🚀 go-pugleaf

rvloon@cv.ruu.nl (Ronald van Loon) writes:

:In <1992Oct26.151002.24470@cadlab.sublink.org> martelli@cadlab.sublink.org (Alex Martelli) writes:
:
:|"I was just about to go ahead and install vile everywhere on this site
:|"(after a trial run on PC's and HP /400 and /700 WS's, all rather
:|"satisfactory), when I saw xvi appear in comp.sources.misc!  Nach,
:|"always too much choice...:-)
:
:Well, I couldn't get xvi to compile on our HPs, while I got (x)vile working in
:no-time at all. Maybe that goes for something.

Thanks for the info, but I don't feel this problem is very important,
actually...  vile goes up straight away because somebody already got
it working on Hp/Ux (there is a specific hpux flag for the make), xvi
apparently never had anybody port it to Hp-Ux yet but it's a no-brainer:

- makefile.hpu can be made from an ALMOST straight copy of makefile.usg,
  EXCEPT -DPOSIX should also be defined;
- in xvi.h, the lines:
#   ifdef   hpux
#   	define	USE_STDHDRS
#   endif   /* sparc */
  should be inserted right after the identical ones for #ifdef sparc
  (using #if defined(sparc)||defined(hpux) would be cleaner, but I assume
  there must be some reason in xvi.h to use the goofy #ifdef in lieu of
  the cleaner #if defined()?)
- in unix.c, the reference to a mysterious "c_ospeed" member of the
  cooked_state structure (which is a struct termios) must be replaced
  by a call of the cfgetospeed() function:
	/* ospeed = cooked_state.c_ospeed; */
        ospeed = cfgetospeed(&cooked_state);
  Note that the POSIX standard does NOT mandate a member called c_ospeed
  in the struct termios, but rather states that "the input and output
  baud rates are stored in the termios structure" and accessed through
  the interfaces named cfgetispeed(), and so on [the BAD use of "baud
  rate" is from the X/Open Portability Guide Issue 3, Dec. 1988, XSI
  System Interfaces and Headers, which I generally prefer to consult
  in lieu of the true Posix standard although it fails here... sorry].

All in all, and though I admit to having performed no deep testing, I
just wish all software from the net went up on our HP boxes with as
LITTLE hassle as this!  (...maybe put these little portability fixes
up in the first patch to xvi...?)

BTW, the summary of what I got by email is coming shortly, I was just
waiting a bit to get more since I received very little feedback on this
request...
-- 
Email: martelli@cadlab.sublink.org                   Phone: ++39 (51) 6130360
CAD.LAB s.p.a., v. Ronzani 7/29, Casalecchio, Italia   Fax: ++39 (51) 6130294