Variety of access logging issues

[thehesiod]

Issue 1

Looks like duplicate log format names was accidentally introduced: both %t and %T resolve to request_time see https://github.com/aio-libs/aiohttp/blob/master/aiohttp/helpers.py#L293-L299

each one should have a unique name, thoughts? Perhaps %T[f] + %D should be request_elapsed[_frac]

another idea is perhaps we should rename this to _LOG_FORMAT_MAP for 3.x?

Issue 2

%t is not the start time of the request, it's just the time the log entry was created (not meaningful). I suggest fixing this to be the time the request was received.

Issue 3

the base AccessLogger is not using self in __init__ so it's impossible for subclasses to replace methods like _format_t

Issue 4

logging of multiple header values doesn't serialize to extras, last one overwrites the others.

[thehesiod]

I've created a PR with a rough idea of a fix

[asvetlov]

Let's discuss issues one by one.

1. Log names was modeled after Apache log format: http://httpd.apache.org/docs/current/mod/mod_log_config.html

2. Access Logger was not intended for inheritance, partially for historical reasons. That's why we don't use self in init but use class-level caching. The proper solution is moving a logger instantiation from web.RequestHandler to web.Server -- logger format parsing should be done once on server startup, not on every connection establishment.
   We can even rewrite the AccessLogger completely from scratch keeping AbstractAccessLogger API. It may speed up log generation as well (the best solution is replacing logger with f-strings as shown in https://docs.aiohttp.org/en/stable/logging.html#format-specification).

[thehesiod]

@asvetlov

1. The issue isn't with the log specifiers, it's with the python logging extras key-name, they need to be unique in the dictionary.
2. (actually 3) I'm not ready to re-write/architect the logger, I can revert this change for now

how about issues 2 + 4 ?

[thehesiod]

sorry, forgot to sync change for 4

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a [new issue] for related bugs.
If you feel like there's important points made in this discussion, please include those exceprts into that [new issue].
[new issue]: https://github.com/aio-libs/aiohttp/issues/new