[2854] in Kerberos-V5-bugs
krb5-libs/525: memory ccache is pretty borken
daemon@ATHENA.MIT.EDU (danw@MIT.EDU)
Tue Jan 13 14:34:22 1998
Resent-From: gnats@rt-11.MIT.EDU (GNATS Management)
Resent-To: krb5-unassigned@RT-11.MIT.EDU
Resent-Reply-To: krb5-bugs@MIT.EDU, danw@MIT.EDU
Date: Tue, 13 Jan 1998 14:31:44 -0500
From: danw@MIT.EDU
Reply-To: danw@MIT.EDU
To: krb5-bugs@MIT.EDU
>Number: 525
>Category: krb5-libs
>Synopsis: memory ccache is pretty borken
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: krb5-unassigned
>State: open
>Class: sw-bug
>Submitter-Id: unknown
>Arrival-Date: Tue Jan 13 14:32:01 EST 1998
>Last-Modified:
>Originator: Dan Winship
>Organization:
MIT Athena
>Release: 1.0
>Environment:
System: IRIX antharia 6.3 12161207 IP32
>Description:
ccache/memory was copied from ccache/file and it shows:
- mcc_gennew uses mktemp to generate `unique' memory ccache
names, which would seem to guarantee that it keeps returning
the same name over and over again since that filename will
never actually get used.
- mcc_gprin returns KRB5_FCC_NOFILE if there's no default principal
for the cache: while there _are_ less appropriate error codes
than that, it's still pretty awful.
- there are lots of comments referring to files and file-based
ccaches that should be talking about memory
- most of the comments saying what errors a function returns either
claim it can return errors it can't, or claim it doesn't return
errors that it can.
>How-To-Repeat:
>Fix:
The comments are easy enough to fix. I'm not sure what the right
error for mcc_gprin is. mcc_gennew needs some new code to actually
guarantee unique names.
>Audit-Trail:
>Unformatted:
Dan Winship
