Re: toolchain/38487 (build.sh install breaks if -D and -O dirs are RO)

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

The following reply was made to PR toolchain/38487; it has been noted by GNATS.

From: Mouse <mouse%Rodents-Montreal.ORG@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: toolchain/38487 (build.sh install breaks if -D and -O dirs are RO)
Date: Sat, 3 Jun 2023 17:48:36 -0400 (EDT)

 > Are you able to reproduce this in a build of -current?
 
 I have not tried.  It has been at least a decade since I tried to build
 -current (specifically, since core let GPLv3 code into the tree).  I
 have built 8.* and 9.* a few times for work, but I don't think I've
 ever tried to install from -D and -O directories which are mounted RO.
 I don't think I've ever tried this since I sent the PR, not even with
 5.2.
 
 That said, if it's not reproducible in -current by doing the analog of
 what I sketch in the PR (do a build.sh -D ... -O ... -x -U build, then,
 with the -D and -O directories on a read-only filesystem, try to
 install into somewhere RW), it can probably be considered fixed.
 
 This strikes me as a practical issue more because it would likely break
 multiple parallel installs from shared -D and -O build directories than
 because it breaks installs with a filesystem mounted RO.
 
 /~\ The ASCII				  Mouse
 \ / Ribbon Campaign
  X  Against HTML		mouse%rodents-montreal.org@localhost
 / \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B