[Xastir-dev] Configure's test for "gv"

[Dan Brown]

On Fri, 12 Jan 2007, Tom Russo wrote:

> On Fri, Jan 12, 2007 at 01:08:05PM -0800, we recorded a bogon-computron collision of the <archer at eskimo.com> flavor, containing:
> > 
> > KD7NM brought up to me today that the configure test for "gv"
> > complains about X11 if there's no X11 running on Cygwin at the time.
> > I believe Tom mentioned this or a similar thing to me before as
> > well.
> > 
> > What would be a good direction to go in this case?  Should we test
> > again for just the presence of the file if it fails the first test?
> > I'd hate to toss out the print customization options just because
> > someone configured without X11 running.
> 
> Well, the main problem is that we need to actually *run* gv to see what 
> version it is in order to determine what sort of options to pass to it (they
> changed them dramatically a couple years ago, and old options are no longer
> recognized).  And "gv --version" tries to connect to the X server even though
> it has no intention of putting up any windows.
> 
> Unless we're going to drop support for the old style GV options (potentially
> shutting some folks out who haven't upgraded in years), we'll need to run
> gv --version in order to figure the version out.  You could, of course,
> always do something like "test for existence of gv, then run it to test 
> version, and if it doesn't run assume it's the new version" or something like 
> that.  


I've had to jump through stupid hoops like this to get silent installs
of <cough> "certain unamed COTS products" to work from a command line. 

fwiw: 

crozet:~$ strings `which gv` |grep GV.version: 
GV.version:                             gv 3.5.8

--
Dan Brown 
brown at brauhaus.org