[Gnash-dev] Re: code question

[Martin Guy]

> This causes trouble with BSD systems where timezone is a function, char
> *timezone(int,int)

Oh, wonderful :-/ This is why I hate date code, as if the stupid
politicians' arbitrary decisions didn't make it hard enough already.

> HAVE_TZSET also appears to be a problem. So we need
> to add a config test for the different timezone types.

I need documentation on BSD and what timezone docs/returns, as well as
a BSD system to test it on. I suggest soeone else will get there
sooner than I do.

On Linux/glibc the only reliable way to get the tzoffset is
tzset/timeone and the GNU manual says "CONFORMING TO  SVr4,
POSIX.1-2001, 4.3BSD". I guess that's not true.

There is yet another way to find the timezone offset on LInux that may
work on BSD too. Anyone caring to look where tzset/timezone is used
will see it is already documented at line 497; if that is available
then use that in preference to tzset/timezone with 2 ifdefs. However I
don't have a BSD system to try that on; I will build one of those too
eventually I guess, unless someone wants to give me access to one
(bjacques never replied to me when I asked).

Or rewrite it to drag in libbosst/datetime I suppose but fixing a tiny
bug in working code seems more sane than rewriting it to add yet
another bloated library module. It was hard enough to get right the
first time, but there is a comprehensive test suite for correct
functioning if anyone wants to try that out.

I've disabled use of tzset/timezone for now so it should compile

  M