[Quickfix-developers] Re: UtcTimeStamp implementation bugs

[Caleb E. <cal...@gm...>]

On 7/11/05, Caleb Epstein <cal...@gm...> wrote:
>     // mktime converts a struct tm (IN LOCAL TIME) to a time_t
>     time_t t =3D mktime( (tm*)this );
>     // Converts a time_t back to a struct tm.  Presumably this is done
>     // for normalization so that a user specifying a date of 31-02-2005
>     // will end up with a more correct 02-03-2005?
>     *static_cast<tm*>(this) =3D time_localtime( &t );

[...]

> My impression is that the mktime and localtime calls (the latter of
> which is redundant - POSIX mktime modifies the passed-in struct tm in
> place so you don't need to call localtime at all) are used only to
> normalize values that might be specified incorrectly by the user (e.g.
> February 30th or the like)

Well, my impression was wrong :-)

There is code in SessionTime which uses the tm_wday (via getWeekDay)
so the call to mktime is necessary to populate this field, otherwise a
couple of tests in SessionTimeTest.cpp suite fail.

> But, because of their localtime semantics and the way that QuickFIX
> populates the structures with GMT values, they actually cause breakage
> in certain edge situations.  These functions are also performance pigs
> on some platforms (notably Solaris before Solaris 9) where I've seen
> QuickFIX spend ~50% of its runtime in mktime calls.
>=20
> For these reasons, I'm going to submit a patch that drops the
> mktime/localtime usage (and some unncessary setCurrent calls) from the
> UtcTimeStamp class which should end up giving a small performance
> boost.  Can anyone see a flaw in this analysis?

My approach now is to completely replace the underpinnings of
UtcTimeStamp and the related classes so they don't use "struct tm" at
all, but instead store a double representing the Julian day number and
the fractional portion of a day.  This eliminates the need for a lot
of the time-related system calls and static_cast<tm*> weirdness in
many places.

I'll send along the patch once I can get the code to stop failing the
bulk of the tests :-)

--=20
Caleb Epstein
caleb dot epstein at gmail dot com