Minor setvideo() bug; SGI bug reporting mechanisms?

Stuart Levy slevy at poincare.geom.umn.edu
Tue Oct 16 09:34:56 AEST 1990


I'm writing to report a minor bug, seemingly too minor and too easily worked
around to be worth bothering SGI customer support, but worth fixing some day:

    Since SGI 3.3, the setvideo() call (on a 4D-240/GTX with genlock card):
	setvideo(DE_R1, DER1_RESET);
    has elicited the message:
	setvideo: unknown video type, 0

I'd consider this a harmless change from previous versions,
EXCEPT that the message is written to standard output.
If library routines want to report errors, they should use stderr!

[This hosed an application that communicated with a child graphics program
 through a pair of pipes connected to the child's stdin/stdout.
 Whenever the child switched the display to/from NTSC mode, this
 message would be mixed with its output and confound the communication,
 but the effects wouldn't appear until some while later.
 Clearly this bug is easy to avoid, but it took lots of hair-pulling
 before the cause of the application's bizarre behavior was discovered.]


Now the other part of the question:  Is there a natural mechanism for
reporting SGI bugs (or requesting features) that's less intrusive than
calling customer support?  I'm thinking of cases like this one,
where the item is low priority and we really don't need a response.

Sun has an e-mail address for this kind of thing:  sunbugs at sun.com.
Does SGI have something similar?  If not, is posting to this newsgroup
sufficient notice, or is it important to register a call to the hot line?
In the latter case, do people not buying software support have any way to
report bugs?

    Stuart Levy, Geometry Group, University of Minnesota
    slevy at geom.umn.edu



More information about the Comp.sys.sgi mailing list