On Wed, 14 Mar 2001 [EMAIL PROTECTED] wrote:
> > at any script that made use of the proposed change to
> > VMS::Stdio::binmode(). Are you prepared to handle that?
>
> Yep as it is only ONE change at the start of an inhouse written script.
> and no changes to whatever modules we do use (like: smtp, ftp, rcmd alike, cgi
> for example.) and I don't want to check and possibly modify code in there
> as that makes future upgrades a nightmare.
OK.
> > Are you here referring to CORE::binmode() as the "not an add on"?
> --8<--
> Yes, it used to be implented as a mybinmode, in the VMS.C specific part/
> The core binmode only checks for the DOS/Windows platform and nothing else
> (#ifdef...) and is a NOOP otherwise. In the past there were check like
> #ifdef mybinmode ... etc. this seems to be deprecated.
Apparently you did not get a chance to comment on the change to
vms.c:mybinmode() that was made quite a while ago.
For anyone who relies on the current behavior of VMS::Stdio::binmode() we
could copy that code into a new xs routine. For illustration let us say
that we turn what is the current VMS::Stdio::binmode() into a new xs
routine called instead VMS::Stdio::oldbinmode(), then we could add your
"ctx=bin" to the freopen in a new VMS::Stdio::binmode() and you'd be able
to write your cross platform scripts in manner that you'd like. Anyone
who was dependent on the old behavior would need to call oldbinmode()
rather than binmode().
Does anyone object to such a change? Can anyone suggest a better name
than "oldbinmode()", "binmode_sans_ctx_bin()",
"binmode_with_ctx_bin()", etc...?
Peter Prymmer
