It's about use of CreateIpcQueue() in Windows10.

[icckitagawa]

Thank you very much, madshi.

MadCodeHook is improved and is wonderful me, too.


Umm, it's strange. We're 15063 builds, too.
The one applied on July 12 is KB4025342. 32bit is also 64bit, too.
What is different?
Feeling of Microsoft isn't understood.


An experimental report.

The next was set on the SendIpcMessage function side.

SetMadCHookOption (SET_INTERNAL_IPC_TIMEOUT and L"500");

I moved with no problem with this.



Thank you very much.

[icckitagawa]

I'm sorry, the parameter of the function is wrong.

X SetMadCHookOption(SET_INTERNAL_IPC_TIMEOUT and L"500");
O SetMadCHookOption(SET_INTERNAL_IPC_TIMEOUT,L"500");


Also there is a thing I'd like to hear.


"2.8.2.9beta" Madshi corrected would like to know the point corrected compared with "2.8.2.0 official".
(I think 9 times do from a build number, and are corrected.)

I think a build will do my product using 2.8.2.9beta.
As well as being official I'd like to know a corrected point.
When there is an easy release note, I'm happy.


Is there a person who doesn't use 2.8.2.9beta or a case that you say?
When I say that it's better to wait for new official, I want advice.
If official release is close, I wait for that, and a build does a product.

Thank you very much.

[madshi]

Please use (LPWSTR) 500 instead of L"500". I know, it's weird, but that's how the option API works.

The next official release is practically done. You can already download it here:

http://madshi.net/madCollectionBeta.exe

I'm just waiting for confirmation from one user, then I'll change the version number of the installer and announce it as the next official version.

[icckitagawa]

Thank you very much for your answer, madshi.

The matter of a parameter was understood.

The situation was understood. I'm thankful for consideration of a pre-release.
I consult with my customer about adoption.

There is much correction this time.
madshi is doing much difficult work.
It's serious, but I'm supporting sincerely.

May the Force be with you.

[madshi]

May the Force be with you.

Haha, thank you! Can't wait for December ("The Last Jedi" movie).

[_NN_]

Accidentally found this topic.
Is there any conclusion here ?
Is there a real bug in Windows ?
If yes, given a small reproduction source, you could ask MS to investigate.

[madshi]

This thread is about a year old, I don't recall all the fine details, but IIRC it seemed to be a bug in Windows. I don't know if it's fixed in the latest Windows 10 build now. I did do some improvements to the IPC logic.

[icckitagawa]

I add an opinion of madshi.

This problem was the bug of WindowsAPI.

When it's a source of reference, it's possible to acquire it from the next page.

viewtopic.php? f=7&t=28385&start=15
http://madshi.net/LpcBug.zip


The bug was reported to Microsoft through my office one year ago.

This bug is Windows10, and when IPC communication beyond 240 bytes is performed, a memory is something to leak.
But it seems that this bug is corrected on the Microsoft side. (Secretly)

I'm repeating a test of memory leak. That continues so far from "nightmare one year before".
At least, memory leak doesn't occur up to now. (It's the environment of Japan.)


This topic is long wastefully, so it's difficult to read.
But the contents of this degree.

When it's helpful, I'm happy.

[madshi]

Thank you icckitagawa, that's very good to hear!