Re: BEAST on FreeBSD: timezone

[Tim Janik <timj gtk org>]

On 14 Feb 2004, Rasmus Kaj wrote:

> I've verifyed thats what causing the assert failure, and I have a
> patch for it (attached).  And I actually thinks that this method is
> better, as it makes it more obvious that the value used is gotten from
> the call to localtime (which seems to be of no use in the original
> code, but actually sets the timezone as a side effekt).

thanks, applied. the possible caveat with that is just that the tm_gmtoff
field of struct tm was introduced only with newer libc versions
on linux. we'll see if that still breaks on someone else's system ;)

> However, I havn't actually got it working on FreeBSD 5.2 yet, as
> testsfi crashes with a segv while checknig Threading (but after
> completing the Time and some other tests).  I hope to find the time to
> debug this one of the nearest days.

oh. beast basically uses the threading layer provided by GLib plus direct
posix function calls for locks/conditions if it finds them. that's so we
can use fast posix locks if they are available. so long as threading
works for GLib for you, you should be able to get beast to work pretty
easily.

> --
> Rasmus Kaj ----------------------------------------------- rasmus@kaj.se
>  \                                      Put no trust in cryptic comments

---
ciaoTJ