12

std::flush right after a std::endl is used all over the legacy code I am looking at. When I first saw this, my thought was it is redundant from looking at the description of std::endl and std::flush at:

http://en.cppreference.com/w/cpp/io/manip/endl

http://en.cppreference.com/w/cpp/io/manip/flush

Here is an example of what I am seeing in the legacy source code:

std::cout << "Generic Notification" << std::endl << std::flush;

But, as many senior software developers have seen this code throughout the years, I am wondering if there is some detail I am missing. Is there any purpose to having a std::flush after a std::endl?

Angew is no longer proud of SO
  • 167,307
  • 17
  • 350
  • 455
9Breaker
  • 724
  • 6
  • 16

3 Answers3

17

There's no purpose for it.

If I had to speculate on why your legacy code contains these lines, there are a few possibilities, starting with (what I consider to be) the most probable scenarios:

  • Someone added the explicit call to std::flush mistakenly, and the senior developers didn't consider it a problem needing fixing
  • The code originates from a time before the C++ standard was widely adopted, and in that time, the local compiler's implementation of std::endl did not trigger a flush, which meant that your senior developers were (correctly) understanding that it was necessary
  • An older version of the C++ standard might not have required std::endl to trigger a flush
  • Your senior developers are mistaken about the behavior of std::endl.
  • Your execution environment is a strange behemoth that actually requires output to be flushed twice before the desired result can be expected.
Xirema
  • 19,889
  • 4
  • 32
  • 68
  • 6
    Silly things like this can also appear when text search & replace is performed on source code, such as: replace `<< '\n'`with `<< std::endl` – Drew Dormann Jun 05 '18 at 19:08
  • 2
    Good answer, but "an older version of the C++ standard…" – no, C++98 was the first standard, and it did require a `flush`. At best, some early version of the standard library may have behaved differently, though I doubt it. BTW, I can think of a reason why they would use `endl` instead of `'\n’` despite not realizing it flushes, which is a system that does not have LF line terminator (most likely Windows). Of course this is not necessary on `cout`/`cerr`, nor a text mode `ofstream`, and then again `endl` doesn't even help for a binary `ofstream`, but that may have been the mistaken reasoning. – Arne Vogel Jun 05 '18 at 21:17
  • I particularly like the last scenario. – Kaiserludi Jul 07 '20 at 20:19
8

I'll add to the other valid answers that, very often, there isn't a good purpose for neither std::flush nor std::endl.

Basically, std::endl = start a new line + flush the stream. A lot of people, though, tend to end their lines with std::endl because it "sounds right" - end the line. But we actually rarely need to flush the output stream. Sometimes we do (e.g. when we're expecting a user's reply to the string, or it's important to monitor our output with minimal delay) - but that's the exception, not the rule.

So, it might take a bit of getting used to, but we should really default to simply:

std::cout << bunch_of_stuff << '\n';

and that's that!

einpoklum
  • 118,144
  • 57
  • 340
  • 684
5

In a standards-compliant environment, std::flush serves no useful purpose in this code.

Whoever wrote this either didn't fully understand the semantics of std::endl, or were working around some limitation of their compiler or execution environment.

NPE
  • 486,780
  • 108
  • 951
  • 1,012