Fix few typos

Author: arkamar

README.md

@@ -17,13 +17,13 @@ ensure that all documentation and information on the wiki is verified to be
 correct and reasonably good quality.
 
 When making edits for submission, you may want to run `make lint` in order to ensure
-that your newly-added documentation or edits are adherant to the markdown
+that your newly-added documentation or edits are adherent to the markdown
 style. However, it's not really necessary to test locally; Travis CI tests all
 pull-requests and runs lint on them, and fails if the new commits introduce warnings.
 
 In addition to editing locally, you can also edit online.
 For creating new pages, click GitHub's "Create New File" button, type the page
 filename (so, \<filename\>.md), and begin editing there. You can also edit pages
 the same way by going into the wiki directory, clicking a page (the filenames
-corrispond to the page titles) and clicking the edit button next to history.
+correspond to the page titles) and clicking the edit button next to history.
 

mathematical-library.md

@@ -15,7 +15,7 @@ IEEE-754 binary128 (ld128).
 Supported arithmetics evaluations (provided by the underlying platform and
 compiler) are FLT_EVAL_METHOD 0, 1 or 2 (when FLT_EVAL_METHOD!=0 the excess
 precision and range should be handled according to the ISO C semantics: rounded
-away at assingments and casts).
+away at assignments and casts).
 
 The four rounding modes (FE_TONEAREST, FE_UPWARD, FE_DOWNWARD, FE_TOWARDZERO)
 are supported on all platforms with fenv support.
@@ -249,7 +249,7 @@ The standard does not require correct rounding for decimal float constants (the
 converted floating-point value can be the nearest or any neighbor of the nearest
 representible number, but the recommended practice is correct rounding according
 to nearest rounding mode), however hexfloats can represent all values of a type
-exactly and unambigously, so the convention is to prefer hexfloat representation
+exactly and unambiguously, so the convention is to prefer hexfloat representation
 in the math code internally especially when it makes the intention clearer (eg
 0x1p52 instead of 4503599627370496.0). (In public header files they must not be
 used as they are not C++ or C89 compatible).