Incorrect use of extended RE in script

[Jens Schweikhardt]

Tom et al,

I'm looking at # $Id: MKterminfo.sh,v 1.10 2002/06/29 20:04:28 tom Exp $
as found in ncurses 5.3.

There's a sed script apparently trying to use an Extended Regular
Expression, \+, but sed only accepts Basic Regular Expressions (as per
SUSv3, see
http://www.opengroup.org/onlinepubs/007904975/utilities/sed.html)

The script in question is (tabs and ^E not preserved):

sed -n <$caps "\
...
s/  [A-Z0-9_()\-]\+ [0-9\-]\+   [Y\-][B\-][C\-][G\-][EK\-]\**   /   T{^E/
...

Thus, the resulting terminfo.5 has horribly broken tables, e.g. on
FreeBSD, where it looks like
...
       These are the boolean capabilities:

               Variable                 Cap-         TCap          Description
               Booleans                 name         Code
       auto_left_margin                              bw            bw
       T}
       auto_right_margin                             am            am
       T}
       back_color_erase                              bce           ut
       T}
       can_change                                                  ccc
       T}
       ceol_standout_glitch                          xhp           xs
       T}
       col_addr_glitch                                             xhpa
       T}
       cpi_changes_res                                             cpix
       T}
       cr_cancels_micro_mode                         crxm          YB


I'm not sure what the \+ is supposed to do. If it's intended as
a quantifier for "one or more", I suggest

s/      [A-Z0-9_()\-][A-Z0-9_()\-]*     [0-9\-][0-9\-]* 
[Y\-][B\-][C\-][G\-][EK\-]\**   /       T{^E/

Regards,

        Jens         <schweikh at FreeBSD.org>
-- 
Jens Schweikhardt http://www.schweikhardt.net/
SIGSIG -- signature too long (core dumped)