Maxim Yegorushkin wrote:
In the real world practicality beats purity. The fact is people do not use boost::lexical_cast because of the aforementioned problems.
I second this! boost::lexical_cast in its current form/implementation is usable in hardly any real project situation because of many downsides. The only thing what is nice about this utility at the moment actually is its pretty interface, which wouldn't be so pretty if it handled all the requirements if should IMO. Those issues araised again and again on the boost list, but nobody felt ever responsible even to comment them in a sufficient way, some people always referring to the 'nice interface' which should never be touched in any way. I don't remember all the issues people complained about in many threads, I remember some of them which are really crucial and need to be dealt with: - performance, performance, performance - error handling - controlling the conversions via facets There may be some others and maybe we should put up a complete list. I'm afraid, if lexical_cast is standardized as is, people will always have to go back to C library functions to do their conversions. I really can't imagine that we are the only ones who have major ressentiments agains lexical_cast in its current form! Stefan