Access response headers

[pavdev64]

I would like to access the response headers if the pdf document request fails.

try {
    const loadingTask = pdfjs.getDocument("https:///doc.pdf");
    const pdfDoc = await loadingTask.promise;
    console.log(`pdf loaded, ${pdfDoc.numPages} pages`);
  } catch (error) {
    const referenceId = error?.response?.headers?.['x-azure-ref'];
    const correlationId = error?.response?.headers?.['x-correlation-id'];

    console.error(`Pdf document loading failed, referenceId: ${referenceId}, correlationId: ${correlationId}`);
  }

Currently, the response object is missing in the error object and it is not possible to read the response headers and get the reference and correlation id the server returns.

An example of the current error object (no response property).

{
  "message": "Missing PDF \"[object Object]\".",
  "name": "MissingPDFException"
}

Some way would be to use custom fetching using the fetch function:

try {
    const response = await fetch("https:///doc.pdf");
    if (response.status !== 200) {
      throw {
        message: "Fetch failed",
        name: "FetchFailed",
        response,
      };
    }
    const pdfData = await response.arrayBuffer();
    const loadingTask = pdfjs.getDocument(pdfData);

    const pdfDoc = await loadingTask.promise;
    console.log(`pdf loaded, ${pdfDoc.numPages} pages`);
  } catch (error) {
    const referenceId = error?.response?.headers?.['x-azure-ref'];
    const correlationId = error?.response?.headers?.['x-correlation-id'];

    console.error(`Pdf document loading failed, referenceId: ${referenceId}, correlationId: ${correlationId}`);
  }

but, that would mean that the pdf document would not be loaded via the pdfjs worker and possibly "PDFDataRangeTransport" would not be used either. It would wait for the entire pdf to be loaded into memory (think of 100MB pdfs).

I would rather use the already implemented loading.

[Snuffleupagus]

Please see PR #18908.