Timeline
10/25/99:
- 10:45 Changeset [13791] by
- Fix typo in merge.
- 10:33 Changeset [13790] by
- Add sendauth prototype, which seems to have disappeared when the prototypes moved from krb-protos.h to krb.h.
10/24/99:
- 14:27 Changeset [13789] by
- Missed a local change in the merge: need $(LOGINLIBS) for klogind since we do libal stuff there.
- 12:35 Changeset [13788] by
- crypto -> k5crypto
10/23/99:
- 15:29 Changeset [13787] by
- compile with warnings enabled
- 15:28 Changeset [13786] by
- convert to autoconf and build with warnings enabled
- 15:28 Changeset [13785] by
- compile with warnings
- 12:07 Changeset [13784] by
- crypto -> k5crypto
10/22/99:
- 12:57 Changeset [13783] by
- Remove the comment about not knowing why the old attach code set an alarm() before getting zephyr subs. We know why now (because libzephyr never timed out if zephyrd/zhm didn't respond) and have fixed it in libzephyr.
- 12:53 Changeset [13782] by
- return ETIMEDOUT after 60 seconds when waiting for a reply from zhm. (in particular, this makes attach eventually time out when trying to get filsrv subs if the zhm or zephyr servers aren't responding).
- 10:35 Changeset [13781] by
- Use free, not krb5_xfree.
10/21/99:
- 00:28 Changeset [13780] by
- Need time() prototype.
- 00:28 Changeset [13779] by
- crypto -> k5crypto
10/20/99:
- 12:44 Changeset [13778] by
- Don't reset a null timer. It's bad. (When removing a packet from the queue.)
10/19/99:
- 22:00 Changeset [13777] by
- This eliminates the pre-4.4 berkeley logout() implementation, and eliminates the getutline stuff in cleanup, because the direct utmp manipulation stuff there should be actually (wonder of wonders) portable. This also elminates a couple of unused variables.
- 21:59 Changeset [13776] by
- This uses library facilities where available, and ought to work on every thing we've ever thought about supporting, and should have the bonus of doing the more correct thing under linux (which has pututline and updwmtp but not pututxline). It takes the more readable approach of implementing logout() on systems that don't.
- 16:39 Changeset [13775] by
- Only use NON_GNU_CPP for Makefiles, not for everything.
- 16:28 Changeset [13774] by
- prevent an unlikely buffer overrun
- 16:27 Changeset [13773] by
- block SIGTSTP, SIGTTIN, SIGTTOU
- 16:27 Changeset [13772] by
- block SIGTSTP, SIGTTOU, SIGTTIN. unlock them after attaching in attachandrun, and check the return value of locker_attach correctly there
- 16:26 Changeset [13771] by
- document requirement that liblocker callers block SIGTSTP, SIGTTIN, SIGTTOU
- 16:25 Changeset [13770] by
- revert 1.7 and 1.8. There can still be problems if signals aren't blocked while individual attachents are locked. Instead, we'll just require liblocker callers to block the signals themselves, since we can't do it safely.
- 16:23 Changeset [13769] by
- compile with warnings
- 13:15 Changeset [13768] by
- crypto -> k5crypto
- 13:14 Changeset [13767] by
- Use regular exit; no reason to use _exit. (Quashes a gcc warning.)
- 10:42 Changeset [13766] by
- Revert 1.7; emacs needs a K&R cpp for its Makefile.
10/18/99:
- 14:43 Changeset [13765] by
- add rcsid, license, descriptive comment
- 13:20 Changeset [13764] by
- autoconf/cleanup/warns
- 13:17 Changeset [13763] by
- on public workstations, rm /etc/athena/cluster.{local,fallback} before the rest of the public workstation cleanup, because we don't want them to affect which packs we end up attaching.
- 11:02 Changeset [13762] by
- crypto -> k5crypto
10/17/99:
- 12:27 Changeset [13761] by
- Don't use broken /lib/cpp on IRIX 6.5.
10/16/99:
- 16:13 Changeset [13760] by
- add athena/etc/writed
- 16:09 Changeset [13759] by
- add (commented out) a rule to queue mail locally rather than forwarding it off to the mailhubs
- 16:08 Changeset [13758] by
- New (non-tainted) implementation of write. writed has moved to athena/etc/writed
- 16:07 Changeset [13757] by
- New (non-tainted) implementation of what was athena/bin/write/writed.c
- 12:13 Changeset [13756] by
- crypto -> k5crypto
- 12:00 Changeset [13755] by
- Get rid of INTERFACE. It doesn't work no more.
10/15/99:
- 22:27 Changeset [13754] by
- don't send bounces with a null envelope from. Violates RFC 1123, but needed with the recent mail system upgrades. [17252] in bugs
- 21:40 Changeset [13753] by
- This patch is entirely cosmetic; it is the result of running indent -npcs -npsl dm.c and tweaking the result slightly.
