Windows 8 compatibility

[Ilya Bizyaev]

Hello!

Recently I've made a simple program to encrypt files. It used a XOR operation. Nothing interesting, you know...

But then I decided to use AES cipher. I've found the Crypto++ library, used it - and got a much more powerful encryption.

My program is supported by such operating systems as Ubuntu, Windows XP, Vista and 7. However, before using Cryptopp, it was also supported by Windows 8. Now it crashes.

What's the problem? Does Cryto++ support Windows 8?

P.S. Excuse me for my English, I'm Russian.

[car...@keepkey.com]

I have been using Crypto++ 5.6.2 daily on Windows 8 for several months. I haven't encountered any issues at all. FYI I'm using Crypto++ to generate test vectors for a crypto library I'm maintaining. AES, ECDSA, MAC, SHA, etc.

[Ilya Bizyaev]

My program crashes on generating random iv; if I comment this line, it crashes on creating an cfbEncryption. This problem occurs only on Windows 8, and even "Compatibility Mode" doesn't help. So, this problem is definitely platform-dependent.
How do you connect this library - as a static or as a shared one? And can it influence this problem?

[Ilya Bizyaev]

You compiled your programs under Windows 8, didn't you?
I compile them under Windows 7.

[Ruben De Smet]

On 06/02/15 19:24, Ilya Bizyaev wrote:
> You compiled your programs under Windows 8, didn't you?
> I compile them under Windows 7.

I compile my software using mingw64 on CentOS using Jenkins. Using them
on Windows 7 and 8 works, as does it on Windows XP, since I patched the
compiler for that last OS.

I don't know about compiling on Windows 8. I tried it, but my compilers
don't like me on that platform.

What compiler do you use?

[car...@keepkey.com]

I compile on Windows 8 using the free Visual Studio Community 2013 edition (a rename of Visual Studio Pro). Furthermore, I use static libraries.

[Ilya Bizyaev]

I compile it under MinGW.

 

So, how can I solve this problem?

суббота, 7 февраля 2015 г., 5:04:52 UTC+3 пользователь car...@keepkey.com написал:

[Ruben De Smet]

On 07/02/15 12:30, Ilya Bizyaev wrote:
> I compile it under MinGW.
>
> So, how can I solve this problem?

I would use something like gdb to debug the problem. This is probably a
MinGW problem. You'll have to contact the MinGW list if you find this to
be a compiler problem: mingw-w6...@lists.sourceforge.net

Please put me in Cc, I'd like to follow up on this issue too, I could
become affected too in the near future.

[Ilya Bizyaev]

суббота, 7 февраля 2015 г., 16:47:30 UTC+4 пользователь ruben.de.smet написал:

> I would use something like gdb to debug the problem. This is probably a
> MinGW problem. You'll have to contact the MinGW list if you find this to
> be a compiler problem: mingw-w6...@lists.sourceforge.net
>
> Please put me in Cc, I'd like to follow up on this issue too, I could
> become affected too in the near future.

Unfortunately, that Windows 8 computer belongs to my friend, so, I can't debug it right now. However, I would definitely test it with GDB next week.

[Ilya Bizyaev]

The problem might have been caused by another library and incrorrect processing of idle events. I'm trying to work it out, will reply in a few days.

[Ilya Bizyaev]

I have debugged it on Windows 8! The error really occures in Crypto++, here is the screenshot from GDB:

[Ruben De Smet]

On 20/02/15 10:12, Ilya Bizyaev wrote:
> I have debugged it on Windows 8! The error really occures in Crypto++, here
> is the screenshot from GDB:
>

> <https://lh3.googleusercontent.com/-zsh9k4Hhgfk/VOb6ZSrnQ8I/AAAAAAAAAug/Y6QpB2jIfhY/s1600/Error%2Bscreenshot.png>

Are you sure the rest of your application is memory-fault free? Anyway,
this indeed looks like a memory fault in cryptopp.

[Илья Бизяев]

You see, my application works fine on Ubuntu and all Windows OS's, except Windows 8. Personally, I haven't found any memory leaks, as well as CppCheck static code analyzer. My program doesn't walk arrays with iterators, so, there can't be any segmentation faults in it. I think that is Cryptopp's internal error. I have already found one, but solved it with dynamical memory allocation (found solution on this forum).

пт, 20 февр. 2015, 18:36, Ruben De Smet <ruben....@telenet.be>: