Re: [Tarantool-patches] [PATCH 5/6] Fix luacheck warnings in src/lua/

[Vladislav Shpilevoy <v.shpilevoy@tarantool.org>]

Thanks for the fixes!

>>> +    },
>>> +}
>>> diff --git a/src/lua/argparse.lua b/src/lua/argparse.lua
>>> index faa0ae130..6c8f10fc1 100644
>>> --- a/src/lua/argparse.lua
>>> +++ b/src/lua/argparse.lua
>>> @@ -91,7 +91,8 @@ local function convert_parameter(name, convert_from, convert_to)
>>>  end
>>>  
>>>  local function parameters_parse(t_in, options)
>>> -    local t_out, t_in = {}, t_in or {}
>>> +    local t_out = {}
>>> +    t_in = t_in or {}
>>
>> 6. Unnecessary diff. At least when I check on top of the branch.
> 
> change is required, see [1]
> 
> 1. https://gitlab.com/tarantool/tarantool/-/jobs/589857153

It reports error 412. It is suppressed globally in .luacheckrc. How
is it possible? I tried it locally, and the warning is not reported.

In the job I see that the .luacheck file is outdated - it does not
contain 412 suppression:
https://gitlab.com/tarantool/tarantool/-/blob/b435c7d040c7ce160d67afde270c336f1b91fddb/.luacheckrc

While on the branch in github the warning is suppressed.

So the diff is unnecessary, after all. Unless the github branch is
outdated.
https://github.com/tarantool/tarantool/blob/ligurio/gh-4681-fix-luacheck-warnings-src/.luacheckrc#L15

The same for all the other places, where you said the change is required.
In the other commits too.

>>>      local stop = fiber.clock() + timeout
>>>      local dns = getaddrinfo(host, port, timeout, { type = 'SOCK_STREAM',
>>>          protocol = 'tcp' })
>>> @@ -1319,7 +1319,7 @@ local function lsocket_tcp_getpeername(self)
>>>      return peer.host, tostring(peer.port), peer.family:match("AF_(.*)"):lower()
>>>  end
>>>  
>>> -local function lsocket_tcp_settimeout(self, value, mode)
>>> +local function lsocket_tcp_settimeout(self, value, mode) -- luacheck: no unused args
>>
>> 10. Mode is really ignored, and this is not some kind of built-in
>> function with 'fixed' signature. And it is not used 'virtually',
>> when there is a variable keeping pointer at one function among
>> many having the same signature. So what is a purpose of keeping it?
> 
> removed mode arg

Yeah, looks like it is needed, after all. Igor says, it conforms
to http://w3.impa.br/~diego/software/luasocket/tcp.html#settimeout
or whatever. Personally, I would remove it anyway. Because in our
sockets and in our code globally this argument is never needed,
and does not conform to anything.