[racket-dev] online check syntax: now disabled by default

From: Casey Klein (clklein at eecs.northwestern.edu)
Date: Fri Sep 16 15:49:12 EDT 2011

On Thu, Sep 15, 2011 at 8:13 PM, Robby Findler
<robby at eecs.northwestern.edu> wrote:
> Yes, certainly! Core files (or just stack traces) are useful, I expect.
>

I've been seeing a freeze roughly once a day. DrRacket gets
permanently stuck with the OS X rainbow pinwheel as the cursor, and
SIGKILLs won't kill it. I have to hard reboot.

It eventually occurred to me to run DrRacket out of a terminal to look
for interesting output, and sure enough, it's a segfault.

I have a core dump, but I'm not getting a useful stacktrace. I'm not
sure if it's because the stack is corrupt or because I'm using gdb
wrong.

I did this:

$ gdb path/to/drracket /cores/core.221

saw this message:

This GDB was configured as
"i386-apple-darwin"..."/Users/clklein/git/new-redex-semantics/bin/drracket":
not in executable format: File format not recognized

then a bunch of messages like this one:

warning: Could not find object file
"/Users/mflatt/tmp/newest-libs/pixman-0.21.6/pixman/.libs/pixman-access.o"
- no debug information available for "pixman-access.c".

and eventually the gdb prompt. The backtrace command shows the following:

#0  scheme_set_cont_mark (key=Cannot access memory at address 0xbfffb9a0
) at eval.c:993
Cannot access memory at address 0xbfffb99c

Let me know if I can provide any more information.


Posted on the dev mailing list.