Re: install/59033: NetBSD 10.1 - duplicate entries in src/distrib/sets/lists/comp/mi

["Martin Husemann via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR install/59033; it has been noted by GNATS.

From: Martin Husemann <martin%duskware.de@localhost>
To: wgstuken%grimme-reitenspezial.de@localhost
Cc: gnats-bugs%netbsd.org@localhost
Subject: Re: install/59033: NetBSD 10.1 - duplicate entries in
 src/distrib/sets/lists/comp/mi
Date: Mon, 27 Jan 2025 11:43:34 +0100

 On Mon, Jan 27, 2025 at 11:28:06AM +0100, wgstuken%grimme-reitenspezial.de@localhost wrote:
 > Hi Martin - please have a look at my update.
 > 
 > copied wrong set into PR first time .....
 > just run statement again - there are plenty of duplicate lines.
 
 The bug that prevents me from reproducing your result is the -d argument
 to cut - the file is supposed to be tab-separated and maybe your original
 command had -d and a quoted tab there, so you got more output than I when
 I copy&pasted the command line (and split fields by a space).
 
 However: you can not check the lists that way, e.g. for -current comp/mi I see:
 
  > cut -d '       ' -f1 mi | sort | uniq -c |grep -v ' 1 ' 
    2 #
    2 ./usr/include/g++/pstl/parallel_backend_serial.h
 
 but:
 
 > fgrep parallel_backend_serial.h mi 
 ./usr/include/g++/pstl/parallel_backend_serial.h comp-cxx-include       gcc,cxx,libstdcxx,gcc=10
 ./usr/include/g++/pstl/parallel_backend_serial.h comp-cxx-include       gcc,cxx,libstdcxx,gcc=12
 
 
 so the first line is only valid for HAVE_GCC=10 and the second line is only
 valid for HAVE_GCC=12 architectures.
 
 You need to examine the conditions in the last field more closely before
 calling a line a duplicate.
 
 Martin