Re: pkg/55806: lang/clisp build fails on NetBSD/i386-9.99.75

[David Holland <dholland-pbugs%netbsd.org@localhost>]

The following reply was made to PR pkg/55806; it has been noted by GNATS.

From: David Holland <dholland-pbugs%netbsd.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: pkg/55806: lang/clisp build fails on NetBSD/i386-9.99.75
Date: Fri, 23 Sep 2022 20:46:23 +0000

 On Wed, Sep 21, 2022 at 11:20:01PM +0000, John D. Baker wrote:
  > From: "John D. Baker" <jdbaker%consolidated.net@localhost>
  > To: gnats-bugs%NetBSD.org@localhost
  > Cc: 
  > Subject: Re: pkg/55806: lang/clisp build fails on NetBSD/i386-9.99.75
  > Date: Wed, 21 Sep 2022 17:02:15 -0500 (CDT)
  > 
  >  As a workaround, I made an exception for "lang/clisp" in my global
  >  application of "GCC_REQD+=10" in "/etc/mk.conf" and it builds without
  >  problems using the base system compiler [GCC 7.5.0 (nb4 20200810)].
  >  
  >  Since -current has GCC 10.4 as the base compiler, the problem remains.
 
 I see that it says to add -falign-functions=4 to the makefile but that
 according to the pasted fragment of the build it's already doing that.
 
 So I guess it's going to require digging...
 
 The symbol it's complaining about is defined in "debug.d"; can you
 check in the full build log whether that's built using
 -falign-functions=4 (and also check .work.log to see if the wrappers
 are causing it to get lost for some reason)?
 
 The code is a maze of cpp abuse but I don't see anything that would
 accidentally override that within the source.
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost