* Although the markdown parser works perfectly,
the official Telegram client does not fully reflect it.
However, if you still think that this is a lie, go check
the markdown parser and test it yourself!
The interactive example allows you to list the top
dialogs and send messages to them until "!q" is entered
More documetation has been added
Fixes when representing TLObjects (lists did not represent well)
The `send_code_request` now allows to use multiple phone numbers at once
README.md was updated to reflect more useful information
More errors from the official Telegrm website have been added
MtProtoSender now handles updates (and doesn't crash!)
Fixes on TcpClient to be able to receive whole large packets
Updated scheme.tl to the layer 55
Session is now saved more often (to prevent damages from crashes)
Fixes to the code generator (generated invalid code for reading "bytes")
The code generator now handles okay the flags using True type
Also, double checking for the flag is now avoided in cases where the
flag was a Vector type
Some fixes include, in more detail:
- Using little over big endianess in some parts
- Flagging all the constructor numbers as unsigned
- Fixed bugs with factorizer
- Implemented TLSharp's RSA
The code generated by the generator now also writes the files
on_response(...) method. Also, all the generated files are
saved in a dictionary containing `constructorId: class`
The code generated by the generator now classifies the output files
in their corresponding categories, also writing their __init__(...)
with documented arguments, and the on_send(...) method
A TLObject class has been added, as well as a method for tokenizen .tl files.
A TLObject can be created by simply calling `TLObject.from_tl(tl_line)`,
this will tokenize the .tl line and return a new TLObject with all
the required properties extracted.
Also added another method to load .tl files and yield TLObjects, including
a SourceBuilder that can be later used to generate Python code.
The initial release contains the most basic implementation of TLSharp core.
This is also fully untested, since no test can be done until more work is done.
