I have WA 7 and I don't understand what the problem is.
My tests:
// check(<input string #1>, <input string #2> as integer, <output string>)

      // tests from acm.timus
      check("10.23", 0, "10")
      check(".04", 1, "0.0")
      check("-0.051e0", 1, "0.0")
      check("1.1e30", 10, "1100000000000000000000000000000.0000000000")
      check("-1.1e-30", 1, "0.0")
      check("2468097632.1358642324268913e-2", 20, "24680976.32135864232426891300")
      check("e23", 3, "Not a floating point number")
      check("1 e3", 1, "Not a floating point number")

      // my tests
      check(" 1 e3", 1, "Not a floating point number")
      check("1e3", 1, "1000.0")
      // http://acm.timus.ru/forum/thread.aspx?id=26503&upd=634425663233423296
      // Posted by Igor Dex 1 Jun 2011 18:28
      // It is clear that we can't have leading spaces in case correct real number. But is it possible to have spaces after correct number till the end of the string?
      // Posted by AterLux 1 Jun 2011 23:05
      // No. Trailing as leading spaces not satisfy the syntactical description.
      check("   ", 1, "Not a floating point number")
      check(" ", 1, "Not a floating point number")
      check("-", 1, "Not a floating point number")
      check(" 1e3", 1, "Not a floating point number")
      check(" .1e3", 1, "Not a floating point number")
      check(".1e-3", 1, "0.0")
      check(" .1e-3 ", 1, "Not a floating point number")
      check(".1e-3 ", 1, "Not a floating point number")
      check(".0e-10000000000000000000", 1, "0.0")
      check(".0e+10000000000000000000", 1, "0.0")

      // int and uint tests
      check("10", 0, "10")
      check("10", 1, "10.0")
      check("-10", 1, "-10.0")
      check("+10", 1, "10.0")
      // simplefloat and usimplefloat tests
      check(".10", 1, "0.1")
      check("10.10", 1, "10.1")
      check("10.01", 1, "10.0")
      check("+.10", 1, "0.1")
      check("+10.10", 1, "10.1")
      check("+10.01", 1, "10.0")
      check("-.10", 1, "-0.1")
      check("-10.10", 1, "-10.1")
      check("-10.01", 1, "-10.0")
      // float tests
      check("10e0", 0, "10")
      check("10e-1", 1, "1.0")
      check("-10e+1", 1, "-100.0")
      check("+10e+3", 1, "10000.0")
      check(".10e-3", 1, "0.0")
      check("10.10e-1", 1, "1.0")
      check("10.01e-2", 1, "0.1")
      check("+.10e-10000", 1, "0.0")
      check("+10.10e-100", 1, "0.0")
      check("+10.01e1", 1, "100.1")
      check("-.10e1", 1, "-1.0")
      check("-10.10e-1", 1, "-1.0")
      check("-10.01e-2", 1, "-0.1")
      // "Not a floating point number" tests

      // But see http://acm.timus.ru/forum/thread.aspx?id=10235&upd=632909844439020000
      // Posted by Kopeliovitch Sergey (1) 30 Jul 2004 21:29
      // Segey say his progamm have AC with other answers, but right answers are
//Update: Now Segey have WA #1 and WA #12 and didn't have AC.
      check("10.", 1, "Not a floating point number")
      check("10.0e", 1, "Not a floating point number")
      check("10.e", 1, "Not a floating point number")

      //
      check("1E1 1", 1, "Not a floating point number")
      check("-+3", 1, "Not a floating point number")
      check("+", 1, "Not a floating point number")
      check("-", 1, "Not a floating point number")
      check(" ", 1, "Not a floating point number")
      check("-10.01ee", 1, "Not a floating point number")
      check("12.e10", 1, "Not a floating point number")
      check("-.e10", 1, "Not a floating point number")
      check("1e", 1, "Not a floating point number")
      check("e1", 1, "Not a floating point number")
      check("e", 1, "Not a floating point number")
      check("se-1", 1, "Not a floating point number")
      check("+e", 1, "Not a floating point number")
      check("-e1", 1, "Not a floating point number")
      check(".e4", 1, "Not a floating point number")
      check("0.", 1, "Not a floating point number")
      check("1e", 1, "Not a floating point number")
      check("0x123", 1, "Not a floating point number")
      check("1E2.1", 1, "Not a floating point number")
      check("-1E2.1", 1, "Not a floating point number")
      check("1E2.1e2", 1, "Not a floating point number")
      check("1e00-1", 0, "Not a floating point number")
      // # tests
      check("1#", 1, "Not a floating point number")
      check("#1", 1, "Not a floating point number")
      check("# 1", 1, "Not a floating point number")
      check("1 #", 1, "Not a floating point number")
      // big power test
      check("1e10", 1, "10000000000.0")
      check("1e10", 0, "10000000000")
      check("-1e10", 0, "-10000000000")
      check("1e-10", 0, "0")
      check("0e+10", 0, "0")
      check("0e+100", 0, "0")
      check("0e+10000000000000000000", 0, "0")
      check("0e-10000000000000000000", 0, "0")
      check("0e10000000000000000000", 0, "0")
      check("1e000000000000000000000000000000000000000000000000000000000000000000000000000000", 0, "1")
      check("1e000000000000000000000000000000000000000000000000000000000000000000000000000001", 0, "10")
      check("1e-000000000000000000000000000000000000000000000000000000000000000000000000000001", 0, "0")
      check("234565432452433e-10000000000000000000", 0, "0")
      check("-234565432452433e-10000000000000000000", 0, "0")

      // An integral part of a number should be non-empty.
      check(".1e+0", 1, "0.1")
      check("-.1e+0", 1, "-0.1")
      // There shouldn't be leading zeroes in a non-zero integral part of a number.
      // A zero integral part of a number should consist of an only zero.
      check("00000.1e+0", 1, "0.1")
      check("-00000.1e+0", 1, "-0.1")
      check("00001.1e+0", 1, "1.1")
      check("-00001.1e+0", 1, "-1.1")
      check("00001.1", 1, "1.1")
      check("-00001.1", 1, "-1.1")
      check("00001.1", 0, "1")
      check("-00001.1", 0, "-1")
      // A fractional part of a number should contain exactly N symbols.
      check("00001.1", 0, "1")
      check("00001.1", 3, "1.100")
      check("-00001.1", 1, "-1.1")
      // There shouldn't be a sign "+" before a positive number.
      check("+00001.1", 0, "1")
      check("+1.1", 0, "1")
      check("+1.1", 3, "1.100")
      check("+1.1e+1", 1, "11.0")
      // You shouldn't round off a number.
      check("+1.5e+1", 1, "15.0")
      check("+1.06", 1, "1.0")
      check("+1.6e-1", 1, "0.1")

Edited by author 29.03.2017 22:42

Edited by author 29.03.2017 22:47

Edited by author 30.03.2017 00:56

Thank's for tests, it was very useful (saved a lot of my time, but didn't help).
I had WA7 too, but i found my bug (forgot about re-reverse in code).
Found mistake after test: check("12345.6789e1", 3, "123456.789");