https://github.com/FI-Mihej/InterProcessPyObjects

https://pypi.org/project/InterProcessPyObjects

This high-performance package delivers blazing-fast inter-process communication through shared memory, enabling Python objects to be shared across processes with exceptional efficiency. By minimizing the need for frequent serialization-deserialization, it enhances overall speed and responsiveness. The package offers a comprehensive suite of functionalities designed to support a diverse array of Python types and facilitate asynchronous IPC, optimizing performance for demanding applications.

Comparison With "multiprocessing.shared_memory"

While both "InterProcessPyObjects" and "multiprocessing.shared_memory" facilitate inter-process communication, there are several key differences to note. Unlike "multiprocessing.shared_memory", "InterProcessPyObjects" offers the following enhancements:

• High-Performance Mutable Objects: Both connected processes can modify shared objects at runtime, and these changes are immediately reflected on the other side. This feature not only increases flexibility but also delivers exceptional performance, with the capability to handle up to several million changes per second.
• Synchronization Features: Ensures that operations are thread-safe and data integrity is maintained across processes.
• Message Queue: Integrates a system for queuing messages, making communication between processes more structured and reliable.
• Extended Type Support: Supports a broad range of data types, including custom classes, which goes beyond the basic types typically handled by multiprocessing.shared_memory.

These features make InterProcessPyObjects a more robust option for developers requiring advanced inter-process communication capabilities.

API State

Stable. Guaranteed to not have braking changes in the future.

License

Licensed under the Apache License, Version 2.0.

No network access from the WSL to the Internet. Both domain and IP based ping commands failed. Other advices from the Internet does not help. Newly installed distro from the Store has same problems as well.

Cure (in my case at least)

1. Backup distros you need (See Backup section)
2. Uninstall an each installed distro
3. Go to Apps -> Programs and Features -> Turn Windows features on or off, make a screenshots of the current state and ensure next features unchecked:
   
   1. Virtual Machine Platform (Since it was on my Windows Home)
   2. Windows Hypervisor Platform (might be required if was in use)
   3. All Hyper-V related features maybe (in order of Windows Pro)
   4. Windows Projected File System
   5. Windows Subsystem for Linux
4. Turn off your PC and then turn in on. Preffered than Reset in order to trully reset thirdparty hardware like network controllers (Wi-Fi/Ethernet)
5. Reset your network settings:
   
   1. Run CMD as an Admin and execute commands from the "Network Reset Commands" section bellow
   2. Hit "Windows" key, type "Network reset" and hit "Enter". Click "Reset now" button.
6. Turn off your PC and then turn in on
7. Go to Store and install not required distro (Debian in my case since I'm using a Ubuntu)
   
   1. Start distro from the Store by clicking an appropriate button
   2. Ensure that ping is working properly
8. Restore previously backupped distros (See Restore section)

Network Reset Commands

ipconfig /flushdns (Clear the dns cache)  
nbtstat -RR (release and refresh NetBIOS names)  
netsh int ip reset (reset ip settings)  
netsh winsock reset (Reset Winsock Catalog)  
netsh interface ip delete arpcache (reset Address Resolution Protocol cache)  

Sources

1, 2

wsl --list  

Result:

Windows Subsystem for Linux Distributions:  
Ubuntu20.04LTS (Default)  
Debian  

We will work with Ubuntu20.04LTS

Shutwown WSL

wsl --shutdown  

Export (Backup) image

Template:

wsl --export <Image Name> <Path to exportable .tar image file>  

An actual command in my case:

wsl --export Ubuntu20.04LTS c:\back\Ubuntu20.04LTS.tar  

Import (Restore) image

Import image

Template:

wsl --import <Image Name> <Directory where you want to store the imported image> <Path to exported .tar image file>  

An actual command in my case:

wsl --import Ubuntu20.04LTS c:\WSL_Dir_With_Read_Write_Permissions\Ubuntu20.04LTS c:\back\Ubuntu20.04LTS.tar  

Distro will be started automatically with a root prompt

Set default user

1. Execute awk -F: '/\/home/ {printf "%s:%s\n",$1,$3}' /etc/passwd and find decimal ID of required user. 1000 in my case (line my_user_name:1000)
2. Enter exit command
3. Ensure you are in CMD without Admin privileges and execute wsl --shutdown command
4. Launch 'regedit'
5. Go to HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Lxss
6. Find your distro ID and go to an appropriate subkey
7. Ensure value of name 'DefaultUid' and of type 'REG_DWORD' with value data 1000 with base Decimal
8. Close Regedit

Set default distro

wsl --setdefault Ubuntu20.04LTS  
wsl --shutdown  

Start WSL:

wsl  

Ubuntu 20.04 LTS will start with your chosen user prompt

Sources

1, 2, 3

We have some third-party project (made with C or C++) with syncronos network code which produces big memory consumption by opening a thread on each incomming connection. As result of this behavior, OS kills our application.

The classic solutions

Actually we have two main different solutions:

1. Rewrite this third-party project in a callback based asynchronous code. This is a hard way. It needs alot of time for investigation, alot of time for development, alot of time for debugging and alot of time for testing. Also, after this work we will be unable to update this third-party project to a newer version without investin the same amount of resources again.

2. We can move from the threads to coroutines and implement coroutine manager and a stack compression code. It need time for development, debugging and testing almost only for a coroutine manager and a stack compression implementation. Ofcourse we still need some time for original sources investigation but only once. Also we will be able to update this third-party project in a future without significant time investment.

Of course coroutines are preffered.

So what's wrong with coroutines on iOS?

Usually we can use POSIX ucontext related calls for coroutines implementation. Or fibers on Windows.
The bad thing is that we can't use ucontext related calls on iOS: although they exist in SDK and can be included and calld, their code is not implemented internally at all - those calls are just stubs for returning an error. This behavior is unfortunatelly undocumented in Apple docs.

Mmm. And what can we do with this?

Actually solution exists.

In a C++ world, there is a well known library - Boost.
For a C++ project we can use Boost.context which is stated and actually has an iOS support.

Boost.coroutine hovewer, does not support iOS at the moment of this post writting.

OK. What about C?

Actually almost the same. Boost.context functions only thanks to several assembly files and a header. We just need to include needed assembly files and a header into our C code build config (and of course include header into our C code sources).

Assembly files are in the folder
src\asm We are interested in three arm64 related files for macho executable file format (Apple made file format):

jump_arm64_aapcs_macho_gas.S  
make_arm64_aapcs_macho_gas.S  
ontop_arm64_aapcs_macho_gas.S  

Header:
include\boost\context\detail\fcontext.hpp

Header should be slightly modified to be compatible with C code of course.

For Xcode we just need to drag all needed files into the list of files of our project.

TimeoutError exception is raised as result of receving 'ETIMEDOUT' socket error which can be generated by connect() and recv() calls.

It can be raised on a connected socket only (TCP connection for example)

This exception can be raised as result of:
1. network cable unplugged (on a client or/and on a server side)
1. client device power off
1. router power off
1. client application crash or kill
1. etc.

This exception will be raised after a while of the actual cause. Usually after a 10 minutes after the actual client disconnection. Actual timeout is OS dependent.

Incomplete list of Python functions that can raise TimeoutError:
socket.connect, socket.recv, socket.recv_into

Also, I believe, asyncio corresponding functions are inherit the same behavior.

I've encountered this exception in my ASockIOCore during the testing phase: as result of the network cable unplugging on a client side.

Conclusion:

During the development I've not realized that Python documentation can lack a such kind of crucial information. It would be nice to add to the description of each function, a complete list of exceptions that can be raised by this function.

More information on https://github.com/FI-Mihej/FI-ASockIOCore

I was going to write a post later. But I replied to a message and then made a topic at the nexusmods.com forums anyway. So...

I made UE3 compiler and decompiler with a handy utilities set. Part of the utilities are using UPKTools as a backend (thanks to wghost81 for a such good tool set!).

I'll release it soon. So no more problems with integers or floats, etc. anymore.

Example tokens:

• float: FloatConst(("0.4059738695"))
• integer: IntConst(("2325839"))
• integer HEX: IntConst((09 0f 04 a0))
• string: StringConst(("色の良いセット Ывдлфо адЦыфо у sljf aeOirJa rfjh "))
• auto label: @to_label(35)
• named label: @to_label("Script Ret")

Code between /DESCRIPTION_BEGIN/ and /DESCRIPTION_END/ tags is for a dev-helping purposes and ignored by a compiler.

Example of formatted code:

Example of the Description part of decompiled code:

Example mod:

Using this compiler I made http://www.nexusmods.com/xcom/mods/663/?. You may use it as an example.
Source code you can see and download here:
Realistic Damage Model

• Diff-installable source code
• Full function source code

UCB source code files can be highlighted by UCB Source Code highlighter

Links:

Topic with discussion on nexusmods' forum

The higher the accuracy is – the higher the damage is. But don’t relax: there is always a chance of maximum damage.

I made my first game mod with my compiler for Unreal Engine 3 bytecode.

Source code can be seen in the "Src" folder content. And in the .ucb file inside of "Universal Release" folder.

Formula can be found in the "Notes" folder content.

Charts been made with Matplotlib for Python 3.

Also I can say that Matplotlib's documentation is unfortunately is not very handy to use in some infrequent situations.

Links:

Mod's repository

Mod's page on Nexusmods

Why this is happens?

Simple:
Illustrator tries to check is user have the last GPU driver by comparing drivers versions: version of installed driver with version of the last generic driver available online.
But. Laptops are like mobile phones: manufacturer modifies generic drivers provided by Intel, nVidia, AMD to be compatible with custom hardware of specific laptop model. In the result, last versions of generic drivers just are not compatible and can't be installed on the laptop.
Thanks to butz for info: Re: When I start illustrator, I get a message that my graphics card drivers are not up to date.

A possible solutions of the problem:

1. Simple: Illustrator should not check drivers. And just run without any tests. That's all. Simple. But it can be used only as a fast fix for a no more than for a couple of weeks.
2. Good: Illustrator should use good well-known practice of software industry. It shouldn't try to check versions of the drivers. It should test for the presence of the required feature itself. And then use available features.
   To be more convincing for non-technical professionals I quote from the MSDN (official Microsoft Developers Network Portal):

   "Identifying the current operating system is usually not the best way to determine whether a particular operating system feature is present. This is because the operating system may have had new features added in a redistributable DLL. Rather than using the Version API Helper functions to determine the operating system platform or version number, test for the presence of the feature itself." Operating System Version (Windows)

A list of some open topics on this subject:

1. When I start illustrator, I get a message that my graphics card drivers are not up to date.
   Created: 08.07.2015 12:34
   Views as on December 13, 2015: 7915
   State: not resolved
2. GPU Acceleration disabled since today with the 2015.2 update
   Created: 01.12.2015 16:45
   Views as on December 13, 2015: 456
   State: not resolved
3. Illustrator GPU acceleration not working
   Created: 12.12.2015 2:47
   Views as on December 13, 2015: 76
   State: not resolved
4. Illustrator Issues after the adobe updates
   Created: 08.12.2015 17:05
   Views as on December 13, 2015: 133
   State: not resolved
5. GPU Acceleration don't activate
   Created: 21.10.2015
   Views as on December 13, 2015: 64
   State: not resolved

My original post on Adobe forum

        with lzma.open(file_name, 'rb') as file:
            data = pickle.load(file)

• 40Mb file
• 90000 of objects inside the file
• File created by CPython-optimized code

Problem:

This code needs only 6.43s when launched under CPython, but in needs more than 5min to complete under PyPy3 (in fact, I did not even wait for the end: I just shut it down after 5 minutes of running)!

Solution:

        with lzma.open(file_name, 'rb') as file:
            raw_data = file.read()
            data = pickle.loads(raw_data)

This code needs 6.35s when launched under CPython and 39s under PyPy (of which 36s is spent on pickle.loads() ).

Also recreating file under PyPy is not the case:

• this is a very slow operation (82s for pickle.dumps() and 74s for file.write() );
• it leads to slightly slower reading (54s for pickle.loads() ) under PyPy.

Conclusions:

It is good to prepare pickled data under CPython.