Lightnews — Scholar-powered news

Shafik Yaghmour

@shafik.bsky.social

970 followers 240 following 720 posts

Compiler Engineer clang front end. Interests: C++, C, and undefined behavior. Martial Artist, Book Worm and Dad. he/him

Mastodon: https://hachyderm.io/@shafik
Blog: https://shafik.github.io/

#cplusplus #compilers

Posts Replies Media Videos

Pinned

Shafik Yaghmour @shafik.bsky.social · Sep 10

C++ quiz #302

Answer here: hachyderm.io/@shafik/1151...

#Cplusplus
#Cpppolls

#302 Given the following in C++

struct S {
void f(int);
private:
void f(double);
};

void g(S* sp) {
sp->f(2); // Does overload resolution
// work even though f(double) is private?
}

Without checking:

A. Yes
B. No

Shafik Yaghmour

@shafik.bsky.social

I meet a microbiologist today

They were way BIGGER than I expected 🥁

January 28, 2026 at 8:56 PM

Shafik Yaghmour

@shafik.bsky.social

New curded phrase just dropped

"Aura attaches to a unique ontic causal chain"

ref: philosophy.stackexchange.com/a/135405

Is a screenshot of a picture the same picture?

Suppose one takes a screenshot of a digital picture (for example, an image displayed on a screen). Is the screenshot the same picture as the original, or is it a distinct picture that merely repres...

philosophy.stackexchange.com

January 28, 2026 at 8:43 PM

Shafik Yaghmour

@shafik.bsky.social

C++ quiz #310 const anonymous bit-fields, what say you?

Answer here: hachyderm.io/@shafik/1159...

#Cplusplus
#Cpppolls

#310 Given the following in C++

struct A {
const int :1; // Well-formed?
};

Without checking:

A. Yes
B. No

January 27, 2026 at 10:01 PM

Shafik Yaghmour

@shafik.bsky.social

"CEOs Say AI Is Making Work More Efficient. Employees Tell a Different Story": www.wsj.com/lifestyle/wo...

We know that workers actually tend to overestimate productivity savings so I would take this as an overestimate and assume reality is actually even worse.

These are really bad numbers.

#ai

Bar graph showing Workers estimate of how much time LLM saves them VS manager

Workers:

No Time: 40%
Less than 2 hours: 27%
2-4 hours: 20%
8-12 hoors: 9%
more than 12: 2%

Managers:

No Time: 2%
Less than 2 hours: 16%
2-4 hours: 33%
8-12 hours: 24%
more than 12: 19%

January 25, 2026 at 6:44 AM

Shafik Yaghmour

@shafik.bsky.social

Curl project brings out the receipts on the AI slop submissions they are getting:

gist.github.com/bagder/07f75...

#ai

AI slop security reports submitted to curl

AI slop security reports submitted to curl. GitHub Gist: instantly share code, notes, and snippets.

gist.github.com

January 25, 2026 at 6:34 AM

Shafik Yaghmour

@shafik.bsky.social

Why did the amoeba fail the math test?
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Because it multiplied by dividing 🥁

January 23, 2026 at 4:30 AM

Shafik Yaghmour

@shafik.bsky.social

"Deep C (and C++)" is a classic but still very relevant presentation about what it means to know C and C++ and what the right mindset should be about learning: slideshare.net/olvemaudal/d...

I still enjoy reading every once and awhile.

#cplusplus
#programming

Conversation on slide

On a scale from 1 to 10, how do you rate your understanding of C++?

Person 1: I rate myself as 8 or 9

Person 2: 4, maybe 5, I have just so much more to learn about C++

Bjarne: 7

Slide

Code:

#include <stdio.h>

void foo(void) {
int a;
++a;
printf("%dn", a);
}

int main(void) { foo(); foo(); foo(); }

Conversation:

Person 1: Now you get1, then 1, then 1

Narrator: Ehm, why do you think that will happen?

