On 02/11/2014 02:23 AM, Edward Diener wrote:
On 2/9/2014 10:58 PM, Vladimir Batov wrote:
... I would strongly suggest updating the docs for the library. Also either add in to the new docs the differences between the original version and the new version, or just include both the old and new version docs so that end-users could easily see what has changed.
Indeed. Docs are must. I started re-writing over the weekend but then thought I'd check with the community first -- maybe there was already a solution that I missed or there was no interest to begin with.
I would be glad to be the review manager of the new library, if necessary, if you wanted the new library added to the Boost review queue once you felt everything was complete. OTOH I would also understand if you wanted someone else to be the review manager.
Thank you for your offer of managing the proposal (if we come to that). It's much and truly appreciated. I was very happy with the way you managed the first proposal and your thorough analysis. I can't fathom why I might want someone else.
...
std::locale rus_locale (rus_locale_name); std::locale eng_locale ("");
// Set locale, case, precision, ccnv(std::setprecision(4))(std::scientific)(std::nouppercase);
string double_rus = convert<string>::from(double_v01, ccnv(rus_locale)).value(); string double_eng = convert<string>::from(double_v01, ccnv(eng_locale)).value();
Offhand this looks better. As I recall a general criticism was that you were trying to do too much in the convert library rather than just focusing in on 'convert' as a better, more flexible replacement than lexical_cast. Sometimes les is more <g>.
"Too much and too messy" was indeed my ultimate feeling back then. Now the split seems to make quite a difference; looks dead simple and more configurability.