From: SourceForge.net <no...@so...> - 2004-04-30 15:03:24
|
Bugs item #736577, was opened at 2003-05-12 19:51 Message generated for change (Comment added) made by broeker You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104664&aid=736577&group_id=4664 Category: None Group: None >Status: Closed >Resolution: Out of Date Priority: 5 Submitted By: Henry Gessau (henry44) Assigned to: Hans-Bernhard Broeker (broeker) Summary: segmentation fault on rebuild Initial Comment: cscope 15.4 on Solaris 2.6 Compiled with gcc 2.95.3 Database is large (148MB) from 17,000 source files. Building database from scratch works fine. Rebuilding after changing a few source files generates a segmentation fault. Stack appears corrupted. (gdb) set args -q -b (gdb) run Starting program: ~/bin/cscope-15.4 -q -b Program received signal SIGSEGV, Segmentation fault. 0x19e7c in build () (gdb) bt #0 0x19e7c in build () Cannot access memory at address 0x6d61649d (gdb) ---------------------------------------------------------------------- >Comment By: Hans-Bernhard Broeker (broeker) Date: 2004-04-30 17:03 Message: Logged In: YES user_id=27517 Lots of questions, no answers in 11 months ;-( I'm closing this as out of date because it concerns a non-current version. ---------------------------------------------------------------------- Comment By: Hans-Bernhard Broeker (broeker) Date: 2003-06-07 13:59 Message: Logged In: YES user_id=27517 That's not exactly very much information to go on. I'll need your help to debug this. Sorry if the number of questions I'm about is a bit overwhelming, but until some of those are answered, we'll be stumbling around in imprenetrable fog. The stack address you see, interpreted as ASCII, reads 'mad' or 'mad', plus some non-ASCII character. Does that word ring any bell? Does this happen only in -q mode, or also without it? What happens if you rebuild with no sources changed? Does it make any difference *which* of those files you edited? Is the cscope.files known for sure to be clean (no non- existing files, no broken links, ...)? Does this happen in -c mode, too? Can you set a breakpoint in build() and see how far, approximately, it gets? The output of gdb command "info locals" after the crash should be helpful, too. Your cscope seems built without debug information --- change that before you re-try in the debugger, so you'll have more precise information where in build() the problem happened. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=104664&aid=736577&group_id=4664 |