Brett asked me to pass this by the list so that it will be archived at Flora. The more [NetBSD] people who know about it, the better. Anyway, it's fixed in the latest NetBSD pkgsrc. (Lynx's ifdef's are too much.)
__Henry > Date: Tue, 20 Aug 2002 21:59:47 +0930 > From: Brett Lymn lt blymn mungedat baesystems mungeddot com mungeddot au gt > Subject: Re: Lynx Bug? [...] > Guys, > One of the other NetBSD developers may have come up with a gem > on this. In LYCurses.c, endwin() is called conditionally on the value > of LYscreen. At the top of LYCurses.c there is a block of ifdefs > starting at line 753. In our case we are not VMS but the default > package build in NetBSD also does not use slang which puts us at line > 758. Here is the bit that causes the problem, older versions of > NetBSD curses does not have newterm, so LYscreen gets set to NULL. If > you skip down to line 911 you will see an ifdef NEWTERM which will > fail, the else clause is at 944 and is a bare initscr(). This means > that LYscreen has been set to NULL and never changed so when lynx is > exited the endwin() is not called hence the tty settings are not > restored. I hope that makes sense. > > I will generate a patch for the NetBSD pkg system that should fix this > problem. > > This is not a problem with the later NetBSD versions because newterm > was added a while back. Just as an aside, the NetBSD version of > curses now has a resizeterm (has for sometime now) so you may want to > look at your ifdefs to make sure you are handling a non-ncurses > library having resizeterm :-) > > -- > Brett Lymn > ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED]
