iraf.net

>From rebull@boron.uchicago.edu Wed Feb 4 08:56:23 1998
From: Luisa Rebull <rebull@boron.uchicago.edu>
To: fitz@noao.edu
Subject: Re: latest version of xgterm?Hi Mike --Thanks for your prompt note.> > noticed that despite the fact that both taupe/rust (at the 0.9m)
> > and my machine (boron) are running SunOS 4.1.4 and IRAF 2.11,
> > the xgterm that you all had running there is subtly different than
> > the one that I have. [...]
> > plot which mine does. Is there another "latest and greatest" version?
> > Or is there a flag I'm not setting somewhere?
>
> This I don't understand at all. If you are running an xgterm with
> V2.11 and did and "stty xgterm" you should see the color plot and have a
> real status line. For completeness, what I get on my system (and what I thought was
normal) is a black window with white writing and red crosshairs. The
additional things I noticed at KPNO which were IMHO improvements
included the status line (in yellow at the bottom of the graph screen);
yellow tick marks (when running the task identify); the fact that when
deleting points from a fit, the red crosshairs would automatically move
to the next point; and the ability to be slightly more robust when I
get impatient and hit a key before the graphics terminal is 100% ready
to accept the input. These changes were subtle enough that I assumed
it was a version change, as I mentioned, as opposed to mistakes in
installation which I figured would have a much more severe effect (e.g.
core dumps).Yesterday I went to your site and noticed that the last time the files
for x11iraf were altered were August; since I installed 2.11 in
December, I figured I had the current version, but to confirm, I pulled
over the tar file, and for lack of a more elegant solution, simply
"diff"ed the two xgterm executables (the one I had just downloaded and
the one I had running on my system). They are indeed identical.I am really running an xgterm -- I'm really pointing to the executables
I grabbed from your website, it looks & behaves differently than an
xterm, and since I'm getting something different than a usual tektronix
window, I assumed I had installed it correctly.I had thought at one point that xgterm had dropped, for example, the
status line because it couldn't grab the right colors from the color
table since Netscape had hogged them; I tried with Netscape -install
and then no Netscape running at all. It didn't help. Color problems
don't seem likely to solve the crosshairs moving to the next point, or
being more robust when accepting input.> Are you perhaps using a private graphcap file which may
> not define xgterm entries? Do a "show graphcap" to see, if it says anything
> other than dev$graphcap No, I am not using a private graphcap file; 'show graphcap' does indeed
return dev$graphcap. However, since I'm my own system administrator, I
wondered this morning if I had corrupted the dev$graphcap file. I did
add some things to my system's graphcap file to allow me to print to my
department's printer, but I would assume that it shouldn't affect
xgterm; I would be happy to send you the additional lines, but they're
all printer-related. The rest of the file is that which came with the
distribution; I didn't touch the xgterm entries.> > Periodically, when my routines try to plot to the xgterm window, I get
> > this error:
> > ERROR: task `buffer' not found
> > dohydra ()
>
> This is another stumper, 'buffer' is a parameter not a real task.Oh, yea, I know; when this first happened, this was a source of great
frustration, since 'help buffer' didn't work!> I suppose that some disk full problem may cause a memory error but this
> hasn't been reported before. Hm, that's an intersting idea. However, It doesn't seem to be behaving
like a memory error; as I said, I concluded it was because it was
trying to work with a partially written image file. I've appended
below my notes to local gurus trying to solve the problem when it first
arose.> Does this
> always happen or just with the disk space problem? Well, it used to happen just when I run up against disk space issues.
But yesterday, when trying some of the stty xgtermc things, I got it
again several times, and I'm floating in disk space at the moment.
Admittedly, it wasn't in a particularly reproducible fashion. But once
it happens, the only way I can reliably fix it is to grab the image
from the backup, reprocess it, and try again... unlearning tasks
doesn't help, although admittedly I haven't tried mkiraf again.thanks for any suggestions,
Luisa
-------------------Original task `buffer' email excerpts:>From rebull Mon Dec 8 10:29:41 1997HI John --I got Iraf 2.11 working last week, on Wed or Thursday. I've been
working on the data I need to process in 2.11 on Thursday and Friday
of last week. On Friday afternoon, I ran out of disk space, and thus
the routine died in the middle. This hasn't been a problem before --
the routine, although it crashes ingloriously, is good at picking up
where it left off, and I've done this successfully on Friday.Now, it crashes when it tries to plot things:
ERROR: task `buffer' not found
dohydra ()There's no such thing as "task `buffer'". I'm using xgterm (recommended by IRAF people), last week and now.I've restarted Iraf. I've restarted the xgterm plotting window.
I've restarted the xgterm in general. I've restarted the openwin
session. And now I've even rebooted boron. Nothing fixes this problem.
The first thing I can think of is that something got changed over
the weekend, but that should only be LAN connections, right, nothing
on oddjob changed, right?-- later --Yes, I'm using things entirely locally; I'm not pointing to any aspect
of oddjob's 2.10.x installation. I'm using the new login.cl and
everything; as I said, it was all working last week...-- later --really, i swear, it ran without any errors last week, and i didn't even
log out over the weekend. when i got in this morning, i tried to
plot, picking up exactly, literally where i left off, and it turned me
over to the xgterm plot window without an error, but without plotting
anything either. the first restarting prompted the error. I get a
blank xgterm plot or tek window, and then the error shows up, so it's
finding a graphcap or termcap entry successfully, but then can't find
the buffer into which to write...-- later still --I grabbed the offending image files again from my backup directory,
reprocessed them, and tried again. This time it worked. The file
must have been corrupted somehow when it ran out of disk space. ??
>From fitz Wed Feb 4 21:38:14 1998
To: rebull@boron.uchicago.edu
Subject: Re: latest version of xgterm?Hi Luisa,
From your mail I'm convinced you're running an xgterm binary, but
the B&W graphics and the bug with hitting a keystroke before the cursor is
ready, the overwriting of status output, etc is old xterm behavior so I'm
still confused. In connection with the DOHYDRA problems I'm thinking there
may be some kind of installation problem. I also think the fastest way to
resolve this would be for me to log into your machine through the iraf
account to check things out, I could verify IRAF is installed correctly
and run xgterm to display on my machine to see if it works as expected.
I could also check your resource definitions and the installed app-defaults
files to see if something here is causing a problem. If this is agreeable
you could change the iraf password while I poke around, please let me know.Cheers,
-Mike