[Rd] sprintf, check number of parameters

[Tomas Kalibera]

On 2/22/21 11:34 AM, Duncan Murdoch wrote:
> This is ugly, but I think it's legal, and it doesn't trigger a 
> warning:  output unused parameters as zero-length strings:
>
>  msnx(T0, mask = '%1$.1f (SD=%2$.1f)%3$.0s%4$.0s')
>
> Perhaps an example using %.0s could be included to show how to skip a 
> value.

Thanks, I've added a sentence to that effect.

Best
Tomas

>
> Duncan Murdoch
>
> On 22/02/2021 5:06 a.m., Tomas Kalibera wrote:
>> Dear Matthias,
>>
>> On 2/6/21 2:11 PM, Matthias Gondan wrote:
>>> Dear developers,
>>>
>>> This is a follow-up from an earlier mail about warnings of unused 
>>> arguments in sprintf:
>>>
>>> 1. This should obviously raise an error (and it does):
>>> sprintf('%i %i', 1)
>>> Fehler in sprintf("%i %i", 1) : zu wenig Argumente [= too few 
>>> arguments]
>>>
>>> 2. This should, in my opinion, raise a warning about an unused 
>>> argument (and I think it does in now R-devel):
>>> sprintf('%i', 1, 2)
>> yes, it does.
>>> 3. From the conversation below, it seems that this also raises a 
>>> warning (in R-devel):
>>> sprintf('%1$i', 1, 2)
>> yes, it does as well
>>> I think that one should be suppressed. When I reported this a few 
>>> months ago, I didn’t really have a use case for (3), but now I think 
>>> I have found something. Suppose I have a function that calculates 
>>> some descriptive statistics, mean, sd, available cases, missings, 
>>> something like the one below:
>>>
>>> msnx = function(x, mask='%1$.1f (SD=%2$.1f, n=%3$i, NA=%4$i)')
>>> {
>>>     m = mean(x, na.rm=TRUE)
>>>     s = sd(x, na.rm=TRUE)
>>>     n = sum(!is.na(x))
>>>     na = sum(is.na(x))
>>>         sprintf(mask, m, s, n, na)
>>> }
>>>
>>> The mask is meant to help formatting it a bit.
>>>
>>> msnx(T0)
>>> [1] "30.7 (SD=4.7, n=104, NA=0)"
>>>
>>> Now I want a „less detailed“ summary, so I invoke the function with 
>>> something like
>>>
>>> msnx(T0, mask='%1$.1f (SD=%2$.1f)')
>>> [1] "30.7 (SD=4.7)"
>>>
>>> In my opinion, in the last example, sprintf should not raise the 
>>> warning in (2) if all arguments in the mask are „dollared“. I am 
>>> still a bit unsure since the example uses a function that calculate 
>>> things that aren’t being used (n and na), and this could be 
>>> considered bad programming style. But there might be other use 
>>> cases, and it is, nevertheless, a deliberate choice to skip 
>>> arguments 3$ and 4$.
>>
>> Thanks for the example. I am sympathetic with your concerns about the
>> programming style in it: the caller needs to know exactly how "mask"
>> will be used, that it would be in a call to sprintf() and what would be
>> the indices of the arguments.
>>
>> The warning has been introduced a while ago and there has not been any
>> report yet that it would break existing good style code (particularly
>> CRAN packages have been tested extensively), which indicates that
>> currently the R code base does not rely on unused $- arguments.
>>
>> It is hence I think wise to keep the warning to prevent R code base from
>> relying on that in the future, because gcc/clang already warn on unused
>> $-arguments. Not only that gcc developers must have been thinking hard
>> about the same thing before us getting to this conclusion: $- arguments
>> are a POSIX extension and gcc/clang are the key compilers for POSIX
>> systems, so it is safer to abide by their rules. In principle POSIX may
>> mandate that $- arguments are used explicitly in the future (now it is
>> rather vague, it seems unused are fine only when last), and even if not,
>> deviations from gcc/clang could cause confusion for applications and
>> developers using both C/C++ and R.
>>
>> Best
>> Tomas
>>
>>
>>
>>>
>>> Best wishes,
>>>
>>> Matthias
>>>
>>>
>>>
>>> Dear Matthias,
>>>
>>> thanks for the suggestion, R-devel now warns on unused arguments by
>>> format (both numbered and un-numbered). It seems that the new 
>>> warning is
>>> useful, often it finds cases when arguments were accidentally passed to
>>> sprintf but had been meant for a different function.
>>>
>>> R allows combining both numbered and un-numbered references in a single
>>> format, even though it may be better to avoid and POSIX does not allow
>>> that.
>>>
>>> Best
>>> Tomas
>>>
>>> On 9/20/20 1:03 PM, Matthias Gondan wrote:
>>>> Dear R developers,
>>>>
>>>> I am wondering if this should raise an error or a warning.
>>>>
>>>>> sprintf('%.f, %.f', 1, 2, 3)
>>>> [1] "1, 2"
>>>>
>>>> I am aware that R has „numbered“ sprintf arguments 
>>>> (sprintf('%1$.f', …), and in that case, omissing of specific 
>>>> arguments may be intended. But in the usual syntax, omission of an 
>>>> argument is probably a mistake.
>>>>
>>>> Thank you for your consideration.
>>>>
>>>> Best wishes,
>>>>
>>>> Matthias
>>>     [[alternative HTML version deleted]]
>>>
>>> ______________________________________________
>>> R-devel using r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>> ______________________________________________
>> R-devel using r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>