[boost] Re: Formal review of "Output Formatters" library begins today

From: Vladimir Prus (ghost_at_[hidden])
Date: 2004-09-20 03:38:03

Roland Richter wrote:

>> I've made a noise about this in the past. I strongly believe that current
>> lexical_cast behaviour does not play nice with existing stream operators
>> (which are commonly written to assume whitespace is skipped). There's no
>> good reason for that, and lexical_cast should not use the "noskipws"
>> flag.
>> 
> 
>   Do you happen to know why it was introduced then in the first place?

Not really. IIRC, KevlinšHenney said he finds it is more clear when
lexical_cast is required to consume the entire input, without skipping
whitespace at the beginning. I don't understand this argument at all, and
the new behaviour only creates practical issues :-(

- Volodya