Re: bin/53548 (here document of /bin/sh of today is broken)

[Robert Elz <kre%munnari.OZ.AU@localhost>]

The following reply was made to PR bin/53548; it has been noted by GNATS.

From: Robert Elz <kre%munnari.OZ.AU@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: bin/53548 (here document of /bin/sh of today is broken)
Date: Wed, 23 Oct 2019 15:14:59 +0700

     Date:        Wed, 23 Oct 2019 01:05:01 +0000 (UTC)
     From:        coypu%sdf.org@localhost
     Message-ID:  <20191023010501.909667A24A%mollari.NetBSD.org@localhost>
 
   |  > If/When the primary change gets pulled up to -8,
   |  
   |  I expect this is getting less likely with netbsd-9 just out the door :-)
   |  -8 is probably about to see mostly security/stability fixes.
 
 Yes, I started several times to make a patch for the changes (the
 "primary change", ie: PR bin/48875) for -8, but never succeeded before
 abandoning the attempt (there have been many structural changes to the
 shell, most unrelated to the changes in question, which means making the
 patch is all essentially repeating the original work, while hoping to
 both do it correctly the old way, and not miss anything - including extra
 mods that are no longer needed in -9).
 
 Since the -8 shell as it is now is correct (wrt this anyway -- just slightly
 surprising in this one aspect) I decided to give up, and just allow that
 shell version to eventually expire when -10 gets released.
 
 I should have closed the relevant PRs (including this one, which was
 for a bug introduced in the fix for 48875, and so irrelevant to -8 as
 it is now) a while ago.   I will get around to that soon.
 
 Of course, anyone else who desires is welcome to attempt to take the
 fixes from the -9 sh (or HEAD, they are essentially the same) and
 apply them to the -8 sh.   It is also possible to take the entire -9
 sh and build it with -8, as I remember with almost no changes, perhaps
 none (though the division of labour between the built in kill, and sh,
 wrt printing signal names with kill -l and trap -l, which do the same
 thing, more or less, and so use the same routine, might have altered,
 and if so, requires a change from #if 0 to #if 1 or vice versa, I forget,
 in one place...)
 
 kre
 
 ps: even after -9 is released, -8 should still get more than just
 security fixes.   If any real bugs (that actually matter) to its sh
 are reported, I will do my best to fix them.