Extending our critical section based on WaitOnAddress to support timeouts

Raymond ChenMay 31, 2017May 31, 201705/31/17

Let's take the critical section we constructed in terms of Wait­On­Address and add two new functions: The first function is a simplification of the existing Enter­Alt­Cs function. It simply gives up if the critical section is not available. The second function is a modification of the existing Enter­Alt­Cs function...

Why isn’t the original window order always preserved when you undo an Aero Shake?

Raymond ChenMay 30, 2017May 30, 201705/30/17

A customer reported that when they used Aero Shake to minimize all the windows on their desktop, and then used it again to restore all the windows, the restored windows didn't always have exactly the same z-order as they did originally. The customer wanted to know whether this was expected. It's not expected, but it's not unexpected either. ...

Does DebugBreak work to launch the debugger, or doesn’t it?

Raymond ChenMay 29, 2017May 29, 201705/29/17

Jorge asked why the Debug­Break function stopped working. Specifically, why it doesn't launch the Visual Studio debugger. Okay, first of all, the primary purpose of the Debug­Break function is not to launch the debugger. The primary purpose of the Debug­Break function is to trigger a break into any attached debugger. That still works the ...

Diagnosing why you cannot create a stable subkey under a volatile parent key

Raymond ChenMay 25, 2017May 25, 201705/25/17

A customer encountered crashes in their program's initialization code. They weren't able to reproduce the problem in-house, but their failure logs suggested it was coming from here: The call was failing with System.IO.IO­Exception: Cannot create a stable subkey under a volatile parent key. The corresponding Win32 error code is ...

Some questions about unflushed data and calling FlushFileBuffers on a new handle to a file

Raymond ChenMay 24, 2017May 24, 201705/24/17

Consider the following sequence of events: First question: Under these conditions, will the data written in step 2 be lazy-written to disk? Or is it at risk of being lost forever because the handle wasn't closed? Let's look at the last part first. Whether the process closed the handle before crashing doesn't affect the story, ...

Debugging a GDI resource leak: Case study

Raymond ChenMay 19, 2017May 19, 201705/19/17

I was asked to help debug a problem. A program was leaking GDI bitmaps like crazy, and after a while, the GDI resource handle count reached 9,999, at which point GDI said, "That's it, I'm cutting you off." The problem isn't discovered until after the limit has been reached. To debug this, I'm gong to use different poor man's way of ...