Dwedit

What does "Browser's Main Thread" even mean when there are over 30 processes running?

Brave Browser offers "Brave-hosted" versions of uBlock Origin and uMatrix.

What about Win32s? Then the same binary can run on Windows 3.1 and Windows 11.

C# does closures by creating a class. When you run the function, captured local variables actually live inside the object instead of in the stack. This makes use of an object, but no memory pages need to become read/write/executable to make C#-style closures happen. You just have a combination function/object pointer (a delegate) instead of a single function pointer.

What do you need the executable stack for? You call using a function pointer, there's no executable read/write memory involved in using a function pointer.

How about allowing browsers to actually display markdown?

That's the nice thing about LLMs, you're always absolutely right.

The topic of an article being AI and an "AI-Generated Article" are two completely different things.

Many many different ways to allocate memory. malloc, new, HeapAlloc, LocalAlloc, GlobalAlloc, CoTaskMemAlloc, etc. GlobalAlloc and LocalAlloc are pretty much obsolete. "malloc" sometimes redirects to HeapAlloc with the default process heap, but only sometimes, it depends entirely on which compiler and CRT is used.

It just causes one big headache that a DLL can't simply return a pointer and have the caller use "free", because malloc/free can be incompatible across modules.

This just led to the COM standard, featuring reference-counted objects that don't care what underlying way was used to allocate the memory. For the situations where you do need memory buffers, COM enforces the use of CoTaskMemAlloc/CoTaskMemFree.

Lossless Webp is still very good, and it decompresses very quickly. Lossless JXL compresses better, but decompresses much more slowly. Lossless AVIF is a joke.

Then you load User32.dll, and you get 2MB of commit and 1MB of private bytes. Throw in a Message Box, and it's 272KB more private bytes and 388K more commit.

Windows 10 makes it literally impossible to write something that uses low memory and uses an actual window.

I made my own 64-byte MS-DOS stub that uses a shorter message "Win32 Only!" instead of the usual "This program can only be run in DOS mode" message. By using the shorter stub, using the secret flag to omit the Rich header, and possibly merging sections together, I can usually get the PE header to fit within 512 bytes, and the linker won't pad out the PE header to 1024 bytes.

A copy can be found at https://github.com/Dwedit/NoCopilotKey/blob/main/stub.bin

I've worked with C compilers that refused the // style before.

Key is Scancode 0x2B, or as a VKEY code: VK_OEM_5 (0xDC)

Javascript says the key I pressed was keycode 220 (same number as 0xDC) or key "\".

I had Tandy 1000 as the selected machine. When I look at the onscreen keyboard, I don't see a backslash key underneath backspace, but I do see one next to left shift. For keyboards that have the alternate backslash there, it is listed as having a scancode of 0x56 and a VKEY code of VK_OEM_102 (0xE2). (I haven't actually checked information specific to the actual Tandy keyboard)

This is about early IBM PCs, not apple computers. And the non-working key is the Backslash key \ , not the backspace key.

Pentium 4 had lids. AMD Athlon XP did not.

Mysteriously, backslash has been mapped to backspace for some reason.

I mean I don't bother reading the dialogs. If there's modifications, "Yes" means save, "No" means close the program without saving, and "Cancel" means go back. And almost every program agrees on that convention, so you don't need to read the dialogs.

Rip John C. Dvorak, writer of the Abort/Retry/Fail column from PC Magazine.

On a different note, never underestimate the value of a modal dialog that gives you a chance to retry an IO operation that failed. Loose USB cables and network shares over Wifi can make those happen a lot more often.

Grouped Details example really needs a hand cursor. An "I" cursor for selecting text doesn't tell you that an element is interactible.

Saw headline, expected branch prediction vulnerability involving Unicode, was surprised at a completely different topic.

I had to do this once. SSH server needed to run on port 443 because the clients had a locked down network firewall and could only make connections to particular port numbers.

Or even better, something that pretends to be a real "Sign in with Google" button but instead phishes your username and password.

Would probably need a full userscript at this point?

Inline assembly is inline. You're not following the platform ABI's calling convention here, you are choosing input registers, output registers, and trashed registers right there. Following the platform's ABI and carrying out the function call has a cost, merely picking registers does not.

This is what happens when you start a separate instance of Chromium everywhere.

For DOS, your choices are basically DJGPP C compiler, or using the HXDOS DOS extender. But how well would Cave Story work on 50MHz?

Until 2023, Edge was sending every URL you visited to bingapis.com.

Microsoft's current privacy statement also says "Depending on your settings, this browsing history is sent to Microsoft"

C++ ABIs are why Win32 settled on things like COM to do cross-module passing of objects. Because nobody could agree on a standard, the COM standard enforced a very specific calling convention, and a very specific vtable layout for COM objects.

Would exposing people to an obviously hallucinating AI help people here? Then they won't blindly trust it so much.