Am 22.09.2016 um 06:01 schrieb Thorsten Froehlich:
> clipka <ano### [at] anonymousorg> wrote:
>> Am 21.09.2016 um 23:08 schrieb clipka:
>>
>>> Try `+a0.3` instead of `+a.3`.
>>>
>>> I'm not sure if POV-Ray /should/ support dropping the leading zero in
>>> command-line options (there's no clear statement about this in the
>>> docs), but apparently it currently doesn't.
>>>
>>> I'll have a look at POV-Ray's bowels to see if there are any hints about
>>> the intended behaviour in this case.
>>
>> The source code didn't give me much of a hint -- but running POV-Ray
>> 3.62 did: That version was fine with `+a.3`. So there's a bug that needs
>> fixing.
> 
> Wait: 3.6 and before had problems with that syntax because it can lead to cases
> that cannot properly be decided when parsing the command line. You will have to
> go back like 15 years to find the post about it ... and I am sorry, I do not
> have that post at hand any more :-(

Let me guess: You don't remember yourself what the problem was about,
and nobody deemed it worth documenting somewhere within the code, or any
other place easily accessible, right?

Fine then. If anybody ever stumbes across the problem again, I'll worry
about it.

From the fact that various other switches have been perfectly fine with
this notation all the time (`+j`, `+ec`, `+er`, `kf`, `+ki`, `+k`, `+sc`
and `+sr`, to name only those already present in 3.62, and to not
mention the `+mv` switch that wouldn't make much sense), my current
presumption is that it was just a historic artifact, that may have led
to conflicts due to the way command line switches had been processed
internally, and that there are no fundamental ambiguities that cannot be
overcome by a smarter implementation.


As a side note, it turns out that POV-Ray 3.62 for Windows didn't
support `+a.3` either; it just simply did not report any error if it
came across an unrecognized command line option.

But since POV-Ray supports dropping leading zeros in the SDL, and also
in many switches, I think it doesn't make sense to /specifically/ deny
this functionality in the `+a` switch.

(It also doesn't make sense to deny it in the `+ef` and `+sf` switches,
but that needs to be addressed separately as the internal handling is a
bit more complicated in these special cases.)

Post a reply to this message