summaryrefslogtreecommitdiff
path: root/server
Commit message (Collapse)AuthorAge
...
* fix tyopGravatar Karl Ramm2007-12-28
|
* syslog what sort of authenticator we gotGravatar Karl Ramm2007-12-28
|
* no, -4 does _not_ take an argumentGravatar Karl Ramm2007-12-28
|
* ostensibly accept both sorts of authentication when offering brain dumps, ↵Gravatar Karl Ramm2007-12-28
| | | | and have a command line switch to choose between the sort you use when getting them
* if the #if ladder doesn't find anything, fall back to uname.Gravatar Karl Ramm2007-12-28
|
* gratuitous blank line removed; some common KRB[45]Gravatar Karl Ramm2007-12-26
|
* protect if statement from macro braindamageGravatar Karl Ramm2007-12-25
|
* ostensibly better diagnostic outputGravatar Karl Ramm2007-12-25
|
* /afs/andrew.cmu.edu/usr18/cg2v/hesiod-fix.diffGravatar Karl Ramm2007-12-25
|
* /afs/andrew.cmu.edu/usr18/cg2v/bdump-privmsg.diffGravatar Karl Ramm2007-12-25
|
* /afs/andrew.cmu.edu/usr18/cg2v/zephyr-krb5-server-20071215+.diffGravatar Karl Ramm2007-12-25
|
* These HAVE_KRB4 things are actually HAVE_KRB5Gravatar Karl Ramm2007-12-25
|
* applied athena-update-branch patchGravatar Karl Ramm2007-12-25
|
* debian patch applied and reconciled, version unfinalizedGravatar Karl Ramm2007-07-20
|
* Add shared library support for libzephyr.Gravatar Greg Hudson2005-07-14
|
* Remove kopt.c and the caching things that depend on it.Gravatar Garry Zacheiss2004-02-29
|
* Don't allow foreign zephyrd to request a triplet with a non-wildcardGravatar Garry Zacheiss2002-07-02
| | | | recipient.
* Don't accept subscriptions from foreign realms for acl'ed classes.Gravatar Garry Zacheiss2001-07-02
|
* Reintroduce checksum enforcement changes.Gravatar Greg Hudson2001-04-10
|
* Reintroduce rev 1.23.Gravatar Greg Hudson2001-03-05
|
* Revert rev 1.22; we're not ready to start enforcing checksums onGravatar Greg Hudson2001-03-05
| | | | | | all operations yet. Also revert rev 1.23, because I got confused. We'll reintroduce it shortly.
* Revert last change; we're not ready to start enforcing checksums onGravatar Greg Hudson2001-03-05
| | | | all operations yet.
* Fix a couple of small memory leaks, and make zephyrd not mistreat peopleGravatar Garry Zacheiss2001-02-28
| | | | who were trying to subscribe to acl'ed classes they weren't on the acl for.
* sender_in_realm() was renamed to realm_sender_in_realm(), and now takesGravatar Garry Zacheiss2001-02-26
| | | | different arguments.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | | * Better interrealm support: Realm state support, add realm state enum and add additional opcode for realm initialization code. * Prototype updates.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | | | | | | | | | | | | | | | | * prototype updates. * Fix some memory leaks. * Better interrealm support: - subscr_subscribe and add_subscriptions need to know what server requests are coming from. - pay attnetion to when we have no more local users listening to a foreign sub and unsubscribe. - Remove unused code. - Call realm_handoff with correct client info. - Add routine to resend our idea of subs in a remote realm to that realm.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | * Some memset 0 cleanliness. * Support enhanced 'zstat -s' output.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | * Replace entire file, resyncing interrealm code between us and CMU.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | * Better interrealm support: Let other realms know we booted, set realm correctly we so can override from the command line, and kill any children that are trying to get tickets.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | * Better interrealm support: be prepared to try old and new style checksumming, and make NOENCRYPTION case work.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | * Remove uneeded casts. * Remove uneeded local versions of Kerberos functions.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | | | | | | | | | | | | | | | | | | * Better interrealm support: - Don't do unneeded check on ack authenticity. - Avoid allowing a remote realm to forge a local user sending to an acl'ed class. - Fix calling conventions. - Refragment packets if we grew them and now need to. - Avoiding failing to ack if packet grew. - Tell other realms we're shutting down. - Avoid nacking bogus packets from broken servers. * Avoid buffer overrun (in #if 0'd code) * memset 0 cleanliness.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | * If class is empty, return useful info so interrealm code can unsubscribe.
* Changes from CMU:Gravatar Garry Zacheiss2001-02-26
| | | | | | | | * Better interrealm support, including setting realm correctly so we can override it with a command line switch, dealing correctly when we don't know a realm, and calling subscr_subscribe with server info. * Some memset 0 cleanliness.
* If we found no servers for a given realm, syslog and abort so we don'tGravatar Garry Zacheiss2001-02-19
| | | | divide by zero later.
* close possible syslog format string holeGravatar Jonathon Weiss2000-10-04
|
* Fix potential buffer overruns.Gravatar Greg Hudson2000-05-17
|
* Remove an old compatibility compromise from back when we started requiringGravatar Greg Hudson2000-02-11
| | | | | checksums on messages. With this change, you can't fiddle with subs if you don't have a valid checksum.
* Reinstate the policy that realm-visible locations cannot be seen byGravatar Greg Hudson2000-01-22
| | | | unauthentic ulocate requests.
* /usr/tmp -> /var/tmpGravatar Greg Hudson1999-11-29
|
* fix things Irix n32 cc complains aboutGravatar Dan Winship1999-08-12
|
* Nuke these files.Gravatar Greg Hudson1999-06-13
|
* Nuke the acl files. They bad.Gravatar Greg Hudson1999-06-13
|
* Some RCS ID cleanup: delete $Log$ and replace other RCS keywords with $Id$.Gravatar Greg Hudson1999-01-22
|
* Revert version 1.50. Empirically, this workaround didn't seem to helpGravatar Greg Hudson1998-11-19
| | | | | | | (I'm not sure why, but it still took ten hours for erato to recover from having an invalid tgt), and we now understand what the problem is (the KDC is handing out identical answers for requests from two different clients; obviously, one of them gets an incorrect address).
* Actually increment the loop pointer in ulogin_find().Gravatar Greg Hudson1998-10-26
|
* If we get RD_AP_BADD from SendKerberosData(), this indicates that weGravatar Greg Hudson1998-10-21
| | | | | have a bad TGT, so discard our tickets. This should never happen, but it does. If we ever figure out why, then this change can go away.
* Don't free nacked until we've finished looking at it (an alternativeGravatar Greg Hudson1998-10-16
| | | | | | would be to save nacked->client in a variable). Do make sure to remove nacked from the list before calling client_deregister(), though.
* Clean up the dump format: report addresses of clients, and fall back oneGravatar Greg Hudson1998-10-13
| | | | | tab. The old dump format was written for when we grouped clients by host manager.
* Search for locations by <user,addr,port> in cases where we were searchingGravatar Greg Hudson1998-10-13
| | | | by <user,machine,tty> (which includes setting exposure and logout).
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-26 14:49:47 +0000