meltdown@MIPS2.PHY.QUEENSU.CA (Robert Heaton) (03/23/91)
I have an encountered an interesting problem in a program using the latest version of SAS/C. The program compiles with no errors, and runs perfectly under 5.10, but crashes under 5.10a. An investigation of the binary using CPR reveals that argc is taking a value of zero; other weirdness includes CPR reporting a valid file pointer (ie a file it actually writes to) as being NULL. The real weirdness is that when SAS technical support tried compiling the code, it worked fine! I have localized my problem down to the translation of the quad file by lc2 into object code; when I replace lc2 5.10a with lc2 5.10, the problem goes away. And yes, SAS did try compiling with a copy of my lc2 (SAS tech has been very helpful, and probably thinks I'm suffering from delusions). What I suspect is that the problem may be hardware dependent. I'm using and A1000 with a Supra SCSI and a Prog. Perifs. 2Mb expansion card. What I would like is to find some volunteers to run the code on a similar system and see if they get the same results. As well, if anyone out there is interested in pulling apart some binary, please feel free to offer. BTW, please Email responses directly, as I usually get C.S.A.* about a week after the posting date. Thanx in advance, Rob