- 21:39 Changeset [13752] by
- Have dm waitpid() for the X server instead of just sleeping for 4 seconds. Oh yeah, and capitalize a sentence in a comment.
- 19:11 Changeset [13751] by
- . remove initialization of the terminal state of a pty that is never used by humans . change the /dev/console-output grabbing on solaris to work more like it does on non-stream-based systems, and ditche the -inputfd arguments to console, making control-flow more similar from platform to platform, and making things a bit less complex . create a function to write pids to specified filenames, and remove a function that duplicates a teeny bit of the functionality of {f,s,}printf, which was probably originally written to avoid pulling in stdio, which seems to have happened anyway.
- 17:26 Changeset [13750] by
- Disable sdpd.
- 17:22 Changeset [13749] by
- warn when ignoring mountoptions because a locker is already attached. (also, fix a return value and an unlikely memory leak)
- 14:33 Changeset [13748]
- This commit was manufactured by cvs2svn to create tag 'athena-8_3-linux-base'.
- 14:33 Changeset [13747] by
- Pull up 1.1 to athena-8_3 branch.
- 14:33 Changeset [13746]
- This commit was manufactured by cvs2svn to create branch 'athena-8_3'.
- 14:33 Changeset [13745] by
- We changed sendmail.cf.
- 14:25 Changeset [13744] by
- Pull up 1.11 to Athena 8.3 branch.
- 14:25 Changeset [13743] by
- Pull up 1.14 to Athena 8.3 branch.
- 14:25 Changeset [13742] by
- Pull up 1.13 to Athena 8.3 branch.
- 13:48 Changeset [13741] by
- Pulup 1.5 to Athena 8.3 branch.
- 13:15 Changeset [13740] by
- Include errno.h and com_err.h; don't declare things that belong to them.
- 13:05 Changeset [13739] by
- Get rid of INTERFACE. It doesn't work no more.
- 12:50 Changeset [13738] by
- Recomend the do -c flag for building current sources when there are dependencies on new changes.
- 12:39 Changeset [13737] by
- Document -t option to do. In the "building from scratch" section, the krb5 build won't bomb out because it won't go into appl. That section is pretty broken right now, though.
- 12:20 Changeset [13736] by
- Nuke the .rconf file we copy into the lisp subdir.
- 10:57 Changeset [13735] by
- disable root password auth
- 00:59 Changeset [13734] by
- Don't error if the WG port is 65535.
- 00:58 Changeset [13733] by
- Cast hostent h_name field to char * before freeing, to pacify Linux.
- 00:57 Changeset [13732] by
- Create install directories.
- 00:56 Changeset [13731] by
- /usr/tmp -> /var/tmp
10/14/99:
- 18:19 Changeset [13730] by
- autoconfiscation of dm
- 18:18 Changeset [13729] by
- Formerly athena/etc/xdm/dm/dm.[8c] autoconfiscate, and aesthetically reorder #includes
- 14:37 Changeset [13728] by
- Don't declare inet_addr(); it doesn't belong to us.
10/13/99:
- 18:32 Changeset [13727] by
- crypto -> k5crypto for krb5 1.1
- 18:29 Changeset [13726] by
- We've upgraded to krb5 1.1, so -lcrypto is now -lk5crypto.
- 16:01 Changeset [13725] by
- Add an MIT copyright statement.
10/12/99:
- 15:59 Changeset [13724] by
- Pull up 1.12 to Athena 8.3 branch.
- 15:57 Changeset [13723] by
- Pull up 1.2 to athena-8_3 branch.
- 15:57 Changeset [13722] by
- Pull up 1.3 to athena-8_3 branch.
- 15:57 Changeset [13721] by
- Pull up 1.43 to athena-8_3 branch.
- 15:57 Changeset [13720] by
- Pull up 1.13 to athena-8_3 branch.
- 15:57 Changeset [13719] by
- Pull up 1.6 to athena-8_3 branch.
- 15:57 Changeset [13718] by
- Pull up 1.23-1.24 to athena-8_3 branch.
- 15:56 Changeset [13717] by
- Pull up 1.16 to athena-8_3 branch.
- 08:38 Changeset [13716] by
- Treat uid as more authoritative than getenv("USER"), since an su with no arguments doesn't set $USER.
10/11/99:
- 10:14 Changeset [13715] by
- We do not want this.
- 10:14 Changeset [13714] by
- Rename "errno" to "code" in krb_get_err_text prototype.
10/10/99:
- 23:40 Changeset [13713] by
- For krb5 1.1, build without Hesiod support (hard to make krb5 find our Hesiod installation). Also, nuke some obsolete tweaks.
