Unfortunately issues with floating-point overloads of std :: to_string have notīeen discussed in the paper that proposed this facility (, ). In fact almost half of floating-point numbers are formatted as zero and there isĪ precision loss or garbage digit output in many other cases.
For example: std :: cout << std :: to_string ( -1e-7 ) And finally we have 6 meaningless zeros after aįormatting of small floating-point numbers is even less useful. Here only the first 17 digits are meaningful, the next 292 are so-called imbue ( loc ) setlocale ( LC_ALL, "C" ) std :: cout :: max ())
SFĬonsider the following example: auto loc = std :: locale ( "uk_UA.UTF-8" ) std :: locale :: global ( loc ) std :: cout. Poll: Fix to_string as proposed by P2587R1 ( to - string or not to_string). Īdded a reference to which proposed deprecatingįloating-point overloads of std :: to_string.Mentioned that the current proposal enables future extension to other typesĪdded an analysis of codesearch results to § 6 Impact on existing code.Īpplied the same changes to to_wstring in the wording. This paper proposes fixing these issues while retainingĮxisting semantics of integral overloads. The choice of the floating-point format makes std :: to_string of very limited Fortunately for integral and unfortunately forįloating-point overloads they are all defined in terms of sprintf inconsistently with C++ formatted output functions ( ). "Though this be madness, yet there is method in ’t."Ĭ++11 introduced a set of std :: to_string overloads for integral andįloating-point types. Appendix A: Sample of code search results for to_string.
0 kommentar(er)
