Boost logo

Boost :

Subject: Re: [boost] [locale] review
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2011-04-17 18:29:27


AMDG

On 04/17/2011 12:25 PM, Phil Endecott wrote:
> I don't have time to answer in much detail, but I'll just re-emphasise
> that I would very much like to see a simple and fast character set
> conversion like this:
>
> string l1_name = charset_conv<utf8,latin1>(utf8_name);
>
> In my own charset conversion code, I generated a large enum of all the
> character set names and aliases based on the IANA registry
> (http://www.iana.org/assignments/character-sets). I then have a
> fallback converter using iconv, and fast specialisations for cases like
> utf-to/from-iso8859. I would happily contribute those specialisations
> if you had a way to plug them in. Currently you offer either conversion
> functions that take character set names as strings, which I dislike
> because of the overhead of the run-time lookup that is required
> (especially for short strings), and stream processing, which I dislike
> because I don't want the overhead (runtime and lines-of-code) of copying
> to and from stringstreams.
>

Of course this only works if the character set is
known at compile time. Boost.Locale is all about
runtime configuration. Thus, specifying the character
set as a string rather than a template parameter
is perfectly reasonable.

In Christ,
Steven Watanabe




Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk