octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: gnulib warnings when compiling Faddeeva.cc


From: Jordi Gutiérrez Hermoso
Subject: Re: gnulib warnings when compiling Faddeeva.cc
Date: Sat, 15 Dec 2012 11:43:54 -0500

On 15 December 2012 08:29, Steven G. Johnson <address@hidden> wrote:
> Jordi Gutiérrez Hermoso wrote:
>>
>> On 14 December 2012 14:13, Steven G. Johnson <address@hidden> wrote:
>>
>>> b) use <cmath> and GNULIB_NAMESPACE::floor; this may be nonportable since
>>> it
>>> assumes that the system <cmath> header includes math.h and is compatible
>>> with the gnulib math.h
>>
>>
>> This isn't any more nonportable than math.h unless you have a perhaps
>> justifiable C++ phobia; cmath is supposed to be exactly the same as
>> math.h and calls the same functions;
>
>
> You are missing the point: gnulib provides a replacement for math.h, not for
> cmath.   It is not cmath that is potentially nonportable, it is the
> assumption that (1) the system cmath includes math.h and (2) the system
> cmath is compatible with gnulib's math.h.

Oh, I see. My mistake.

>> On an unrelated note, I was thinking, I want to stomp all over the
>> style in your Faddeeva file, such as for example fixing the whitespace
>> (no tabs!) and getting rid of the needless C macros. Since this is our
>> function now and you are tracking our hg repo so you can easily merge
>> back, this is ok, right?
>
> I'm not going to merge those changes upstream.   When I make a change
> upstream, it is not "easy" to merge the change downstream if you have
> changed all the whitespace etc, unless you have some version of diff and
> patch that I don't.

Well, this is why we have a DVCS, right? We can easily maintain two
branches of this file. Probably not worth the effort for such minor
cosmetic differences, though.

Can we at least get rid of the tabs? Mixing tabs and spaces is a bit
ugly. I don't like having to guess which tab width makes the code
align properly.

- Jordi G. H.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]