austin mccartney

"The proper use of comments is to compensate for our failure to express ourself in code" - Robert C. Martin

[austin mccartney @boolmoos]

@PhilosophyTube The vibe:
https://t.co/NpLCetx8g4

[Tim McNamara @timClicks]

You know it's a going to be a hard day as a consultant when you see JSON embedded in a database's string field and the customer wants to know what the best indexing strategy should be.

[austin mccartney @boolmoos]

@CplusplusVille Ada had such a strange origin. Imagine a department of the US government trying to mandate what language software is written in...

[austin mccartney @boolmoos]

@incomputable > I'm sorry, I can't assist with creating images of specific, real-life people such as Scott Meyers based on their photographs.

<Proceeds to produce Leslie Lamport>

[austin mccartney @boolmoos]

@daddyhosen @ericniebler The Rule of Chiel is a good rule of thumb for the relative cost of compile-time operations.

https://t.co/Qm0jd7loAR

Keep that in mind and design to make the most of memoized instantiations.

[austin mccartney @boolmoos]

@BMenkuec @MaxClerkwell > I only use c++ templates occasionally and only in simple ways. I have to look up the syntax everytime.

> regarding cout, I used it but were never a fan of it, it's worse than printf imho. fmt or std::format is much cooler.

@vzverovich
https://t.co/8c105RZoQa

[austin mccartney @boolmoos]

@ericniebler Probably rhetorical but...
What would the callsite for `g` look like?

What if the argument name did not refer unambiguously to the function template of interest (e.g. overloads)?

What if multiple concept-constrained function templates with otherwise matching signatures exist?

[austin mccartney @boolmoos]

@jurshsmith @nice_byte I think it's better to argue that Rust's design makes engineering trade-offs that, at scale, yield more robust software with little or no performance overhead.

[austin mccartney @boolmoos]

@jurshsmith @nice_byte To argue that Rust's design doesn't sometimes complicate the implementation otherwise valid constructs seems ignorant or disingenuous.

[austin mccartney @boolmoos]

@ericniebler That would be somewhat surprising. I would have expected the dispatching for that customization to be deferred until the construction of the operation state. Does this dispatching potentially impact downstream decision-making (thereby necessitating more eager evaluation)?

[austin mccartney @boolmoos]

@ericniebler The signature of bulk is what I anticipated, but from the motivating discussion, I understand that the scheduler passed is not retrieved from the downstream receiver but instead from the composed sender. Is that correct?

[austin mccartney @boolmoos]

@RichFelker @pati_gallardo @ColossalCaveTBG Respectively -> Respectfully* 😑

[austin mccartney @boolmoos]

@RichFelker @pati_gallardo @ColossalCaveTBG > some guru grifter type I've never heard of
> someone like that with nothing of value to contribute

Respectively, your lack of familiarity with Authur and his contributions to the C++ ecosystem is a blind spot.

[austin mccartney @boolmoos]

@RichFelker @pati_gallardo @ColossalCaveTBG I believe a component of the hesitance to exclude him from these organizations and events, for a subset of the community, stems from the quality and quantity of his previous contributions.

Whether those factors have a place in this sort of consideration is another matter

[austin mccartney @boolmoos]

@berglerma @pati_gallardo @Cor3ntin Copying files between repositories feels... not great (especially if it's someone else's content, given copyright and license notice distribution obligations) and the upsides don't strike me as particularly compelling.

But that's just, like, my opinion. You do you.

[austin mccartney @boolmoos]

@berglerma @pati_gallardo @Cor3ntin > Yeah but then you have to manually add each individual test case with add_test yourself.

It's true

As I said earlier, grabbing a dependency for a project's build system to avoid calling add_test directly seems heavy to me, but tastes vary.

[austin mccartney @boolmoos]

@berglerma @pati_gallardo @Cor3ntin That's available with vanilla CTest.

A given executable may be passed as part of the COMMAND argument of many invocations of add_test. Use commandline arguments of that executable to select a subset of tests appropriate for each add_test call.

1 add_executable != 1 add_test

[austin mccartney @boolmoos]

@pati_gallardo @Cor3ntin From the documentation for CLion, I don't think that's a problem. If I'm reading the docs correctly, the tests found by CLion are populated by interpretting the CTestTest files generated by CMake.
I expect calling add_test for each test (not test executable) will accomplish that

[austin mccartney @boolmoos]

@pati_gallardo @Cor3ntin Grabbing a dependency for a project's build system to avoid calling add_test directly seems heavy to me, but tastes vary.

Is the project in question open source or free software? Easy enough to demonstrate.