Segfaults in MIT libkrb5

[Fredrik Tolf]

On Sun, 2006-12-17 at 00:43 -0500, Theodore Tso wrote:
> If you are willing to recompile e2fsprogs (which is the source of the
> system com_err library), please try applying this patch and try
> running your program (without the LD_PRELOAD hack) with the
> environment variable COMERR_DEBUG set to the value 1 and let us know
> what you see.  That should help us see if the remove_error_table is
> getting called, and with what arguments.

I got around to doing it, but the results have me a bit perplexed. It
seems there is no adding or removing of error tables at all during
dynamic loading. It is especially strange since the program doesn't link
at all against libkrb4 until the PAM module is loaded, at which point
its error tables would have to be loaded. ldd verifies that libkrb4
properly links against libcom_err. Any ideas?

By the way, if you use that patch yourself for whatever reason, you may
want to add a PID field for each line of debug output. It had me
speechless for a while with what seemed like all error tables were
getting removed multiple times, until I realized it was an NSS module
that was forking. :)

Fredrik Tolf