On 01/26/2013 09:03 PM, Dmitry Bogatov wrote:
Hello! I found in seq.c:152 following line:if (! arg[strcspn (arg, "xX")] && isfinite (ret.value)) where arg is checked to be valid double (check for `xstrtold' success) just before. I checked for `git blame`, it refers to 86e4b778, and commit message talk about scentific notation. Can you point me to documention about "xX" symbols in double? I found only about "1e-6" -- like style.
That line is basically avoiding the equal width logic in the presence of "hex float constants". Before 86e4b778, numbers in scientific notation also bypassed the equal width logic, but 86e4b778 removed that limitation (and 64d4a280 recently improved this support) Hex consts are complicated and have the form: 0xA.BpC where A=hex, B=hex/16 and C = 2^-dec, for example... $ seq 0x10.8p0 18 16.5 17.5 thanks, Pádraig.