Person 1: Because you said they where initialized to 0

Narrator: But this is not a static variable

Person 1: ah, then you get three garbage values

Person 2: the value of a will be undeﬁnded, so in theory you get three garbage values. In practice however, since auto variables are often allocated on an execution stack, a might get the same memory location each time and you might get three consecutive values... if you compile without optimization.

Narrator: on my machine I actually get, 1, then 2, then 3

Person 2: I am not surprised... if you compile in debug mode the runtime might try to be helpful and memset your stack memory to 0

Slide

Code:

#include <stdio.h>

void foo(void) {
int a;
printf("%dn", a);
}

void bar(void) { int a = 42; }

int main(void) {
bar();
foo();
}

Execution results:

$ cc foo.c && ./a.out
42

Conversation:

Narrator: I am now going to show you something cool!

Person 2: Nice! I love it!

Person 2: You now want me to explain about execution stack or activation frames?

Narrator: I guess you have already demonstrated that you understand it. But what do you think might happen if we optimize this code or use another compiler?

Person 2: A lot of things might happen when the optimizer kicks in. In this case I would guess that the call to bar() can be skipped as it does not have any side effects. Also, I would not be surprised if the foo() is inlined in main(), ie no function call. (But since foo () has linker visibility the object code for the function must still be created just in case another object ﬁle wants to link with the function). Anyway, I suspect the value printed will be something else if you optimize the code.

Slide

Code:

#include <stdio.h>

void foo(void) {
int a = 41;
a = a++;
printf("%dn", a);
}

int main(void) {
foo() ;
}

Conversation:

Person 1: a gets an undeﬁned value

Narrator: I don’t get a warning when compiling it, and I do get 42

Person 1: Then you must increase the warning level, the value of a is certainly undeﬁned after the assignment and increment because you violate one of the fundamental rules in C (and C++). The rules for sequencing says foo(); that you can only update a variable once } between sequence points. Here you try to update it two times, and this causes a to become undeﬁned

Narrator: So you say a can be whatever? But I do get 42

Person 1: Indeed! a can be 42, 41, 43, 0, 1099, or whatever... I am not surprised that your machine gives you 42... what else can it be here? Or perhaps the compiler choose 42 whenever a value is undeﬁned ;-)

January 22, 2026 at 6:11 PM

Shafik Yaghmour

@shafik.bsky.social

Retro C++ quiz #53 The Final frontier

Answer here: hachyderm.io/@shafik/1159...

#Cplusplus
#Cpppolls

$Retro C++ quiz #53 Given the following in C++: #include <iostream> struct A { operator int() {return 1;} }; bool operator==(A, int){return 0;} What is the result of: int main() { std::cout <<(1 == A{})<<"\n"; } A. 1 B. 0 C. Ill-formed D. Depends on standard version$

January 22, 2026 at 5:30 PM

Shafik Yaghmour

@shafik.bsky.social

"'Everybody Loses' chronicles the rise of America's sports betting boom": www.npr.org/2026/01/19/n...

This is such a corrupt industry.

'Everybody Loses' chronicles the rise of America's sports betting boom

NPR's A Martinez speaks with journalist and author Danny Funt about his new book, "Everybody Loses: The Tumultuous Rise of American Sports Gambling."

www.npr.org

January 19, 2026 at 4:29 PM

Shafik Yaghmour

@shafik.bsky.social

"Extracting books from production language models": arxiv.org/abs/2601.02671

#ai

6 Conclusion
With a simple two-phase procedure (Section 3), we show that it is possible to extract large amounts of in-copyright text from four production LLMs. While we needed to jailbreak Claude 3.7 Sonnet and GPT-4.1 to facilitate extraction, Gemini 2.5 Pro and Grok 3 directly complied with text continuation requests. For Claude 3.7 Sonnet, we were able to extract four whole books near-verbatim, including two books under copyright in the U.S.: Harry Potter and the Sorcerer’s Stone and 1984 (Section 4).

January 17, 2026 at 3:58 AM

Shafik Yaghmour

@shafik.bsky.social

"C Is Not a Low-level Language
Your computer is not a fast PDP-11.": queue.acm.org/detail.cfm?i...

