r') as fobj:
             output = json.load(fobj)
         yield pytest.param(data, output, id=input_path.stem)