http://d.puremagic.com/issues/show_bug.cgi?id=5781
Unfortunately, _all_ of the unit tests are now disabled on Windows
(instead of just most of them). Issue #5454 seems to be worse at present.
Hopefully Don's changes to CTFE will help.
toISOExtendedString is annoyingly long - albeit wonderfully accurate -
and toISOExtString seems accurate enough and enough shorter to be worth
the switch. The old names are scheduled to be deprecated.
putenv actually keeps a pointer to the string that it's passed, which
setTZEnvVar didn't take into account, so putenv risked having an invalid
pointer as soon as a garbage collection cycle was run. This is an
example of a bug that's easier to find in C++. The inderministic nature
of garbage collection makes it much more insidious. But it's the kind of
thing that's really only going to happen when interacting with code that
isn't D code.
http://d.puremagic.com/issues/show_bug.cgi?id=5761
A subtle math error involving leap years caused Dec-30 and Dec-31 of A.D.
leap years and Jan-01 and Jan-02 of B.C. leap years to be off by one day.
The revised tests catch the bug.
They tests that were changed are a bit scattered, because I was trying
to pick the tests that were related to the test failure in SysTime's day
getter property (which is a manifestation of bug# 5761).
There should be no functional changes to the code - only the unit tests.
Also, the lines which have been changed are now mostly within 80
characters, and those that are over are not over by much. Only a portion
of the tests have been updated, but I'm doing it in chunks.
assertThrown, assertNotThrown, and collectExceptionMsg were removed
from std.datetime, since they're in std.exception now, and assertPred
was renamed to _assertPred in case anyone is trying to use the function
which was under review in their own code (since all templates are
currently public even if declared private).
There should be no functional changes to std.datetime. It's just
adjusting which versions of the unit test functions get called.
It will still use daylight if it's there, but since FreeBSD doesn't have
it for some reason (in spite of the fact that the daylight variable is
part of Posix), we can't assume that it's there.
Also, now that tzname is in druntime, it has been removed from
std.datetime.
This should fix bug# 5616.
I mostly cleaned up and made what was already there more consistent, but
I also reduced some of the documentation in cases where there were
multiple template specializations which differed only by template
constraints and not by signature, so it looked like the same function
was showing up multiple times in the documentation (since the
documentation doesn't show the template constraints). Also, most, if not
all, of the documentation is now within 80 characters per line, though I
haven't done anything to reduce the width of the code lines yet.
I'll work on finding ways to reduce the amount of documentation without
degrading its quality later.
Also, the Windows-specific unit tests have now been enabled. So, while
unfortunately, we can't currently have them all enabled on Windows,
since it causes dmd to run out of memory, we can at least get the
Windows-specific stuff tested. Hopefully, the fact that the rest are
running on Linux will be enough for now as far as the other tests go.
std.gregorian, std.date, and std.dateparse have all been marked as
scheduled for deprecation. Everywhere (except for std.file) which was
using std.date is now using std.datetime. std.file is now using
std.datetime but has a number of functions still using d_time but which
are marked as scheduled for deprecation. I tried to give as many as I
could pragmas indicating that they were scheduled for deprecation, but
at the moment, that requires that a function be a templated function, and
I couldn't templatize all of them. So, some functions in std.file are
only marked as scheduled for deprecation in their documentation and will
not give any warning on compilation.
I had to rename several functions in std.file in order to avoid making
any breaking changes. And since I was already having to mess with
function names, it seemed like a good time to change the names of a
number of the functions in std.file to use proper capitalization
(such as changing isdir to isDir) as has been discussed and overwhelmingly
supported in the newsgroup with regards to std.string. And since I was
making those changes, it seemed like a good time to fix bug# 3848
(functions in std.file don't take symbolic links into account) as well.
So, std.file should now deal with symlinks properly.
The issue which Andrei brought up with +VERSION causing the std.datetime
unit tests to fail on OSX has been fixed as well.
For some reason, the std.datetime unit tests compile and pass without
problems on Linux and in Wine on Linux, but on real Windows, the
compiler runs out of memory when compiling them. So, for now, all of the
std.datetime unit tests are versioned with testStdDateTime. On Posix,
version=testdStdDateTime is used in std.datetime, so the tests will run,
but that version is not set on Windows, so the unit tests don't run on
Windows. So, if you want to run the std.datetime unit tests on Windows or
under Wine or Linux, then compile with -version=testStdDateTime.
At the moment, all of the tests pass on both Linux and under Wine, and
the normal code does compile on proper, so I'm fairly certain that it's fine
for use on Windows, but for now, I can't enable the tests on Windows. My
best guess is that it's some sort of weird compiler bug involving system
calls that leak memory on Windows-proper but not in Wine, but I really
don't know.
Nothing in Phobos has been made to use it yet and nothing has yet been
marked as deprecated or scheduled for deprecation due to the new
std.datetime being checked in. That's for a future commit.
This function came from std.date.
But now, benchmark returns Ticks[N] and don't use result argument.
I removed need to use GC for by doing it in this way.
