Re: pkg/56936

["David H. Gutteridge via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR pkg/56936; it has been noted by GNATS.

From: "David H. Gutteridge" <david%gutteridge.ca@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: pkg/56936
Date: Mon, 07 Oct 2024 19:25:12 -0400

 On 2024-10-05 01:30, Mayuresh via gnats wrote:
 > The following reply was made to PR pkg/56936; it has been noted by 
 > GNATS.
 > 
 > From: Mayuresh <mayuresh%acm.org@localhost>
 > To: gnats-bugs%NetBSD.org@localhost
 > Cc:
 > Subject: Re: pkg/56936
 > Date: Sat, 5 Oct 2024 10:58:31 +0530
 > 
 > On Sat, Oct 05, 2024 at 09:59:50AM +0530, Mayuresh wrote:
 > 
 > Further I confirm that setting 'bl_cv_pty=streams' in configure script
 > still solves the problem.
 > 
 > Just that it's not clear to me what is a clean way to do it. For above
 > trial I just replaced bl_cv_pty=bsd with bl_cv_pty=streams at 3 places 
 > in
 > the configure script.
 
 The upstream project modified a configure.in definition in the 3.9.3
 release here:
 
 https://github.com/arakiken/mlterm/commit/abe3282fc7b6952450cf6000222a19729f3cb7cc
 
 In summary, they removed this:
 
 -AC_CACHE_VAL(bl_cv_pty,
 [etc.]
 
 This means they no longer allow builders to override the value of
 bl_cv_pty as an argument to configure. I'm not sure if that was done
 deliberately, or just something cut for simplicity's sake when other
 changes were being made. Later on, they also moved the file location of
 this check itself, in its newer form.
 
 (Again, another fix would be simply to populate a pbulk sandbox with
 the expected pseudo-terminal devices so the runtime check succeeds.)
 
 Dave