Re: [OT] Re: OT: Gawk match() and numbers in scientific notation

On Saturday 10 May 2008 10:58, Hermann Peifer wrote:

>  From a semantics point of view, it is also surprising that, when using
> de_DE.UTF-8 locale, e.g. <LATIN_SMALL_LETTER_A_WITH_ACUTE>  is included
> in the range [A-C], but <LATIN_SMALL_LETTER_C_WITH_ACUTE> is not, as it
> sorts after uppercase C.

Well, considered the general strangeness of the thing, from a certain
point
of view this makes sense to me. If [A-C] expands to AaBbCc, the "c" is not
matched because it comes after the "C" (which is the last character in the
bracket expression). On the other hand, if [A-C] expands to aAbBcC,
then "a" is not matched, again because it's outside the range.

> In summary: using character ranges in combination with a non-POSIX
> locale is a good recipe for surprising results. To a somewhat lesser
> degree, this is also true for character classes.

Agreed. That's why I'm trying to find a way to print character classes and
collating sequences, so that when someone comes asking "why doesn't sort
work correctly" or "why doesn't grep/awk/sed. etc. match what I expect",
they can be instructed to run the (non-existent yet) command or script to
see for themselves what the programs' idea of what they waht to do is.

-- 
All the commands are tested with bash and GNU tools, so they may use
nonstandard features. I try to mention when something is nonstandard (if
I'm aware of that), but I may miss something. Corrections are welcome.