Steven,
>The trailing semi-colon may matter little,
On the contrary, it matters a LOT. The trailing semi colon is an explicit "most recent version number". Whereas if it's missing, the version is unspecified, and can be therefore be inherited from somewhere else.
Consider (using your symbol names):
$ newfile1= F$PARSE("NEWFILE", this_proc_novers, some_other_filespec)
$ newfile2= F$PARSE("NEWFILE", this_proc_simple, some_other_filespec)
newfile1 will always map to the latest version of NEWFILE, whereas newfile2 can pick up a version number from some_other_filespec, which may result in a filespec which doesn't exist.
Most of the time it doesn't make sense to inherit version numbers between file specifications, so it's better to use an explcit ";" than remove the version number altogether.
FYI: I've put logged a request with OpenVMS engineering for F$PARSE to be enhanced to allow multiple fields to be returned together (there are precedents for this type of syntax in other lexical functions). It's a bit silly having to invoke $PARSE multiple times, when each instance has the complete set of fields available. If they ever action my request, this:
$ this_proc_novers = f$parse( this_proc, , , "device")+ -
f$parse( this_proc, , , "directory")+ -
f$parse( this_proc, , , "name")+ -
f$parse( this_proc, , , "type")
could become:
$ this_proc_novers = f$parse(this_proc, , , "device,directory,name,type")
which would make this kind of filename manipulation much simpler.
(though my perception of OpenVMS engineering at the moment is, unfortunately, they seem far to scared to change anything, so I'm not optimistic of ever seeing it happen).
A crucible of informative mistakes
