The explicitly defaulted or implemented move constructors and assignment
operators are made "noexcept".
Bugfix:
* src/stream.cpp Stream::Stream() char_traits::int_type intro[4] is
now aggregate-initialized (to zero) to avoid UB.
Minor changes:
* Using std::isinf() and std::signbit() instead of comparing for
equality with infinity.
* src/streamcharsource.h: Added #include "stream.h".
* src/stream.h: Forward declaring "class StreamCharSource".
* Some implicit casting changed into static_cast's.
Signed-off-by: Ted Lyngmo <ted@lyncon.se>
This is in preparation for other patches that will make use of the
macro. The patch also removes #undef:ing the macro after its been
used to not make the header inclusion order critical. Otherwise,
the new header would always have to be the last of the yaml-cpp
headers to be included.
Invalid access via operator[] or as<> will now print the offending key, if possible.
For example:
a:
x: 1
y: 2
node["a"]["z"].as<int>()
will say that the key "z" was invalid.
* Add compilation flags: -Wshadow -Weffc++ -pedantic -pedantic-errors
* Delete implicit copy & move constructors & assignment operators
in classes with pointer data members.
* An exception to the above: Add default copy & move constructors &
assignment operators for the Binary class.
* Convert boolean RegEx operators to binary operators.
* Initialize all members in all classes in ctors.
* Let default ctor delegate to the converting ctor in
Binary and RegEx
* Don't change any tests except regex_test (as a result of the change
to binary operators).
Note: https://bugzilla.redhat.com/show_bug.cgi?id=1544675 makes
-Weffc++ report a false positive in "include/yaml-cpp/node/impl.h".
The issue is that numbers like
2.01 or 3.01 can not be precisely represented with binary floating point
numbers.
This replaces all occurrences of 'std::numeric_limits<T>::digits10 + 1' with
'std::numeric_limits<T>::max_digits10'.
Background:
Using 'std::numeric_limits<T>::digits10 + 1' is not precise enough.
Converting a 'float' into a 'string' and back to a 'float' will not always
produce the original 'float' value. To guarantee that the 'string'
representation has sufficient precision the value
'std::numeric_limits<T>::max_digits10' has to be used.
* Removed an expression which is always true
* The second expression (ch is space) is removed because the first one contains space 0x20
* nextEmptyLine is always false so it is removed from the expression
This way it's possible to build using older MSVC (<13) that don't
support this yet. Macro is undefined in each file where it is used so it
should stack well with other libs and sources.
* Change node_map type from map<ptr,ptr> to vector<pair<ptr,ptr>>
Map nodes are now iterated over in document order.
* Change insert_map_pair to always append
Always append in insert_map_pair even if the key is already present.
This breaks the behavior of force_insert which now always inserts KVs
even if the key is already present. The first insert for duplicated keys
now takes precedence for lookups.
This improves performance on the test.yaml attached to #158 by about
25% on my machine as compared to the previous commit (0.25s -> 0.20s),
as measured by `time build/util/parse < test.yaml > /dev/null`.
This improves performance on the test.yaml attached to #158 by about
35% on my machine (0.39s -> 0.25s), as measured by
`time build/util/parse < test.yaml > /dev/null`.
