This is the mail archive of the mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: LFTP: cygwin and setupterm

On Tue, Dec 17, 2002 at 12:37:45AM -0500, wrote:
>Just checked on both of my servers, they symlink /usr/include/term.h with
>ncurses/term.h, running RedHat 6.1 and 8.0. Should cygwin/ncurses do the
>  The QNX proprietary term_* functions have been deprecated in favor of
>  ncurses. ncurses is a set of terminal-independent routines for painting
>  screens and handling input events......
>  .....The file /usr/include/term.h is now an ncurses header file; you'll
>  find the old <term.h> in /usr/include/sys/term.h. An error message is
>  displayed if you combine the old term_* and ncurses header files.

Since the code clearly checks for the situation of finding the needed headers
in /usr/include/ncurses, I don't see any reason why this package should be
used as a justification for adding a symlink.

>Btw, I put the below __CYGWIN__ mention in because it still complained even
>when when it was passed the sufficient defines in CPPFLAGS and CXXFLAGS
>before running configure.

Usually when I see things like HAVE_NCURSES_CURSES_H, it suggests to me
that the author of the source code is correctly attempting to set things
up in a general, non-system-specific way and is testing for various
capabilities and oddities of the operating environment.  System-specific
checks like "#ifdef ultrix" in source quickly become unmanageable.
That's why HAVE_NCURSES_CURSES_H is nice.

There's no reason why it couldn't work in this scenario.  Of course, if it
"complained" maybe all you need to do is rebuke it sternly.


Unsubscribe info:
Bug reporting:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]