LUBE (Networking Library)

[TechnoCat]

I believe nested tables will require more effort on your part.
There are lots of options already though:
http://lua-users.org/wiki/TableSerialization
http://love2d.org/wiki/TLTools

[Deif]

ah i think i've got this now, im just having trouble sending a big table back with all the X,Y coords of people.

Code: Select all

connected = {
    {10, 10},
    {20, 20}
}

function send()
	lube.bin:setseperators(string.char(30),string.char(31))
	local data = lube.bin:pack( connected )
	server:send(data)
end

That is basicly what i'm sending, but for some reason it doesn't allow it?

You would need to pack the {10,10} table and the {20,20} tables first with those seperators. Then pack them again using different seperators. Basically you have to turn each table into a string (which is what the packing does) and the seperators are an easy way to tell where a table begins and where a table ends so you can turn them back into tables on the other side.

[Robin]

You would need to pack the {10,10} table and the {20,20} tables first with those seperators. Then pack them again using different seperators. Basically you have to turn each table into a string (which is what the packing does) and the seperators are an easy way to tell where a table begins and where a table ends so you can turn them back into tables on the other side.

TechnoCat's answer is better. What if the nesting goes even deeper? Serialization can handle those cases, just using different separators, not so much.

[bartbes]

Except this is serialization. I never claimed it was a good implementation though, quite the opposite.

[Deif]

You would need to pack the {10,10} table and the {20,20} tables first with those seperators. Then pack them again using different seperators. Basically you have to turn each table into a string (which is what the packing does) and the seperators are an easy way to tell where a table begins and where a table ends so you can turn them back into tables on the other side.

TechnoCat's answer is better. What if the nesting goes even deeper? Serialization can handle those cases, just using different separators, not so much.

Then change the separators again. I think making this example too complicated is going to confuse people - besides, what's the point in the library serialisation methods if people are going to use more robust methods, when in reality you're not going to be sending massively complicated nested tables across a network anyway.

[TechnoCat]

I personally find Taehl's solution to look the most elegant. https://github.com/Taehl/TLTools
Only 21 lines! But it looks like it is incredibly easy to break with including certain characters in strings.

[slime]

For TSerial I had to change line 9 of TSerial.lua from

Code: Select all

if type(k) == "string" then k = k

to

Code: Select all

if type(k) == "string" then k = string.format("[%q]", k)

, and line 13 to

Code: Select all

if type(v) == "string" then v = string.format("%q", v)

for it to work with all strings.


EDIT: also, line 14 needs to be

Code: Select all

elseif type(v) == "table" then v = TSerialize(v)

instead of

Code: Select all

elseif type(v) == "table" then v = Tserialize(v)

[bartbes]

Don't you mean

Code: Select all

if type(k) == "string" then k = string.format("[%q]", k)

?

[Robin]

Except this is serialization. I never claimed it was a good implementation though, quite the opposite.

Well, yeah. But I specifically meant the serialization implementation TechnoCat linked to.

Then change the separators again. I think making this example too complicated is going to confuse people - besides, what's the point in the library serialisation methods if people are going to use more robust methods, when in reality you're not going to be sending massively complicated nested tables across a network anyway.

If you want more than a simple table with only string and number values, changing separators gets hairy and unwieldy quickly. Better use a library that's easy to use right away in that case.

[Bambo]

Gah, decompression doesn't seem to work

Code: Select all

		
		local compressed = data                                      -- this is the compressed table
		local decompressed = TLibCompress.DecompressLZW(data)            
		tablei = loadstring(decompressed)()