pooya-shams / massspring Goto Github PK

when receiving data from the server created by handle_client, I sometimes can not get the full data mostly because I just try to receive 4096 bytes of data and if I get all 4096 bytes of it, I assume that there is more and try to receive again until the received data is less than 4096 bytes.
this approach causes tons of bugs so I thought maybe I could change the way handle_client work so that it sends the length of the data first and then sends the actual data.

there is just the decoder for mass_lis response and not for spring or mass|spring, there should be such functions that implement decoding the responses similar to implementation of encoding them.

there are bunch of asserts in functions in all files that use type(obj) == type to check if the passed object is from the right type or not. python suggests that it's better to use isinstance instead of this.
the code will change from something like this:

assert type(obj) == type, TypeError("error msg")

to this

assert isinstance(obj, type), TypeError("error msg")

you can not receive both mass positions and spring positions in one place that causes a lag and difference in positions of masses and springs that are connected to those masses.

the response that analyse_request returns is just simple concatenation of all little responses from functions. they are not really easy to split and determine which part of the response is for which request.
I can do 2 things to solve this.

• 1 : use another character as delimiter of the responses and concatenate them using this special character.
• 2 : send multiple packets. client can not send all it's requests for one at once and receive all of them in that frame but it can send all of them at one request and receive in multiple responses. so we will send a response for each of the requests that the client sent us. However, there is a little problem with this approach. there are not a fixed number of responses being transferred so if the client makes a mistake and doesn't do a recv and/or do an extra one, both programs will halt and misfunction. but it is possible to manage it.

this function won't respond to the dictionary passed as the kwargs for the mainloop function, whatever they are. seems to be because the dictionary which is passed to the start_server_mainloop will not be used in the function and is not passed anywhere.

it is designed to work with earlier versions of analyse_request and should be updated.

analyse_request function in networklib.py can only handle three types of requests which are all hardcoded. also the delimiter ('|') is not optional and can not be changed.
this function should be refactored in a way that can handle every possible request for the value of any variable specified by the user.
this can be done in a way that uses eval function but I hate eval and exec so I will probably implement it in a way that the function gets a dictionary with bytes objects as keys and functions as values. the function will act like getters for specified variable.
for example assume that the user wants the server to be capable of sending the width of the window. in this case the dictionary that will be passed to analyse_request and the function that is given as the value for "WINW" should be something like this:

def get_winw():
    return ms.WINW
{b"WINW": get_winw}