[clang-tidy] Add missing check documentation.
authorHaojian Wu <hokein@google.com>
Fri, 17 Aug 2018 19:50:22 +0000 (19:50 +0000)
committerHaojian Wu <hokein@google.com>
Fri, 17 Aug 2018 19:50:22 +0000 (19:50 +0000)
llvm-svn: 340075

clang-tools-extra/docs/clang-tidy/checks/abseil-duration-division.rst [new file with mode: 0644]

diff --git a/clang-tools-extra/docs/clang-tidy/checks/abseil-duration-division.rst b/clang-tools-extra/docs/clang-tidy/checks/abseil-duration-division.rst
new file mode 100644 (file)
index 0000000..b7c8635
--- /dev/null
@@ -0,0 +1,36 @@
+.. title:: clang-tidy - abseil-duration-division
+
+abseil-duration-division
+========================
+
+``absl::Duration`` arithmetic works like it does with integers. That means that
+division of two ``absl::Duration`` objects returns an ``int64`` with any fractional
+component truncated toward 0. See `this link <https://github.com/abseil/abseil-cpp/blob/29ff6d4860070bf8fcbd39c8805d0c32d56628a3/absl/time/time.h#L137>`_ for more information on arithmetic with ``absl::Duration``.
+
+For example:
+
+.. code-block:: c++
+
+ absl::Duration d = absl::Seconds(3.5);
+ int64 sec1 = d / absl::Seconds(1);     // Truncates toward 0.
+ int64 sec2 = absl::ToInt64Seconds(d);  // Equivalent to division.
+ assert(sec1 == 3 && sec2 == 3);
+
+ double dsec = d / absl::Seconds(1);  // WRONG: Still truncates toward 0.
+ assert(dsec == 3.0);
+
+If you want floating-point division, you should use either the
+``absl::FDivDuration()`` function, or one of the unit conversion functions such
+as ``absl::ToDoubleSeconds()``. For example:
+
+.. code-block:: c++
+
+ absl::Duration d = absl::Seconds(3.5);
+ double dsec1 = absl::FDivDuration(d, absl::Seconds(1));  // GOOD: No truncation.
+ double dsec2 = absl::ToDoubleSeconds(d);                 // GOOD: No truncation.
+ assert(dsec1 == 3.5 && dsec2 == 3.5);
+
+
+This check looks for uses of ``absl::Duration`` division that is done in a
+floating-point context, and recommends the use of a function that returns a
+floating-point value.