Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!bu.edu!snorkelwacker!tut.cis.ohio-state.edu!ASYLUM.GSFC.NASA.GOV!chris
From: chris@ASYLUM.GSFC.NASA.GOV (Chris Shenton)
Newsgroups: gnu.gcc
Subject: Problem with #include for gcc on VMS (also, getting g++ running)
Message-ID: <9002282057.AA19478@asylum.gsfc.nasa.gov>
Date: 28 Feb 90 20:57:24 GMT
Sender: daemon@tut.cis.ohio-state.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 29

I recently got gcc running on VMS 4.4 (yech!), and have two problems with
the #include directive:

1.  My logical SYS$LIBRARY doesn't get expanded properly when looking for
    things like curses.h. It's currently defined as ``SYS$SYSROOT:[SYSLIB]''
    but if I redefine it without the embedded logical to be the fully
    qualified name (``PITCH$DUA0:[SYS0.SYSLIB]'') then it works OK. Any 
    suggestions? 

2.  Some of the things included by the VMS #include files use the non-ansi
    brain-damage ``#include modulename'' (note lack of quotes or brackets); the
    VMS C compiler looks in a special VMS text library for the proper module.

    Has anyone hacked the preprocessor to do this for you? The only alternative
    I can see is to un-library all the VMS text libraries and change this
    abomination by hand or program, possibly introducing bugs and
    incompatibilities with future VMS updates.

Also:
    I'd like to get g++ running on VMS. My plan is to merge the g++ and gcc
    sources on my unix box, then move the distribution to VMS and compile
    there. Any other suggestions? has anyone gotten it to work? I imagine I'll
    still have the same problems mentioned above...

TIA.
_______________________________________________________________________________
Internet:  chris@asylum.gsfc.nasa.gov (128.183.10.155)      NASA/GSFC: Code 735
UUCP:      ...!uunet!asylum.gsfc.nasa.gov!chris             Greenbelt, MD 20771
SPAN:      PITCH::CHRIS                                            301-286-6093