home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.gcc.help
- Path: sparky!uunet!clsi!mitch
- From: mitch@clsi.COM (Mitchell N. Perilstein)
- Subject: Need advice debugging 2.1
- Message-ID: <MITCH.92Dec18160719@pollux.clsi>
- Sender: usenet@clsi.COM
- Organization: CAD Language Systems, Inc.
- Distribution: gnu
- Date: 18 Dec 92 16:07:19
- Lines: 34
-
- I'm giving all the following information not as a bug report but as a
- request in hopes that someone can suggest a debugging technique that
- will result in understanding the cause of the crash. Once understood, I
- will submit a report.
-
- I'm trying to debug gcc2.1 on a Sun 4/110 running SunOS 4.1.2. The error is
-
- gcc: Internal compiler error: program cc1 got fatal signal 4
-
- while compiling an 11meg C file. (It's automatically generated C for
- simulation which needs to be that big. Anyway, it's a good torture test
- for gcc.)
-
- cc1 runs for a long time. Watching the sizes of the /tmp/cca files, the
- .i appears very quickly, then the .s grows steadily in size, then stops
- growing. Things sit for about fifteen minutes in this state, with cc1
- in `D' state and and gcc in `IW'. Finally, the message appears and it
- leaves a 45 meg core. gdb4.6 is confused and prints only one stack frame
- when examining the core.
-
- (gdb) i st
- #0 0xf76e3e54 in bzero ()
-
- It works the same way when run under gdb4.6 by attach'ing to cc1.
- Resourcewise, there is 50m of swap free at the lowest point, and 80m
- /tmp.
-
- If anyone thinks upgrading to more recent gcc would change anything, I
- will do so.
-
- TIA,
- ---
- Mitchell N. Perilstein CAD Language Systems, Inc. 410-992-5700 x225
- Member, League for Programming Freedom. Mail lpf@uunet.uu.net for info.
-