Re: LYNX-DEV HP-UX weirdness with 2-7ac0.28 solved

[T.E.Dickey]

> The generic, dgux, freebsd, and netbsd targets have 
> LYFLAGS=$(CFLAGS) -I../../.. $(SITE_LYDEFS)
right.  Those targets have been modified to support the configure script.
 
> Prepending LYFLAGS and MCFLAGS with $(CFLAGS) -I../../.. in the 
> snake3 target seems to "make things work."
> 
> So the snake3 target which works for me with ac-0.28 is:
> snake3:
>         cd WWW/Library/snake; $(MAKE) CC="cc" LYFLAGS="-Ae $(CFLAGS) \
>                 -I../../..  $(SITE_LYDEFS)"
>         cd src; $(MAKE) all CC="cc" MCFLAGS="$(CFLAGS) -I../../.. \
>                 -O -Ae -DFANCY_CURSES -DUNIX \
>                 -DSNAKE -I../$(WWWINC) $(SITE_DEFS)" \
>                 LIBS="-lcurses -ltermcap \
>                 $(WAISLIB) $(SOCKSLIB) $(SITE_LIBS)" \
>                 WWWLIB="../WWW/Library/snake/libwww.a"
> 
> Any reason why this shouldn't be done for all the targets in the
> ac code set? 
yes/no.  I was trying to keep the initial diffs from Makefile down to the
places where I could test.  I'll incorporate your mods into the next
patch (but adding a test for the -Ae option).

(what is "SNAKE"?)

I'm proceeding with makefile.in by testing the platforms that
I know about, moving the -D's and other special options out of the
logic into the configure-script's logic.  The ones I've hit are marked
TESTED.  The 'snake' targets afaik are oriented around the ANSI compiler
(I don't have that, but can make some guesses).  So they won't work
with the configure script.

So I built HP-UX last week by configuring, using the linux target
(since it's the closest to being "clean").  Once they're clean, I'll
combine targets, and try to get testers for the remaining stuff.
 
> One other little thing with the snake3 target -- HP-UX 9's unbundled
> cc won't handle -O and -g together [although its optimizer is quite
> nice].  It generates a warning for each invocation.  It'd be nice
> if use of -g was a configure-time option.
you can set CFLAGS to override.  (The configure script tries to, but cannot
always determine if that combination is legal -- HP-UX is one where it warns,
but does nothing else).

> Blah.  It'd be nice if ``we'' could simplify the current mess of makefile 
> stuff significantly.
that's my goal.
 
> BTW, lynx with color curses is neato.
which curses? (HP-UX 10.x has curses_colr, or was that ncurses?)

-- 
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey
;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;