Could we postpone this change for after 5.10.0 ?
On 16/11/2007, John E. Malmberg <[EMAIL PROTECTED]> wrote: > There is no reason anymore to have VMS::Filespec::vms_realpath be > dependent on symbolic links being present. Same for > VMS::Filespec::vms_case_tolerant. > > It looks like another API is needed. vms_realpath will always return a > UNIX format pathname, and for for compatibility with existing code, we > need a routine that returns a VMS format pathname. > > To keep things simple, I would like that routine to have a different > name than vms_realpath, such as vms_realname. > > This is so that vms_realname() will always return the "primary link" in > VMS format to a file. > > vms_realpath() first attempts to return the realpath() in UNIX format to > a file, and if it is unable to do that for any reason, it falls back to > returning the name of the primary link to a file in UNIX format. > > I will update the VMS::Filespec pod after I get the vms_realname api > implemented.
