Re: pkg/32991

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

The following reply was made to PR pkg/32991; it has been noted by GNATS.

From: David Holland <dholland-pbugs%netbsd.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: pkg/32991
Date: Mon, 8 Aug 2011 04:02:49 +0000

 On Thu, Jul 21, 2011 at 12:10:06PM +0000, Robert Elz wrote:
  >    |  This sort of problem can also be caused by updating a lower-layer
  >    |  library and recompiling the game, but *not* recompiling some
  >    |  intermediate library package it depends on that's also linked to the
  >    |  same lower-layer library.
  >  
  >  Yes, I know, but I can't quite see how that could have happened here.
  >  When it first occurred it was an obvious possibility, but in the
  >  intervening period I don't believe there's anything that hasn't been
  >  recompiled - including at least one occasion where I deleted everything
  >  (which includes everything X related of course, as I am using modular x-org
  >  on NetBSD 4) and recompiled the lot.   Yet this problem persists.
 
 Hmm. Odd...
 
  >    |  I'm going to mark this PR netbsd-4 only, so if you don't get around to
  >    |  it before -6 comes out we can just let it die quietly.
  >  
  >  That's fine, for whatever reason, maybe thread related (???) it only
  >  seems to be a problem there.   [Aside: I have no idea at the minute if
  >  there's anything linked with the tread libraries in this game - that's
  >  just pure guesswork of one obvious difference between NetBSD 4 & later.]
 
 Yeah, that's the first thing I thought of too. The backtrace suggests
 no threads but instead something OpenGL-related, but that doesn't make
 much sense.
 
 On the other hand, the backtrace also asserts that libglut was being
 called (and then calling into) libstdc++, which doesn't make any sense
 at all. Maybe that's an indication that the problem is libstdc++? Or
 maybe that's just gdb being insane.
 
  >  If anyone else is seeing this problem on NetBSD 4, or can confirm it, then
  >  just marking the package as broken on NetBSD 4 and closing the PR would
  >  also be satisfactory, NetBSD 4, while still supposed to be supported, is
  >  getting so ancient now that it is unlikely there are many people left who
  >  care - and for those who do, some random game not working is very unlikely
  >  to be a serious issue.
 
 Yeah.
 
 Well, I'd probably look into it if I still had a netbsd-4 machine, but
 I don't...
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost