"VAXINO::DAVID"@SCRI1.SCRI.FSU.EDU (08/02/89)
I am using dbx and edge to debug some rather large fortran programs (~20k lines) on a 3130. I have been able to get dbx to die several times with the following error: dbx: internal error: skimsource: too many lines This error has occurred with both the List command and when the debugger attempts to display the source code of a break point. I have also been able to get the debugger to die in such a way that it actually HANGS the machine! This error normally occurs while the debugger is loading in the symbolic information but before it prints "(dbx)". These errors both seem to go away when I run my code through egrep just before compilation to remove of the comment lines (code size ~10k lines). A third error that I have come across is not being able to print out the contents of arrays that are passed in to a subroutine. The error is some gibberish about null pointers. The only way that I have found around this error is to reference the variables from the calling procedure. For example: print main.var1[1] - David LaSalle SCRI Florida State University Tallahassee, FL 32306-4052 (904)644-8532 arpanet: minuit@scri1.scri.fsu.edu (128.186.4.1) or minuit%fsu@nmfecc.arpa
blbates@AERO4.LARC.NASA.GOV ("Brent L. Bates AAD/TAB MS294 x42854") (08/06/89)
I have noticed that for most of my work, dbx works most of the time on our 3130 as long as I make a call to the forground subroutine when using graphics. However, I have had some similar problems to yours. I don't use edge, because it doesn't work on a 3130. Just on of the many things you give up on, because the 3000's are orphaned machines. -- Brent L. Bates NASA-Langley Research Center M.S. 294 Hampton, Virginia 23665-5225 (804) 864-2854 E-mail: blbates@aero4.larc.nasa.gov or blbates@aero2.larc.nasa.gov