tonhowtf/omniget · error

ffmpeg nao converteu a imagem

Error message

ffmpeg nao converteu a imagem {}

What it means

ensure_jpeg converts a downloaded image (any format) to JPEG using ffmpeg at quality -q:v 2. If the conversion process exits non-zero, it throws 'ffmpeg nao converteu a imagem {idx}' (ffmpeg did not convert image idx), including the image index. This indicates ffmpeg could not decode the input image or write the JPEG output.

Solutions

1. Validate the downloaded bytes are actually an image (content-type/magic bytes) before conversion.
2. Log ffmpeg stderr around the call to see the exact decode failure.
3. Add a pre-check on Content-Type and reject non-image responses before calling ensure_jpeg.
4. Ensure ffmpeg has decoders for the source format, or convert via an alternative tool (e.g. ImageMagick).

Example fix

// before
let bytes = http_get(url).await?;
ensure_jpeg(&bytes_path, idx).await?;
// after
let resp = http_get(url).await?;
anyhow::ensure!(
    resp.content_type().starts_with("image/"),
    "URL nao retornou uma imagem"
);
ensure_jpeg(&bytes_path, idx).await?;

// Rust
async fn looks_like_image(bytes: &[u8]) -> bool {
    bytes.starts_with(b"\xFF\xD8") // jpeg
        || bytes.starts_with(b"\x89PNG")
        || bytes.starts_with(b"GIF8")
}

fn is_image_response(ct: &str) -> bool { ct.starts_with("image/") }

match ensure_jpeg(&input, idx).await {
    Err(e) if e.to_string().contains("nao converteu a imagem") => {
        eprintln!("imagem {idx} invalida ou formato sem decoder");
    }
    r => r?,
}

When it happens

Trigger: Calling download() with a slide image whose format ffmpeg can't decode (corrupt download, HTML error page saved as image, unsupported/proprietary format), or an output path that can't be written.

Common situations: Remote URL returns an error page or rate-limits and the body is fed to ffmpeg; SVG/HEIC input without decoder support; zero-byte or truncated downloads; disk full.

Related errors

• HTTP
• {}
• download de falhou: HTTP
• nao foi possivel buscar
• corpo ilegivel de

AI-assisted analysis of tonhowtf/omniget@8600b91f42 (2026-09-12). Data as JSON: /api/errors/f3783b84cde594e0. Report an issue: GitHub.

Appendix: source

Thrown at src-tauri/omniget-core/src/core/tools/slides.rs:77

/// Converte qualquer imagem em JPEG pelo ffmpeg (webp/png escapam do DCTDecode).
pub(crate) async fn ensure_jpeg(data: Vec<u8>, work: &Path, idx: usize) -> anyhow::Result<Vec<u8>> {
    if super::jpeg_pdf::is_jpeg(&data) {
        return Ok(data);
    }
    let ffmpeg = crate::core::dependencies::ensure_ffmpeg().await?;
    let input = work.join(format!("{}.img", idx));
    let output = work.join(format!("{}.jpg", idx));
    tokio::fs::write(&input, &data).await?;
    let o = crate::core::process::command(&ffmpeg)
        .args(["-y", "-hide_banner", "-loglevel", "error", "-i"])
        .arg(&input)
        .args(["-q:v", "2"])
        .arg(&output)
        .output()
        .await?;
    if !o.status.success() {
        return Err(anyhow!("ffmpeg nao converteu a imagem {}", idx));
    }
    Ok(tokio::fs::read(&output).await?)
}

pub async fn download(
    url: &str,
    dest_dir: &str,
    progress: super::ProgressFn,
) -> anyhow::Result<SlidesResult> {
    if !url.contains("slideshare.net") {
        return Err(anyhow!("cole um link do slideshare.net"));
    }
    let client = super::client()?;
    let html = client
        .get(url)
        .send()
        .await?
        .error_for_status()?

View on GitHub (pinned to 8600b91f42)