The repeated sequence of allocating a CharString and CharStringByteSink,
before calling some function that writes into this, can be moved into a
single shared helper function which then is used to give all ulocimp.h
functions that write to ByteSink an overload that instead returns a
CharString, to make call sites look like perfectly normal C++ code.
Some of this code was originally written as C code and some of this code
was originally written as C++ code but made to resemble the then already
existing code that had once been C code. Changing it all to normal C++
now will make it easier and safer to work with going forward.
· Use unnamed namespace instead of static.
· Use reference instead of non-nullable pointer.
· Use bool instead of UBool.
· Use constexpr for static data.
· Use U_EXPORT instead of U_CAPI or U_CFUNC.
· Use the default calling convention instead of U_EXPORT2.
This is the normal standard way in C, C++ as well as Java and there's no
longer any reason for ICU to be different. The various internal macros
providing custom boolean constants can all be deleted and code as well
as documentation can be updated to use lowercase true/false everywhere.
Remove the ICU macros ATOMIC_INT32_T_INITIALIZER and U_INITONCE_INITIALIZER,
which made use of C++ ATOMIC_VAR_INIT, which has been removed from C++20.
With modern C++ features being available, these macros no longer served
any real need.
Revise uses of UVector in Formatting related code to better handle memory
allocation failures. This is one of an ongoing series of commits to address
similar problems with UVector usage throughout ICU.
The changes primarily involve switching uses of UVector::addElementX() to the
new adoptElement() or addElement() functions, as appropriate, and using
LocalPointers for tracking memory ownership.
Now when all equality operators return standard bool (commit 633438f),
it no longer makes any sense to use the ICU4C constants TRUE & FALSE
or local variables of type UBool for their return value.
This is the first step towards improving the error handling and out-of-memory
behavior of UVector::addElement(). A followup PR will add back a new addElement()
with corrected error handling, then additional followups will switch call sites
from the original (renamed) function to the new addElement().
This commit includes no logic or behavior changes; it only renames the existing functions.
If you call the API getDefaultHourCycle on an empty DateTimePatternGenerator
instance (ie: no locale) then it calls UPRV_UNREACHABLE which calls abort().
We should return an error code instead of aborting.
DateTimePatternGenerator needs to consider the hour-cycle preferred by
Locale. This means that we need to to override the hour-cycle when a
locale contains "hc" keyword. This patch is adding such functionality.
In addition, "DateTimePatternGenerator::adjustFieldTypes" should adjust
hour field to properly follow tr35
spec(https://www.unicode.org/reports/tr35/tr35-dates.html#dfst-hour).
This fixes a regression introduced by commit
b12a927c93 for issue ICU-13778.
The above commit improved the error checking in the
DateTimePatternGenerator class, adding checks for errors/failures
where there previously was none at all. This was done in order to
catch catastrophic errors like out-of-memory (OOM), and properly
report them to the caller, rather than ignoring/hiding these errors.
However, in doing so it exposed a case where the code was depending
on ignoring errors in order to fall-back to the Gregorian calendar
when the default ICU locale is set to root.
This restores the previous behavior, by allowing the error of
U_MISSING_RESOURCE_ERROR to fall-though and continue without
reporting back an error to the caller.
Note: This regression was technically introduced in ICU 63, and
also effects ICU 64 as well.
This eliminates the need for scratch buffers in any code path that ends
with these functions and also eliminates the need for counting bytes,
something that ByteSink will now handle correctly when needed.
Existing calls to uloc_addLikelySubtags() and uloc_minimizeSubtags()
throughout ICU4C implementation code are also updated to instead use
either the Locale or ulocimp_* functions with the new API.
None of this should have any externally visible effect, it's all about
cleaning up implementation internals.
