Reinderien

When constexpr is sent to the compiler, it can make a whole family of guarantees about evaluation simplicity, and optimisation benefits.

Keep researching. If you care about inline, you should care much more about constexpr.

@bruno re. constexpr - I think your interpretation is that it only matters for quantities that are compile-time-constant, and that interpretation is wrong.

true about the runtime params.

cheers!

You're welcome. And I disagree with the age take. I think we as professionals need to operate in an atmosphere of hope: change for the better is possible.

"Two lines when one is enough" without being coloured by other design virtues can lead to some really, really bad code.

You don't need to change, and gdb is perfectly fine. Users of gdb will thank you if you sanitise your mutations.

(a) we don't live in that world anymore, and (b) a situation like that is actually much more justification to emphasize a simple execution model.

Great. I don't think that length of a career is necessarily the best metric for people being effectual; and also, this is less about tools and more about humans.

Depending on your definition, let's say 1995

The decrement is a trivial example, but there are more severe examples where being able to easily separate a mutation and the evaluation of its effect are very, very important.

Then you haven't debugged enough.

if (!n1)

--n1;

And... don't do it. Instead write:

That's a bad thing to force on a programmer.

When you step over the line I showed above, it will be executed all at once.

Whatever, it doesn't matter. Now we're talking about you running gdb in a shell. That's enough.

if (! --n1)

When you do a step-over command in GDB, a line like this:

Great.

OK great. So.

Can you do a step-by-step debug of the program?

Well that's... bad.

OK. Does it support GDB sessions?

Are you running the program in an IDE?

Just like, why it should be done?

Which part of the mutator predicate comment?

@bruno constexpr has almost nothing to do with constant values. I strongly encourage you do to some reading on e.g. stackoverflow.com/questions/14116003/…

No, that's not what I mean; (a) I meant fixing the indexDiag functions and then do the same for isUsable; (b) there being no perceptible difference in execution time doesn't particularly matter; it's still the right thing to do to tell the compiler what's possible and unblock other optimisations upstream.

Make indexDiag1 and 2 constexpr and then try again.

I doubt I'll ever beat this record from my work account

That would have bothered me all day

Yes. Thank you

I'm dying here. What's the name for a class of question where it doesn't address the root of the question, is superficial, and tends to produce answers that don't materially improve the situation? aka. "what are you actually asking"

it's true

ty!

yeesh. can you please link the meta?

My longest-standing gripe with SE is that downvoting is silent and anonymous. I think if it required an accompanying comment the incentive structures could improve.

and (b) yes, that is weird. My mental threshold for down-voting sits somewhere between "abusive content" and "spam/AI slop"; and if the poster is acting in good faith there's basically never a case where I think a downvote is a useful tool.

(a) me speaking naively: do mods not have tools to see vote attribution?

> I've noticed a lot of posts seem to get bizarre -1s. Lots of posts will enter the 2nd with a random -1 which doesn't make much sense. I would ask a CM... but the team seem to take 6-8 weeks to respond.

Re. @Peilonrayz 's

lol. if you aren't seeing the same behaviour, then I will have to blame my family's router

:(

NS_ERROR_UNKNOWN_HOST

for e.g. gravatar.com/avatar/… I get

nor ctrl+F5