GCC 12 Adds Support For New #elifdef #elifndef Directives

[phoronix]

Phoronix: GCC 12 Adds Support For New #elifdef #elifndef Directives

With C2X and potentially a future version of C++, there is finally the #elifdef and #elifndef directives. The GNU Compiler Collection is preparing its support...

https://www.phoronix.com/scan.php?page=news_item&px=GCC-12-elifdef-elifndef

[discordian]

Cnat wait to see code like this:

#ifdef HUGO

dothehugo();

#if __GNUC__ >= 12 || __cplusplus >= 20120000
#elifdef FOO
#else
#elif defined(FOO)
#endif

dothefoo();

#endif

Seriously, it should've been there from the start, but adding it now it ridiculous

[coder]

Cnat wait to see code like this:

LOL. That obviously won't work, of course.

Seriously, it should've been there from the start, but adding it now it ridiculous

I take a different view. I think #ifdef is too easily confused with #if. I'd rather they left it to just #if. It's a frequent point of confusion for neophyte C/C++ programmers, and even something you occasionally see more experienced types mess up.

#if defined( FOO )

isn't an unreasonable amount of typing.

[ReinUsesLisp]

It has to be added at some point for it to be widespread. We'll be able to use it on public headers in 20 years, like we can use `//` and stdint.h for the most part today.

[hoohoo]

Honestly it is time for the committee to recess. For five or ten years. Until there is something important to worry about.

[arQon]

Seriously, it should've been there from the start, but adding it now it ridiculous

Exactly so. Its absence has had literally no negative impact on the language for decades, and adding it is of absolutely no benefit at all at this stage.

But, meh - it's not actively HARMFUL either, just petty and stupid. If the committee is so idle and so bored that they need to chase make-work like this, then hopefully everything that DOES actually matter is in pretty good shape.

[carewolf]

Exactly so. Its absence has had literally no negative impact on the language for decades, and adding it is of absolutely no benefit at all at this stage.

But, meh - it's not actively HARMFUL either, just petty and stupid. If the committee is so idle and so bored that they need to chase make-work like this, then hopefully everything that DOES actually matter is in pretty good shape.

It can cause headers to become accidentily restricted to specific modern compilers, when there is no reason for that. There is a clear risk of harm.

[pal666]

I think #ifdef is too easily confused with #if. I'd rather they left it to just #if.

it's hard to leave it to just

#if defined

when

#ifdef

was invented before

#if defined

#if defined( FOO )


isn't an unreasonable amount of typing.

it contains 3 unnecessary characters(parens and space)

[pal666]

If the committee is so idle and so bored that they need to chase make-work like this, then hopefully everything that DOES actually matter is in pretty good shape.

all things which are both exciting and easy to do were done long ago. now exciting things are hard. subj is pretty easy

[pal666]

It can cause headers to become accidentily restricted to specific modern compilers

like every other new feature. so instead of crying you build your project with oldest supported compiler on ci. and btw this feature will not be even enabled without relevant -std= option. if you build your project with -std=c++23, then you clearly want to restrict to specific modern compilers

, when there is no reason for that.

if there's no reason you wouldn't use feature

There is a clear risk of harm.

only for people who have no chances anyway