Not new but always worth a read.

#programming

C also requires padding at the end of a structure because it guarantees no padding in arrays. Padding is a particularly complex part of the C specification and interacts poorly with other parts of the language. For example, you must be able to compare two structs using a type-oblivious comparison (e.g., memcmp), so a copy of a struct must retain its padding. In some experimentation, a noticeable amount of total runtime on some workloads was found to be spent in copying padding (which is often awkwardly sized and aligned).

If an unspecified value for flow control is used (for example, using it as the condition in an if statement), then the result is undefined behavior: anything is allowed to happen. Consider the loop-unswitching optimization, this time in the case where the loop ends up being executed zero times. In the original version, the entire body of the loop is dead code. In the unswitched version, there is now a branch on the variable, which may be uninitialized. Some dead code has now been transformed into undefined behavior. This is just one of many optimizations that a close investigation of the C semantics shows to be unsound.

Since then, implementations of C have had to become increasingly complex to maintain the illusion that C maps easily to the underlying hardware and gives fast code. A 2015 survey of C programmers, compiler writers, and standards committee members raised several issues about the comprehensibility of C.3 For example, C permits an implementation to insert padding into structures (but not into arrays) to ensure that all fields have a useful alignment for the target. If you zero a structure and then set some of the fields, will the padding bits all be zero? According to the results of the survey, 36 percent were sure that they would be, and 29 percent didn't know. Depending on the compiler (and optimization level), it may or may not be.

This is a fairly trivial example, yet a significant proportion of programmers either believe the wrong thing or are not sure. When you introduce pointers, the semantics of C become a lot more confusing. The BCPL model was fairly simple: values are words. Each word is either some data or the address of some data. Memory is a flat array of storage cells indexed by address.

January 14, 2026 at 9:53 PM

Shafik Yaghmour

@shafik.bsky.social

C++ Retro quiz #52 No one expected that code

Answer here: hachyderm.io/@shafik/1158...

#CPolls
#cpppolls
#Programming

$Retro C++ quiz #52 Given the following C code #include <stdio.h> int main(void) { for (size_t i=0;i<6;++i) { typedef int VLA[i++]; printf("%zu",i); } } W/o checking what is the output? A. 012345 B. 135 C. ¯\_(ツ)_/¯ undefined behavior D. Varies on standard version$

January 14, 2026 at 4:09 PM

Shafik Yaghmour

@shafik.bsky.social

Two muffins where in an oven

One muffin says to other "I am getting baked in here!"

The second muffin goes "WOAH!!! a talking muffin!!!!"

🥁

January 14, 2026 at 4:07 PM

Shafik Yaghmour

@shafik.bsky.social

A magician said he would disappear on the count of three.

He wrapped himself in his cloak and counted, Uno, Dos

and then disappeared w/o a Tres

🥁

January 10, 2026 at 7:00 PM

Shafik Yaghmour

@shafik.bsky.social

C++ the language in which the edge cases have edge cases 🧐

#cplusplus

January 10, 2026 at 3:59 AM

Reposted by Shafik Yaghmour

Shafik Yaghmour

@shafik.bsky.social

Cursed Code of the day

Cursed Hello World in C: gcc.godbolt.org/z/GEsEGh34T

How many tools did it take to write this 🤔

#programming

Code:

const int main[] =
{440,113408,-1922629632,
4149,1423872,84869120,
15544,266023168,1920287493,
543450483,1819043144,
1867980911,543452274,663858};

Results:

Program returned: 1
Cursed Hello World 2!

January 8, 2026 at 6:47 PM

Shafik Yaghmour

@shafik.bsky.social

If C++ were an Orchestra, what would be the loudest instrument?

