logo_kerberos.gif

Release Meeting Minutes/2009-11-24

From K5Wiki
< Release Meeting Minutes
Revision as of 16:08, 24 November 2009 by TomYu (talk | contribs) (New page: Steve Buckley, Sam Hartman, Thomas Hardjono, Luke Howard, Greg Hudson, Zhanna Tsitkova, Tom Yu Branch deletion? Tom used to say that branches shouldn't be deleted because it makes histor...)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Steve Buckley, Sam Hartman, Thomas Hardjono, Luke Howard, Greg Hudson, Zhanna Tsitkova, Tom Yu

Branch deletion? Tom used to say that branches shouldn't be deleted because it makes history harder to find. It's less of a problem now that we have FishEye, which allows browsing through deleted directories.

Greg
get_init_creds touched by FAST; also affected by IAKERB stuff. Not merging TGS patch changes because there are currently code duplication issues.
Sam
A 1.8 client will fail to do IAKERB but later clients will succeed (?)
Luke
new GSS-API -- gss_acquire_cred_password
Sam
Cross-realm will cause AS-IAKERB to fail.
Greg
Should only integrate new krb5 APIs for now.
Luke
Heimdal only has *_step functions, not GSS.
Sam
Want to integrate test case for FAST. Might have time to integrate into dejagnu. Complexity in kinit procedure in default.exp? Actually that's to support some strange "standalone" stuff.
Greg
crypto -- working on consolidating internal interface down to input/output IOVs. single-DES IOV implementation not complete? [ Luke only made the bits that were needed to implement WrapEx. ] Also v1_seal calls into RC4 internals. [ Some debate about whether calling through function pointers (accessor etc.) is better than using a functional interface. ]
Tom
xorkey situation [krbdev.mit.edu #6584] -- test suite checks that the relevant checksum functions talk to themselves, not whether they validate against test vectors.