[gmx-developers] Re: gmx-developers Digest, Vol 96, Issue 16

[Szilárd Páll]

On Wed, Apr 11, 2012 at 3:16 PM, Peter Kasson <kasson at stanford.edu> wrote:
>> Date: Wed, 11 Apr 2012 14:33:09 +0200
>> From: Erik Marklund <erikm at xray.bmc.uu.se>
>> Subject: Re: [gmx-developers] Coordinate scaling in pdbio.c
> [...]
>>>>>>> What happened to A2NM and NM2A?
>>>>>>> In my (4.0.5, yes that is very old) includes/physics.h I still see:
>>>>>>>
>>>>>>> #define A2NM (ANGSTROM/NANO) /* NANO */
>>>>>>> #define NM2A (NANO/ANGSTROM) /* 10.0 */
>>>>>>>
>>>>>>> Why aren't these still used for that?
>>>>>>>
>>>>>>>
>>>>>> At some stage macros were removed from the code, even though these seem rather harmless. I agree that it is confusing to hard code these numbers. If we cannot use macro's like this we should probably replace them by
>>>>>> static const real ANGSTROM=1e-10;
>>>>>> static const real NANO=1e-9;
>>>>>>
>>>>>> etc.
>>>>>>
>>>>>> Comments?
>>>>>
>>>>> include/physics.h still has these macros. IMO, hard-coded constants are a greater evil than macros to prevent that, though I expect we will transition to const values at some stage soon.
>>>>
>>>> But are these the kind of macros we want to avoid? Aren't function-like macros the ones to kill in the first place?
>>>  They're taking no parameters, so they're hardly function-like. The code fragments above are compiled into constants by the pre-processor. A macro that is used as
>>>
>>> dist_in_nm = dist_in_angstrom * A2NM;
>>>
>>> is much less evil than a macro
>>>
>>> dist_in_nm = A2NM(dist_in_angstrom);
>>
>> My point exactly. I'm just not seeing the harm in using macros in cases like these.
>
> Consts are preferred (not #defines but actual consts).
>
> http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Preprocessor_Macros

That's C++ and we're not there yet. For 4.6 IMO it makes sense to
clarify code by replacing "x/10.0" with "x/A2NM". In fact, this will
help with the transition to constants later.

--
Szilárd


> --
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.