Path: ...!weretis.net!feeder9.news.weretis.net!news.quux.org!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: David Brown <david.brown@hesbynett.no>
Newsgroups: comp.lang.c
Subject: Re: Suggested method for returning a string from a C program?
Date: Mon, 24 Mar 2025 13:09:00 +0100
Organization: A noiseless patient Spider
Lines: 63
Message-ID: <vrri0s$qila$2@dont-email.me>
References: <vrd77d$3nvtf$2@dont-email.me>
 <87a59hvgyk.fsf@nosuchdomain.example.com> <vrdi0g$47cb$3@dont-email.me>
 <vrdifm$4n2o$1@dont-email.me> <vrdj4b$47cb$4@dont-email.me>
 <vrdkcb$4n2o$2@dont-email.me> <86ecyronqa.fsf@linuxsc.com>
 <vrh548$3a4q5$1@dont-email.me> <86a59fo910.fsf@linuxsc.com>
 <865xk0ljcb.fsf@linuxsc.com> <vrogin$1ut0v$2@dont-email.me>
 <861punleus.fsf@linuxsc.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Mon, 24 Mar 2025 13:09:00 +0100 (CET)
Injection-Info: dont-email.me; posting-host="aff7e84aa47b82b9c39de36ffb8a5c58";
	logging-data="871082"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19cQESltKPmCPXdQyv6JhIsur/ROeOCOo4="
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.11.0
Cancel-Lock: sha1:Q8BuW18EjuH3uvpTvtyxq6HTPTA=
In-Reply-To: <861punleus.fsf@linuxsc.com>
Content-Language: en-GB
Bytes: 4667

On 23/03/2025 20:35, Tim Rentsch wrote:
> Richard Heathfield <rjh@cpax.org.uk> writes:
> 
>> On 22/03/2025 23:46, Tim Rentsch wrote:
>>
>>> Here is my summary of the corresponding list in the C11 standard
>>> (descriptive headings reprepresent my own views on each area):
>>
>> Thank you, Tim.  What I'm taking away from this is that I'm not
>> personally affected by the changes.
> 
> I had a similar reaction when C99 came out.  In fact it was several
> years before I started looking at C99 as a suitable alternative to
> using just C90.  Gradually I began to see that the new features of
> C99 offered substantial advantages over what C90 offers.  My
> impression is the C community at large followed a similar timeline.
> 
> For C11 I am somewhere in the middle.  Most of the time I find C99
> adequate, and don't need the capabilities added in C11;  so C99 is
> still my default choice, with C11 being the exception.  However,
> for building software libraries rather than just programs, C11
> allows some significant benefits, so I am turning more and more
> to C11 as I add functionality to the library projects I'm working
> on.  (I should add that I try to write code that can benefit from
> C11 but still can be called from C99 or C90, perhaps with reduced
> functionality.)
> 

I can somewhat agree with those sentiments.

In my field, C99 support was not an option for many compilers for 
several years - some toolchains have never had good support for it.  (I 
remember seeing a toolchain in the 2010's advertised as "now with 
partial C99 support" as a feature.)  So it took time before I could use 
it properly.

However, once I got used to it, the difference to the way I code was 
significant.  I greatly dislike the occasions when I have to go back to 
the middle ages coding in C90 - though not as much as the dark ages of 
large-scale assembly programming.  I have not the slightest doubt that 
with C99, my code is clearer, better structured, has a lower risk of 
errors and is easier to maintain than with C90.  I won't claim it is 
/hugely/ better, but it is definitely significant.  C99 is when C grew up.

The changes in C11 are much less, but there are some that are useful to 
me.  C17 is just a bug-fix (and slightly nicer typography).  C23 has a 
number of useful features, but again is not like the step from C90 to 
C99.  (And of course newer standards always have features that I dislike 
too.)

One thing I generally like with newer standards is that it can mean 
fewer implementation-specific features in my code, or fewer ugly 
workarounds.  I expect to see at least some of my gcc __attribute__'s 
replaced by [[C23 attributes]].  C11 came with a real _Static_assert, 
instead of having to use ugly macros for the task.  These sorts of 
things don't make a big impact in my code (and are often tidied away in 
macros anyway), but are nice to see.

I don't expect to be relying on C23 features for a while yet, however - 
I expect most of my new C code to be compatible with C99 (or at least, 
"gnu99").  And of course I won't change existing projects to newer 
standards.