archive/zip reader memory usage

[MA LIG]

Hi, does anyone know whether the "archive/zip" reader does (or can) operate without reading the entire file into memory? In particular, I would like to process a large zip file by just streaming the requested decompressed content to the calling process (or function) and using a limited, fixed-size internal buffer that never requires any heap usage (per concurrent execution) beyond a predetermined, fixed size.

[Ian Lance Taylor]

The archive/zip Reader does not read the entire file into memory.
However, it requires an io.ReaderAt: it needs the ability to read data
from specific points in the input. So it can't read from an
unseekable stream. But it can produce a stream from a seekable file
using only a fixed amount of memory. I think the amount of memory
required will depend on the compression method.

Ian

[krolaw]

I'd be curious to know if https://github.com/krolaw/zipstream suits your needs.

Thanks.