develooper Front page | perl.perl5.porters | Postings from April 2019

Re: Config options d_strtod_l & d_strtold_l

Thread Previous
Karl Williamson
April 25, 2019 20:56
Re: Config options d_strtod_l & d_strtold_l
Message ID:
On 4/25/19 2:50 AM, H.Merijn Brand wrote:
> On Thu, 25 Apr 2019 14:28:01 +1000, sisyphus <>
> wrote:
>> Hi,
>> Previously unaware of these options, I learned of their existence
>> while reading
>> But I can't find any C code in the perl source that actually does
>> things differently depending upon whether these symbols are defined
>> or not.
>> I mean, I can see that defining them leads to HAS_STROD_L &
> Missing a T there :)

I make that mistake all the time too.
>> HAS_STRTOLD_L being defined, but that's it.
> It was added back in in April 2018 because Karl asked for it.
> It is a kind of chicken-egg problem: one cannot use feature until they
> are there. So we (the Configure maint team) add features on request,
> but we do not control its use.
>> AFAICS, there's in no place in the perl source that makes purposeful
>> use of these symbols; hence whether they are defined or not really
>> makes no difference at all..
> That is a correct conclusion, but now that the probes are in, you *can*
> use them.
>> Have I missed something ?
> I don't think so, except maybe the way this process works, but that is
> not important

I did ask for them, and it's been in the back of my mind to add code to 
use them, but I didn't get around to it in this release.  On systems 
that have them, we don't have to go through gyrations to get the locale 
correct before calling them.  We just call them with the locale we want, 
and libc takes care of it.

In my response on
I forgot that the '_l' versions weren't used, so undeffing them should 
have made no difference, but it's good to do now so that when they do 
get used we won't have had to remember to do this.
>> Rob

Thread Previous Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About