Seems to be working here. Notice, when you are using unicode characters, you need to add # -*- coding: UTF-8 -*- to your file first line:

#!/usr/bin/env python3
# -*- coding: UTF-8 -*-

a=[]

for _ in range(5):
    a.append('你好你好')
    print('\n'.join(a))

thanks for your reply.
but, sorry, the point is 5000 times and move 'print' out from 'for'. (and the first two lines in your code are meanless in python 3 IMO)

so please fix.

Maybe you need add "directwrite" to your font options. Preferences -> Settings - User

https://i.imgur.com/lZEmHAg.png

Thanks for your reply.
tried but still failed.
sublime console always shows:
"

[Finished in 0.3s]
"

Interestingly, changing the 5000 to 2340 works for me, but any higher values and the print isn't reflected in the build output, nor any prints after it. Python version 3.6.1 (v3.6.1:69c0db5, Mar 21 2017, 17:54:52) [MSC v.1900 32 bit (Intel)] sys.version_info(major=3, minor=6, micro=1, releaselevel='final', serial=0), Windows 7 x64.

The Sublime Text build system expects UTF-8 from the running program – this can be seen in the file Packages/Default/exec.py using https://packagecontrol.io/packages/PackageResourceViewer.

The os.read() call reads up to 32,768 bytes at a time from the program. The string 你好你好 is 12 bytes long in UTF-8, implying 3 bytes per char. 5000 * 12 = 60,000 bytes. 32,768 is not divisible by 3, so the read ends up reading a partial UTF-8 byte sequence. A decode error occurs, which leads to a bug in exec.py causing the error to not be printed.

The fix for this is to:

1. Keep reading data if the read results in 32,768 bytes – this should help prevent reading a partial byte sequence
2. Ensure that the decoding error message makes it to the output, but not resetting the proc var

so is this bug going to be fixed in next release? thanks

This has been fixed in build 3153