On Wed, 3 Oct 2001, Rasmus Lerdorf wrote:
> Hrm.. So Sybase does not need nulls escaped? Are you sure?
>
> Isn't the correct fix to make stripslashes() check for \0 ?
>
> -Rasmus
I don't have a sybase setup on which to test, but oracle uses the same
addslashes method and should not have nulls
addslashes currently backslash escapes \0 regardless of the value of
magic_quotes_sybase. if magic_quotes_sybase is on, it should not do so. a
patch to change this is very small and simple. see attached.
this is a particularly bad thing to leave unpatched, as stripslashes will
not remove backsl
ok, this is a little more bizarre than I thought it was.
apparently, tzset is called in mktime or when putenv('TZ=foo') are
called. so this works, and putenv'ing TZ isn't thread safe already, but
should the mktime tzset call be taken out? and maybe some thread safe
method of setting a timezone
there are quite a few bug entries that deal with timezones and the TZ
environment veriable... they're marked as closed and reentered in each
version release. from what I can gather from the code, it's still not
fixed.
localtime_r does not set the global tzname from TZ when called. I believe
