How standard libraries read GZ files

2 GZ files given to 5 readers in 19 runtime versions. Cells of the ordinary files: reads 35, differs 3, refuses 0, crashes 0. How this is measured. The results as JSON.

GZ files, by reader
FileThe manifest says Go compress/gzip adapter Java java.util.zip.GZIPInputStream adapter Node.js zlib adapter Python gzip adapter Ruby Zlib::GzipReader adapter
gz/lorem-1mb.txt.gz 252.3 KB (252,288 bytes) members 1, uncompressed_bytes 1000000

reads reports uncompressed_bytes 1000000

reads reports uncompressed_bytes 1000000

reads reports uncompressed_bytes 1000000

reads reports uncompressed_bytes 1000000

reads reports uncompressed_bytes 1000000

gz/multi-member-3.gz 839 B members 3, uncompressed_bytes 1440

reads reports uncompressed_bytes 1440

reads reports uncompressed_bytes 1440

reads reports uncompressed_bytes 1440

reads reports uncompressed_bytes 1440

differs reports uncompressed_bytes 605 where the manifest has 1440

The GZ files themselves, and the measured properties each row is compared with: GZ sample files.