🚀 go-pugleaf

tph@longhorn.uucp (Tom Harrington) wrote:

>Billy Chambless (bi...@cast.msstate.edu) wrote:
>
>: Another guy comes to my office and says: "I keep getting this
>: "Out of Memory Error" .... I tried delting a buncha files, but it didn't
>: seem to help much....
>
>I've actually used systems where that might have been a valid approach.
>This was on an old LTX analog IC test system (the model "DX/90" rings a
>bell, but I could be wrong); I think it was based on some ancient Control 
>Data hardware, but it's been a few years.  Anyway, once the hard drive 
>filled up, most common commands became disabled, giving "disk full" 
>errors.  When this happened, you couldn't do *anything* with the
>system until you removed a few files.  Including, as it happened, 
>deleting files.  So you'd get an error message saying something like 
>"Cannot delete file.  Disk full."  At which point the mind would 
>boggle briefly at what to do next.  We had two of these beasts 
>networked together somehow, and it was a good thing, too, or I 
>don't know how we'd ever have recovered from full hard drives.

Well, this reminds me of a problem I had to diagnose for a highly paid
professional programmer (a bit a sarcasm there).  A program (DOS/370
COBOL) abended with a strange error code.  After finding the appropriate
manual (another IBM quirk, there was no manual with ALL the error codes,
you had to find the master list to find which manual to look in), it was
"command reject" on a disk write.  After the usual flailing around with
register saves, etc., I found that the write was issued from a paragraph
which closed the file, and instructed the operator to dump (what was
that utility, DEBE?) the file.  Tracing further, it came from a WRITE...
ON SIZE ERROR GOTO statement.  So, when there was no more space in the
file, the program attempted to close the file, and the close routine did
not handle the problem gracefully.  I asked the programmer if the error
routine had been tested.  'Of course, I inserted a jump to the routine,
and it worked fine'.  I calmly pointed out that testing an error routine
in the absence of the error didn't really do the job.

>
>--
>Tom Harrington ------- t...@rmii.com ------- http://rainbow.rmii.com/~tph
>   "Welcome to the real world - an interesting place to visit, but I
>                 refuse to live there." -Andy Rozmiarek
>-> Fractal Kit:  http://rainbow.rmii.com/~tph/fractalkit/fractal.html <-

---------------
Larry McKnight
(this space unintentionally left blank.....