Hi, 2013/7/19 Peter Lind <peter.e.l...@gmail.com>
> On 19 July 2013 11:18, Dan Cryer <d...@dancryer.com> wrote: > >> What's the intended use case for string increment / decrement? >> >> > Personally, I instantly think of mirroring spreadsheet columns - works > quite well in that context. > ++/-- 'XYZ00001234' would have use cases. > > >> It just seems like madness to me, using mathematical operators with >> strings, producing seemingly arbitrary results in some circumstances (C -> >> B -> A -> NULL / False ?). >> >> > Throw a warning and don't decrement A/a any further - that's not arbitrary. > -- is more problematic. --'XYZ0000' would be 'XYZ0000' or 'XYY00000' or 'XYY'? It would be better not to change length of string, IMO. e.g. --'XYZ0000' became 'XYY0000' All chars would stop decrement at lowest chars of [0-9], [a-z], [A-Z]. e.g. Lowest value of 'XYZ0000' is 'AAA0000'. This is not a symmetric operation of ++, but it's impossible to achieve symmetric operation ++/-- on strings anyway. > >> Also what happens in other languages? Take for example German, in which ß >> comes after S, Ü after U, and so on. >> >> > Nothing, works purely on ascii, as is currently the case. Perl handles > other character sets though - see the link I posted for details on how. > There may be need for this, but it will be very complex code than now. [0-9a-zA-Z] would be enough. PHP doesn't have default multibyte char module, too. Regards, -- Yasuo Ohgaki yohg...@ohgaki.net
