Re: 9.1 xdpyinfo oddity

>> [...] right up until it's about to start printing things.  It does

>>    507      1 xdpyinfo CALL  __fstat50(1,0x7f7fffbd4de0)

> It sounds to me as if xdpyinfo is checking to see if stdout is a tty
> and is doing something wrong if it isn't.  Doesn't sound like it has
> anything to do with the server at all.

That would be my own guess too.  The only reason I sent it to tech-x11
is that the only thing I've seen manifest this behaviour is an X
program.

Looking at the only .c file in 9.1's sources whose name contains
"xdpyinfo", xsrc/external/mit/xdpyinfo/dist/xdpyinfo.c, at a first
quick glance I see nothing that would explain this.  The only instance
of "tty" in the entire file is coincidental

      printf  ("\ttype:         %s\n",

and the only occurrences of "stat" are all substrings of "static"
(feeding it to sed -e 's/static/st@tic/g' | fgrep stat produces
nothing).  Looking at the control flow, I don't see anything glaring.

I may go digging more at some point, but at the moment I'm not strongly
inclined to; I doubt work is interested in putting my time into chasing
this oddity, and I'm not personally interested in going grubbing around
in 9.1's sources unless I'm being paid to.

I was a bit curious whether anyone else saw the same oddity,
secondarily to throw out a heads up in case someone with more interest
in modern NetBSD than I have wanted to investigate, and tertiarily in
case someone else trips over the same problem.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B