Feature/support prepare statement

[membphis]

do support mysql's prepare statement.

there are two ways for calling:

-- first way
mysql:new
db:connect
db:run

-- another way
mysql:new
db:connect
db:prepare
db:execute
db:set_keepalive

PS: It's not implement completely.

[agentzh]

I think we can require packet_type to be always defined, thus saving this branch.

[membphis]

it'll be more simple if we can store data on special cosocket object.

for example:

local ok, err = sock:set_bound_data(key, value)
local value, err = sock:fetch_bound_data(key)

then we can store k-v like this:

key: prepare_sql
value: statement_id

-- inside db:prepare(prepare_sql)

local statement_id = self.sock:fetch_bound_data(prepare_sql)
if statement_id then
    return statement_id
end

... -- do prepare as current 

self.sock:set_bound_data(prepare_sql, statement_id)
return statement_id

@agentzh how about we take # attach arbitrary data to cosocket instances to ngx_lua_module?

[agentzh]

I think we already have Lua functions reading the result set. Can we reuse that?

[agentzh]

That is, reusing the read_result function and etc?

[tmiskoEconomia]

Hello when will this be official release?

[forkfork]

Would love to know whether this one seems likely to get merged.

[agentzh]

This will definitely get merged at some point since this PR was the result of my request to @membphis :) He's been doing me a favor.

[denisu]

Is there anything preventing this to merge? It would be cool to see it merged soon.

[agentzh]

Just for the record, we'll redo this atop a new OpenResty builtin API for better performance. We already use the new way to add preparing statement support for PostgreSQL.

[yukixia]

hello, when is added to the built-in function of this feature??

[nafey]

Are there any plans to complete this feature?

[vinsonzou]

hello, are there any plans to complete this feature?

[sunlin7]

Just for the record, we'll redo this atop a new OpenResty builtin API for better performance. We already use the new way to add preparing statement support for PostgreSQL.

Really look forward the PREPARE statement support for MySQL. : )