Identify corrupted entries before reading their data

[ksidlauskas]

Hi, is it possible to identify which files are corrupted inside the zip before you start reading their data?

I have a zip file that is causing issues when extracting (local file header not found), but I would like to know that at the time i check the contents of the zip, rather than when i'm already trying to read the data.

I compared the entry information for all the contents of the zipfile, but they don't seem to have any obvious issues.

The error is thrown on const blob = await file.getData(new zip.BlobWriter(file.type));
below is the error info that's logged out. I am using the zip-full.js;

[gildas-lormeau]

The local header of an entry in a zip file must begin with 0x504b0304 (in big-endian). The error you mentioned is thrown when this value is not found (cf. chapter 4.3.7 in the spec).

zip.js/lib/core/zip-reader.js

Lines 260 to 262 in a8f0189

	if (getUint32(dataView, 0) != LOCAL_FILE_HEADER_SIGNATURE) {
	throw new Error(ERR_LOCAL_FILE_HEADER_NOT_FOUND);
	}

This test cannot be run when listing entries because the ZipReader#getEntries method would run much slower, especially when retrieving the zip content with a HttpRangeReader instance or when reading a zip file with hundreds of entries.

Note however that you can abort the execution of Entry#getData if you want to test an entry without reading the data. For this, you can use the signal option, e.g.

const entries = await zipReader.getEntries();
const controller = new AbortController();
try {
  const promiseEntryData = entries[0].getData(new zip.BlobWriter(), { signal: controller.signal });
  controller.abort();
  await promiseEntryData;
} catch (error) {
  if (error.message != zip.ERR_ABORT) {
    throw error;
  }
}