chunk mode send json

[hyzeer]

Hi , thank you very much for your wonderful framework , but I regret to find an error in processing NON-ASCII characters by using the framework .

for chunk in res:
     await resp.send('{:x}\r\n'.format(len(chunk)))
     await resp.send(chunk)
     await resp.send('\r\n')

the above codes in server.py at line 337-340 , "len(chunk)" refer to the numbers of the character variable "chunk" , not the bytes length . It's OK with the return of the function "len( )" , when the character variable is ASCII type , but for the NON-ASCII type , such as a Chinese character , the character number is 1 in function "len( )" , and the length is 3 bytes . So , this function will cause a loss in transmission .

[belyalov]

Hi @hyzeer,

thanks for reporting this issue - I never tried to send non ASCII chars through JSON, so you're totally right! :)

It actually easy to fix by

            chunk_len = len(s.encode('utf-8'))
            await resp.send('{:x}\r\n'.format(chunk_len))

The question here is - do you guys always using utf8 as default encoding? In Russia we do use utf8 as default, so I wonder do we need to make encoding configurable or we can just hardcode it - in order to save bytecode / RAM.

[hyzeer]

Hi @belyalov , thanks for your reply. Yes, most of us do using utf-8 as default encoding anytime. :)

Could you release new firmware after fixed? I can do it, but I not cross compile environment. Thanks!

[belyalov]

Fix just landed and new firmware is ready.

Enjoy!

[hyzeer]

Thanks! :)