[comp.sys.sgi] More /tmp woes

SOFPJF@VM.UOGUELPH.CA (Peter Jaspers-Fayer) (06/05/91)

I know this has already been discussd (with FORTRAN), but after having
sweated this one out for about an hour, I'd like to relate this war story
in case someone else should be so unlucky (and in case someone at SGI
is "listening"):

I have this C program that writes some messy PostScript.  I was testing
it, found a bug, removed a "printf", and attempted to re-compile:

ugen: internal  : line 82 : build.p, line 1740
                unexpected u-code

it sez.  Whattheheckizzat? Say WHAT?  Line 82 is no where near my change.
Quick, put that printf back! To no avail.  In trying to "vi" the .PS file
I spooted this strange error message which started me on the right track.
A long while later, I finally found in SYSLOG:

unix: dks0d1s0 (/): Out of space

Similar messages appeared in my (iconified - tsk) console window.

Seems psview left some crud in /tmp, and so...


1) The above compiler message is not too informative (Hello SGI?)
2) Is there a  way to "open" and "pop" a hidden or iconified
   wsh window whenever there is any output to it?

 /PJ                                __o
                                   -\<,
                             ......O/ O             SofPJF@VM.UoGuelph.Ca

davea@quasar.wpd.sgi.com (David B.Anderson) (06/06/91)

In article <9106051132.aa28547@VGR.BRL.MIL> SOFPJF@VM.UOGUELPH.CA (Peter Jaspers-Fayer) writes:
[stuff deleted from various places to shorten the reply]
>I have this C program that writes some messy PostScript.  I was testing
>it, found a bug, removed a "printf", and attempted to re-compile:
>
>ugen: internal  : line 82 : build.p, line 1740
>                unexpected u-code
>A long while later, I finally found in SYSLOG:
>unix: dks0d1s0 (/): Out of space

Two problems with the compiler back-end tools (ugen, as1, uopt):
1) They never report file i/o problems helpfully (as the example above
   demonstrates).
2) Problem reports tend to be a confusing mixture of information
   for the user and for the tool-maintainers.

In 4.0 the ugen messages have been made slightly clearer (in many cases)
but no one had time to do the rather large amount of work that would be
needed to really *fix* the problems.  Sorry.

We will fix these problems in some future release.

[ David B. Anderson  Silicon Graphics  (415)335-1548  davea@sgi.com ]
[``What could possibly go wrong?''                      --Calvin
 `` Whenever you ask that, my tail gets all bushy.''    --Hobbes    ]