search

python / cpython

The Python programming language
https://www.python.org
Other
63.1k stars 30.22k forks source link

Read and write gzip header and trailer with zlib #103477

Open iii-i opened 1 year ago

iii-i commented 1 year ago

Feature or enhancement

Replace manual gzip format handling with zlib's inflateGetHeader() and deflateSetHeader().

Pitch

RHEL, SLES and Ubuntu for IBM zSystems (aka s390x) ship with a zlib optimization [1] that significantly improves deflate and inflate performance on this platform by using a specialized CPU instruction.

This instruction not only compresses the data, but also computes a checksum. At the moment Pyhton's gzip support performs compression and checksum calculation separately, which creates unnecessary overhead on s390x.

The reason is that Python needs to write specific values into gzip header; and when this support was introduced in year 1997, there was indeed no better way to do this.

Since v1.2.2.1 (2011) zlib provides inflateGetHeader() and deflateSetHeader() functions for that, so Python does not have to deal with the exact header and trailer format anymore.

Previous discussion

https://discuss.python.org/t/read-and-write-gzip-header-and-trailer-with-zlib/25703/2

[1] https://github.com/madler/zlib/pull/410

Linked PRs

serhiy-storchaka commented 8 months ago

On one hand, using the C code makes code faster, and it may help to solve #89672.

On other hand, see #112346. zlib implementation can produce different result. We should solve that issue first, and ensure that similar errors will not happen here.

iii-i commented 7 months ago

@serhiy-storchaka, @rhpvorderman suggested that we should use a simpler approach to achieve the same result: instead of using the new C functions, let zlib generate both gzip header and gzip trailer, and strip the gzip header. I have implemented it here: https://github.com/python/cpython/pull/112199. It passes the #114116 test. Could you please take a look?