chapman@lll-lcc.UUCP (Carol Chapman) (05/11/89)
I'm writing a program that displays digitized color images using X11R3, Xt, and Athena widgets. It was working fine using SUN O/S 3.5. We just upgraded to O/S 4.0.1 and now when I run my code the core widgets that hold each image appear on the screen, but they're blank! When I run the code using the Saber-C debugger, it runs fine, except for the blank images. When I don't use the debugger, the program is bombing in what appears to be random places involved with the image creation, usually at XPutImage. I don't think I changed anything relevant in my code, but at this point I may be losing my mind... Any ideas? carol -- ------------------------------------------------------------------------------- Livermore Lab (chapman@lll-crg.llnl.gov or chapman@lll-lcc.llnl.gov) P.O. Box 808, L-153 Tel. (415) 423-7876 Livermore, CA 94550 "Never own anything that you have to FEED or PAINT."
meo@stiatl.UUCP (Miles O'Neal) (05/12/89)
In article <2471@lll-lcc.UUCP> you write: | I'm writing a program that displays digitized color images using | X11R3, Xt, and Athena widgets. It was working fine using SUN O/S 3.5. | We just upgraded to O/S 4.0.1 and now when I run my code | the core widgets that hold each image appear on the screen, but | they're blank! When I run the code using the Saber-C debugger, it | runs fine, except for the blank images. When I don't use the | debugger, the program is bombing in what appears to be random places | involved with the image creation, usually at XPutImage. I don't know for sure if this is it, but will tell you about our experience. We run X11R2 on a Sun 386i, SunOS 4.0. We also use Xt and Xaw. We had our application about half done, featurewise, but mostly debugged as far as we had gone. Between the Sun server and our own lack of Xperience, we had some memory leaks, but nothing major. We upgraded to 4.0.1, and chaos became our project manager. Both the server and our application grew at ridiculous rates, at least two orders of magnitude faster than under 4.0. We got core dumps where before we had none. Events disappeared all too often (both input and output). Strange behavior was the order of the day. The typical send-too-many-of-the-same-key problem was MUCH worse. Needless to say, we backed down to 4.0 really quickly. The BIG problem with doing that (there are several others) is that, at least on the 386i, the Sun installation INSISTS on completely reformatting the disk! In the process, one of our backup tapes lost its mind, and we lost ANOTHER week's worth of work. We now (a) have redundant backups, and (b) refuse to upgrade until we are satisfied that the upgrade is worthwhile. I HATE being on the trailing edge of the leading edge! Reading comp.systems.sun is a must to keep up with this sort of thing, even though there is a LOT of stuff in there we never need. -Miles