[krbdev.mit.edu #6787] S4U memory leak

Greg Hudson via RT rt-comment at krbdev.mit.edu
Wed Sep 29 13:39:57 EDT 2010


The other solution I can imagine is creating an anonymous memory ccache 
variant which destroys itself on exit.  There would be implications for 
the current implementation of krb5_cc_dup(), as well as any code which 
tries to use the ccache name.

A flag in the gss-krb5 creds structure is probably more conservative, and 
is fine for now.



More information about the krb5-bugs mailing list