Major performance issues with bytea performance #2240

Hi,

I encountered the exact same issue as #1286 wheres a simple query to retrieve a 100mb bytea column from a one table, one row DB takes forever and causes the cpu to spike at 100%, rendering my node instance unresponsive.

After investigating, I think I found the cause in the pg-protocol package

Which suffers from the same thing as the original issue - too many allocations and buffer copies going around. I believe the fix introduced back then in brianc/node-packet-reader#3 should solve the problem here.