tdlight/README.md
Andrea Cavalli 45e855f89d Remove most memory related features
I can't maintain anymore this amount of features while keeping the library constantly updated and without bugs. Every merge was taking me multiple hours of revisioning the code. I give up.
From this commit onwards TDLight will only have small useful customizations that are easy to maintain.
Now the people relying on the OptimizeMemory method can restart the session every N hours to free up the memory.
The real way to keep a low memory usage must involve a huge refactoring to allow the unloading of the caches into the sqlite database, similar to what's already happening with messages data. Only Levlam has the ability to implement this without needing to merge the upstream everytime.
2021-09-25 22:11:42 +02:00

10 KiB

TDLight

TDLight is a fork of TDLib, a cross-platform library for building Telegram clients. It can be easily used from almost any programming language.

Table of Contents

Features

TDLib has many advantages. Notably TDLib is:

  • Cross-platform: TDLib can be used on Android, iOS, Windows, macOS, Linux, FreeBSD, OpenBSD, NetBSD, illumos, Windows Phone, WebAssembly, watchOS, tvOS, Tizen, Cygwin. It should also work on other *nix systems with or without minimal effort.
  • Multilanguage: TDLib can be easily used with any programming language that is able to execute C functions. Additionally it already has native Java (using JNI) bindings and .NET (using C++/CLI and C++/CX) bindings.
  • Easy to use: TDLib takes care of all network implementation details, encryption and local data storage.
  • High-performance: in the Telegram Bot API, each TDLib instance handles more than 24000 active bots simultaneously.
  • Well-documented: all TDLib API methods and public interfaces are fully documented.
  • Consistent: TDLib guarantees that all updates are delivered in the right order.
  • Reliable: TDLib remains stable on slow and unreliable Internet connections.
  • Secure: all local data is encrypted using a user-provided encryption key.
  • Fully-asynchronous: requests to TDLib don't block each other or anything else, responses are sent when they are available.

TDLight extra features

TDLight extra options

  • disable_minithumbnails (true/false) This setting removes minithumbnails everywhere. It reduces memory usage because tdlib keeps them in RAM
  • disable_document_filenames (true/false) If you don't care about having the original filenames of every file stored in RAM, you can disable them using this option. It reduces memory usage
  • disable_notifications (true/false) In TDLib pending notification updates are stored in ram until you "read" them. This option disables completely notifications and keeps the pending notifications queue empty, reducing memory usage
  • ignore_update_chat_last_message (true/false) If you don't care about have updateChatLastMessage updates enable this
  • ignore_update_chat_read_inbox (true/false) If you don't care about have updateChatReadInbox updates enable this
  • ignore_update_user_chat_action (true/false) If you don't care about have updateUserChatAction updates enable this
  • ignore_server_deletes_and_reads (true/false) If you don't care about receiving read receipts and remote deletes from other users, enable this, it will reduce memory usage
  • receive_access_hashes (true/false) Receive chats and users access hash as updates

TDLight extra API functions

TdApi.GetMemoryStatistics

This method is used to read the size of all the internal TDLib data structures. The output contains a string that can be parsed as a JSON.

  • Options:
    • ignore_inline_thumbnails: true
    • disable_top_chats: true
    • ignore_platform_restrictions: true
    • ignore_sensitive_content_restrictions: true
  • Disable all the databases (messages_db, users_db, files_db)

Examples and documentation

See our Getting Started tutorial for a description of basic TDLib concepts.

Take a look at our examples.

See a TDLight build instructions generator for detailed instructions on how to build TDLib.

See description of our JSON, C++, Java and .NET interfaces.

See the td_api.tl scheme or the automatically generated HTML documentation for a list of all available TDLib methods and classes.

Dependencies

TDLib depends on:

  • C++14 compatible compiler (Clang 3.4+, GCC 4.9+, MSVC 19.0+ (Visual Studio 2015+), Intel C++ Compiler 17+)
  • OpenSSL
  • zlib
  • gperf (build only)
  • CMake (3.0.2+, build only)
  • PHP (optional, for documentation generation)

Building

The simplest way to build TDLight is to use our TDLight build instructions generator. You need only to choose your programming language and target operating system to receive complete build instructions.

In general, you need to install all TDLib dependencies, enter directory containing TDLib sources and compile them using CMake:

mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .

To build TDLib on low memory devices you can run SplitSource.php script before compiling main TDLib source code and compile only needed targets:

mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build . --target prepare_cross_compiling
cd ..
php SplitSource.php
cd build
cmake --build . --target tdjson
cmake --build . --target tdjson_static
cd ..
php SplitSource.php --undo

In our tests clang 6.0 with libc++ required less than 500 MB of RAM per file and GCC 4.9/6.3 used less than 1 GB of RAM per file.

Using in CMake C++ projects

For C++ projects that use CMake, the best approach is to build TDLib as part of your project or to install it system-wide.

There are several libraries that you could use in your CMake project:

  • Td::TdJson, Td::TdJsonStatic — dynamic and static version of a JSON interface. This has a simple C interface, so it can be easily used with any programming language that is able to execute C functions. See td_json_client documentation for more information.
  • Td::TdStatic — static library with C++ interface for general usage. See ClientManager and Client documentation for more information.

For example, part of your CMakeLists.txt may look like this:

add_subdirectory(td)
target_link_libraries(YourTarget PRIVATE Td::TdStatic)

Or you could install TDLib and then reference it in your CMakeLists.txt like this:

find_package(Td 1.7.8 REQUIRED)
target_link_libraries(YourTarget PRIVATE Td::TdStatic)

See example/cpp/CMakeLists.txt.

Using in Java projects

TDLib provides native Java interface through JNI. To enable it, specify option -DTD_ENABLE_JNI=ON to CMake.

See example/java for example of using TDLib from Java and detailed build and usage instructions.

Using in .NET projects

TDLib provides native .NET interface through C++/CLI and C++/CX. To enable it, specify option -DTD_ENABLE_DOTNET=ON to CMake. .NET Core supports C++/CLI only since version 3.1 and only on Windows, so if older .NET Core is used or portability is needed, then TDLib JSON interface should be used through P/Invoke instead.

See example/csharp for example of using TDLib from C# and detailed build and usage instructions. See example/uwp for example of using TDLib from C# UWP application and detailed build and usage instructions for Visual Studio Extension "TDLib for Universal Windows Platform".

When TDLib is built with TD_ENABLE_DOTNET option enabled, C++ documentation is removed from some files. You need to checkout these files to return C++ documentation back:

git checkout td/telegram/Client.h td/telegram/Log.h td/tl/TlObject.h

Using from other programming languages

TDLib provides efficient native C++, Java, and .NET interfaces. But for most use cases we suggest to use the JSON interface, which can be easily used with any programming language that is able to execute C functions. See td_json_client documentation for detailed JSON interface description, the td_api.tl scheme or the automatically generated HTML documentation for a list of all available TDLib methods and classes.

TDLib JSON interface adheres to semantic versioning and versions with the same major version number are binary and backward compatible, but the underlying TDLib API can be different for different minor and even patch versions. If you need to support different TDLib versions, then you can use a value of the version option to find exact TDLib version to use appropriate API methods.

See example/python/tdjson_example.py for an example of such usage.

License

TDLib is licensed under the terms of the Boost Software License. See LICENSE_1_0.txt for more information.