Surely the unresolved external cymbols 🥁

#cplusplus

January 8, 2026 at 6:53 PM

Shafik Yaghmour

@shafik.bsky.social

Cursed Code of the day

Cursed Hello World in C: gcc.godbolt.org/z/GEsEGh34T

How many tools did it take to write this 🤔

#programming

January 8, 2026 at 6:47 PM

Shafik Yaghmour

@shafik.bsky.social

Why should version numbers always be palindromes?
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
B/c then they will always be backward compatible 🥁

#programming

January 8, 2026 at 7:22 AM

Shafik Yaghmour

@shafik.bsky.social

Retro C++ quiz #51 Extra virtual?

Answer here: hachyderm.io/@shafik/1158...

#Cplusplus
#Cpppolls

Retro C++ quiz #51

Given the following in C++:

struct A {
virtual void f() = +0;
};

Without checking, is this:

A. Well-formed
B. Ill-formed (requires a diagnostic)
C. Depends on the standard version

January 7, 2026 at 4:54 PM

Shafik Yaghmour

@shafik.bsky.social

"Call for contributors for C++20 Modules on Windows": discourse.llvm.org/t/rfc-call-f...

#llvm

[RFC] Call for contributors for C++20 Modules on Windows

This comes from [C++20] [Modules] [Windows] Collections for C++20 Modules for windows in clang · Issue #174067 · llvm/llvm-project · GitHub This is a meta issue collects the issues or missing feature...

discourse.llvm.org

January 5, 2026 at 6:23 PM

Shafik Yaghmour

@shafik.bsky.social

"On the Calculation of Volume" (which is not about math, at least not directly)

Was a book I was interested in and turned out to be really excellent. It kind of has to set in but it is really a fascinating read.

#books

January 5, 2026 at 5:10 PM

Shafik Yaghmour

@shafik.bsky.social

Multi-Dimensional Analog Literals is my “favorite” abuse of C++ operator overloading

eelis.net/C++/analogli...

🤯🧐

*Chef kiss*

#Cplusplus #Programming

$// Note: The following is all standard-conforming C++, this is not a hypothetical language extension. #include "analogliterals.hpp" #include <cassert> int main () { using namespace analog_literals::symbols; // Consider: unsigned int a = 4; // Have you ever felt that integer literals like "4" don't convey the true size of the value they denote? If so, use an analog integer literal instead: unsigned int b = I---------I; assert( a == b ); // Due to the way C++ operators work, we must use N*2+1 dashes between the I's to get a value of N: assert( I-I == 0 ); assert( I---I == 1 ); assert( I-----I == 2 ); assert( I-------I == 3 ); // These one-dimensional analog literals are of type analog_literals::line<N>, which is convertible to unsigned int.$

// In some cases, two-dimensional analog literals are appropriate:

unsigned int c = ( o-----o
| !
! !
! !
o-----o ).area;

assert( c == (I-----I) * (I-------I) );

assert( ( o-----o
| !
! !
! !
! !
o-----o ).area == ( o---------o
| !
! !
o---------o ).area );
// Two-dimensional analog literals are of type analog_literals::rectangle<X, Y> which exposes static member constants width, height, and area.

/* As an example use-case, imagine specifying window dimensions in a GUI toolkit API using:

window.dimensions = o-----------o
| !
! !
! !
! !
o-----------o ;
Who said C++ was unintuitive!? */

January 4, 2026 at 7:50 PM

Reposted by Shafik Yaghmour

Shafik Yaghmour

@shafik.bsky.social

Cursed code of the day

godbolt.org/z/1z87qaW9r

I am very sorry 🧐

#programming

Godbolt screenshot

Code:

int main(){
switch(0){
struct a {
int i:({case 0:; 1;});
};
}
}

Resulting Assembly:

main:
push rbp
mov rbp, rsp
xor eax, eax
pop rbp
ret

January 3, 2026 at 11:29 PM

Shafik Yaghmour

@shafik.bsky.social

Cursed code of the day

godbolt.org/z/1z87qaW9r

I am very sorry 🧐

#programming

January 3, 2026 at 11:29 PM

Add to Home Screen

Light up
your news

Add to Home Screen

Light upyour news

Sign in to Lightnews

Sign up to start reading

Connect Bluesky

Connect with Bluesky

Light up
your news