In Python3.5 asyncio.StreamReader.readline() can separate lines only by b'\n'. For some task, I want actually "read until separator", where separator is another byte.

Since I cannot create pull-request, I will write just here what I want:

@coroutine
def readline(self, separator=b'\n'):
    if not isinstance(separator, bytes) or len(separator) != 1:
        raise ValueError(separator)
    ....
    ichar = self._buffer.find(separator)
    ....

I like the idea. It's common in protocols to have different kinds of
"terminators" to mark the end of a message. C strings use a null byte
for example. I would be trivial to accept other terminators, since we
don't store received bytes as lines, but as a raw byte string
(technically in a bytearray).

> if not isinstance(separator, bytes) or len(separator) != 1:

Yes, we should only reject empty terminator string. But terminator
strings of more than 1 character are fine, no? I don't think that we
need to put an arbitrary limitation here.

> Since I cannot create pull-request, I will write just here what I want:

Oh, what is your problem? You can use https://github.com/python/asyncio

When I reported that bug, I want to suggest multibyte separators, but implementation is much complex than current one. For example, we should decide what to do if partial separator is read and EOF occur.

So, I think we should merge ability to use custom one-byte separator, and next create another pull-request to make support of multibyte separator (like '\r\n').

Okay, I can make pull request on asyncio, but asyncio now is the part of Python. Does not it ?

If you hit EOF in the middle of a partial separator, treat it as a premature EOF just as if you hit EOF before the separator -- the record is not properly terminated so it shouldn't be accepted.

I prefer a single PR that has the full code.

We keep the CPython Lib version and the asyncio repo synchronized, a PR for the asyncio repo is easier to handle for me.

Well, https://github.com/python/asyncio/pull/297 is ready :)

ReadStream.read_until() has been implemented instead enhancing readline().

The same Git Hub pull request is mentioned in Issue 26050, so I am guessing this is now a duplicate.

During development, we decide not to change readline() function. So, yes, this issue is closed.