DB Dump Corrupting Unicode Characters

[JetUni]

Affected area

Database

Supabase CLI version

v2.115.0

Operating system

Windows 11

Installation method

Other

Command

supabase db dump --data-only > supabase/dump-data.sql

Actual output

These are the values from the dump-data.sql file generated by the provided command
"d├╢Terra"
"Oranges 🍊"
"First 💳 Bonus"

Expected behavior

Here are the unicode strings that were in the columns. I expected the unicode to be preserved when dumping the DB.
"döTerra"
"Oranges 🍊"
"First 💳 Bonus"

Steps to reproduce

Insert these values into a text column in your DB, then dump the DB and observe the corrupted unicode

Crash report ID

No response

Docker and service versions

Additional context

No response

[Cypher-Aura-19]

I reproduced this on Windows 11 and traced the output bytes. The CLI and pg_dump both emit valid UTF-8; the corruption happens specifically with Windows PowerShell 5.1 > redirection, which decodes native stdout using the console code page before writing the file.

Using the CLI's file option preserves the dump correctly:

supabase db dump --data-only -f supabase/dump-data.sql

cmd.exe and Git Bash redirection also preserved the UTF-8 output in my testing.

[7ttp]

hey folks  💚 super thanks for this!

reproduced and it's not the cli tbh

windows PowerShell 5.1's > re-encodes

(🍊 → ≡ƒìè is exactly UTF-8 read as CP437)

which we can't really prevent from inside the process. as @Cypher-Aura-19 said:

supabase db dump --data-only -f "supabase/dump-data.sql"

writes the file directly and keeps everything intact

tho am thinking to maybe add a heads-up pointing at --file for exactly this....