readline issues?

[Fairlight <fairlite at iglou dot com>]

I have the latest readline that is available for cygwin.  Readline 4.1-2
according to the CYGWIN-PATCHES directory in the source.

I was trying to port tintin++ 1.84 this evening.  bzero seemed to get
complained about but I simply did a (char *) cast on arg1 in both
instances.

The critical failure seems to be in the linking to readline:

---- Linking...
gcc -Iprotos  -O2 -D_GNU_SOURCE -DHAVE_CONFIG_H      -o tt++ action.o alias.o an si.o antisub.o files.o glob.o help.o highlight.o ivars.o llist.o log.o main.o mi sc.o net.o parse.o path.o rl.o rlhist.o rltab.o rltick.o rlvt100.o scrsize.o ses sion.o substitute.o text.o ticks.o utils.o variables.o function.o bindkey.o walk .o chat.o -lreadline -lhistory -ltermcap
action.o(.text+0x584):action.c: undefined reference to `readline_echoing_p'
parse.o(.text+0x43):parse.c: undefined reference to `readline_echoing_p'
rl.o(.text+0x54):rl.c: undefined reference to `_rl_last_c_pos'
collect2: ld returned 1 exit status
make: *** [tt++] Error 1

I've looked at the cygwin patched source for readline--and both are included.

I've also noted that I tried an old trick (thinking maybe it wasn't seeing
libreadline at first) and explicitly declared /lib/libreadline.a  ...It
actually complains MORE.  I get identical results to the above if I link
against the .dll.a though.  So it's seeing it...BUT--

Why are they considered undefined when they're so obviously in the library?

Anyone shed some light on this?  I'm displeased with wintin.  :)

mark->
-- 
Fairlight->   |||                                    | Fairlight Consulting
  __/\__      |||     "I'm talking for free,         |                        
 <__<>__>     |||      It's a New Religion..."       | http://www.fairlite.com
    \/        |||                 --Le Bon           | info@fairlite.com                     

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple