[1647] in Athena Bugs
xmath bugs -- more detail
daemon@ATHENA.MIT.EDU (custer@ATHENA.MIT.EDU)
Fri Jan 13 14:44:13 1989
From: <custer@ATHENA.MIT.EDU>
Date: Fri, 13 Jan 89 14:43:05 EST
To: henry@ATHENA.MIT.EDU, bugs@ATHENA.MIT.EDU
Cc: custer@ATHENA.MIT.EDU
My problems occur on vaxes both VaxII and VS2000s.
I get several problems.
1) floating point error
I get this when clicking on the squares that expand and
contract the range of the axis. The clicks work almost all of the time. I
don't know what I am doing differently when it breaks. I doubt that I
am getting too big or too small as I have played with the ranges well
beyond the bounds of the number I am interested in. When it breaks,
you lose your graphs.
1.5) I think I also get the floating point error when I am burying
xmath windows with the <compose-charachter> left or right mouse.
2) segmentation violation
I get this when I am trying to print out graphs. I type the
printer name (usually electra) and it claims that it is postscripting
and spooling and then the xmath windows disappear. I thought it might
be that I had typed the printer name incorrectly, but it seems to
handle typos ok (mumlbe: unknown printer). The root of the problem
may be that the net has been flakey, and that will make this bug hard
to fix.
3) well, in trying to make xmath break in one of the above ways, I got
another bug that I have experienced before. I played with the sizing.
I plotted a datafile. I ran through plot style to get lines through
my points. I set up my axis so all my data would really get graphed
and clicked on the OK square. Where the xmath> prompt was began
giving the error
'expecting number, not array ^ '
and would not stop, filling up my screen and slowing down all my
processes. ^c would not stop it. I could ^z and then kill the job.
When I try to duplicate the error, I can not make it break.
4) There is a strange thing that happens (does not happen) when I try
to use xmath's shell command to change directories. I enclose a copy
from an xmath session:
________________________________________________________________
xmath> help shell
reading help file... done.
File: gnuplot.info Node: shell, Prev: setshow, Up: top, Next: startup
shell
=====
A single shell command may be spawned by preceding it with the !
character at the beginning of a command line. Control will return
immediately to GNUPLOT after this command is executed. For example,
! ls
prints a directory listing and then returns to GNUPLOT.
xmath> ! cd reps/4coms
!
xmath> !pwd
/mit/custer
!
xmath> ! cd /mit/custer/reps/4coms
!
xmath> ! pwd
/mit/custer
!
xmath> !cd /mit/custer/pres
sh: /mit/custer/pres: bad directory
system() failed
(Error 0)
xmath> !cd /mit/custer/reps
!
xmath> !ls
21730 comments food print.evals suspects
21777 coop grab print.evals2 suspects~
21780s core history print.evals2~ tt
Mail dribbles jpbu print.evals~ turb
RMAIL dumps manual print.kirsh turbabs
RMAIL~ e menu print.kirsh2 video
annots eload motion print.kirsh2~ words
art eload~ oldrmail reps ww
bozo emacsload p rs1home z
c emacsload~ papers rsuserhome zombii
clasproc foo postprev stack
!
xmath> !pwd
/mit/custer
!
x
________________________________________________________________
You can't change directories. This bug is just a pain. It means that
you have to type an entire directory path name if your data files are
spread out in different directories. (I'd like to see a line editor
with a history command (like ! in unix or ^y in emacs) so that you
don't have to type
plot wind 4 [dah hum] something or other "/mit/mumble/mumble/mumble"
all evening.)
So) I can't make xmath break in any of these modes when I try to break
it. Using xmath is like walking around waiting to fall in a tiger
trap. I think I have a very high tolerance for this breaking business
-- I just curse at athena, send out flaming bug reports, and make sure
I print graphs out every 30 seconds so that I don't loose too much
when my windows go hyperspace. I can't imagine that very many other
people will derive such satisfaction from this experience. I
don't think that athena will win any friends with xmath in its current
state.
><
