Re: [Plplot-devel] Current test_diff_psc results

[Geoffrey F. <geo...@at...>]

Alan W. Irwin writes:
 > On 2010-05-11 12:29-0500 Geoffrey Furnish wrote:
 > 
 > > Geoffrey Furnish writes:
 > > > Alan W. Irwin writes:
 > > >  > C++ and OCaml are perfect, and here is the status of the remaining
 > > >  > non-perfect bindings:
 > > >  >
 > > >  > tcl
 > > >  >    Missing examples            :
 > > >  >    Differing postscript output :  19 21 28
 > > >  >    Missing stdout              :
 > > >  >    Differing stdout            :  21
 > > >
 > > > I believe I've fixed Tcl #19.  Let me know if anyone finds otherwise.
 > >
 > > BTW, the only difference I see in example 21 postscript output, is the
 > > time stamp:
 > 
 > As I said earlier in this thread my x21 result for Tcl is an artifact of
 > not having Tcl8.5 which also explains the different stdout in my case.
 > That message was
 > 
 > This example require Tcl 8.5 or later: use of NaNs

I've taken a look into Tcl example 21, and I'm not very convinced.  What I
find is that when I comment out the return which currently bails if you don't
have Tcl 8.5, well, everything works fine, as far as I can tell.

In particular, the isnan proc appears to me to have been coded in a way that
anticipates the non-recognition of the NAN construct by Tcl.

However, peering into "svn blame" and "svn log", it seems to me that the
lines in questsion were authored by Arjen, in August 2008.  So I think it's
really up to Arjen to render the final verdict on this.

My only input is that I tried running the example on Tcl 8.4, and don't see a
problem.  It would seem to me we could drop the requirement for 8.5.

-Geoff