- 22:34 Changeset [13712] by
- Revert rev 1.89; we're turning off the krb5 hesiod support for now.
- 22:25 Changeset [13711] by
- Fix merge; we have to pull some old code back in.
- 17:19 Changeset [13710] by
- Fix merge: krb4-proto.h is gone.
- 13:39 Changeset [13709] by
- krb5 1.1 uses Hesiod in a few places. Not really sure why.
- 12:02 Changeset [13708] by
- Get a couple more places where srvtab is mentioned but any isn't.
10/09/99:
- 09:22 Changeset [13707] by
- Add keytab/any to the list of Makefiles to generate.
- 09:18 Changeset [13706] by
- Oops, back to local mods. Need to descend into any subdir.
- 09:14 Changeset [13705] by
- Add keytab/any to the list of Makefiles to generate.
- 09:10 Changeset [13704] by
- Not needed in krb5 1.1.
- 09:09 Changeset [13703] by
- Update for krb5 1.1.
10/08/99:
- 18:14 Changeset [13702] by
- if invoked with "show -file" on a MIME message, invoke mhshow with the -file flag too. Fixes, among other things, display of MIME messages with MIME-encoded message/rfc822 subparts
- 10:18 Changeset [13701] by
- krb5 1.1 took this feature; go to vendor branch state.
- 10:16 Changeset [13700] by
- Not needed with krb5 1.1.
10/07/99:
- 13:48 Changeset [13699] by
- invoke console with -nosession so it doesn't get session-managed
- 13:06 Changeset [13698] by
- In get_nfs_quota(), add break's to the case blocks handling the status of the getquota RPC, and add a default case. In callaurpc(), use pmap_rmtcall() both as a ping of the remote host, and to obtain the server's UDP port number. This gives us control of the timeout in the initial contact of the host, and lets us skip the portmapper exchange done by clntudp_create().
- 12:59 Changeset [13697] by
- Specify the correct mount table to setmntent().
- 09:39 Changeset [13696] by
- Fix merge: include libkrb524.a again.
- 09:36 Changeset [13695] by
- Fix typo in merge.
10/06/99:
- 23:21 Changeset [13694] by
- Merge local changes.
- 23:20 Changeset [13693] by
- These went away. We'll have to do the changes one level up.
- 23:20 Changeset [13692] by
- Merge local changes. Except we might be leaving out the krb524 library now; might have to fix that.
- 23:18 Changeset [13691] by
- Revert to vendor branch state.
- 22:52 Changeset [13690] by
- Revert to vendor branch state.
- 22:42 Changeset [13689] by
- Merge local changes. This is a bit tricky; they added an "autoencrypt" flag which does what we want, but is off by default. I took out the test for autoencrypt, but that seems kinda wrong. Maybe we should revert to the vendor branch and deal in main.c.
- 22:14 Changeset [13688] by
- Merge local changes.
- 22:14 Changeset [13687] by
- Revert to vendor branch state; problem kinda fixed in stock krb5.
- 22:12 Changeset [13686] by
- Gone now, local changes no longer needed.
- 22:12 Changeset [13685] by
- Merge local changes. ("srvtab" is now stock; "any" is still a local change.)
- 22:11 Changeset [13684] by
- Revert to vendor branch state. (Why did we ever make that local change? Anyway, it's in the mainline now.)
- 22:10 Changeset [13683] by
- Merge local changes. (They added "srvtab" but evidently didn't like "any." Would have been nice if they sent me some mail about it, maybe.)
- 22:09 Changeset [13682] by
- Merge local changes. We should look into making krb5 build against the Athena com_err library, but this will do for now.
- 21:47 Changeset [13681] by
- Revert to vendor branch state. They fixed our bug.
- 21:45 Changeset [13680] by
- Merge local changes.
- 21:42 Changeset [13679] by
- Revert to vendor branch state. Stock krb5 fixed our bug... and then seems to have introduced a resource leak; kt_id doesn't get closed in the success case. Oops.
- 21:21 Changeset [13678] by
- Revert to vendor branch state.
- 21:15 Changeset [13677] by
- Revert to vendor branch state.
- 21:12 Changeset [13676] by
- Merge local changes. Stock krb5 renamed getst() to getstr(), so our local change might not be needed any more.
- 21:10 Changeset [13675] by
- Revert to vendor branch state.
- 21:08 Changeset [13674] by
- Merge local changes.
- 21:08 Changeset [13673] by
- Gone now. Our change is in the stock configure script one level up.
- 21:06 Changeset [13672] by
- We changed it. It's gone now. Our change is in the stock krb.h now.
- 21:00 Changeset [13671] by
- Merge local changes.
- 20:59 Changeset [13670] by
- Revert to vendor branch state; krb5 1.1 fixed our problem in a different way.
- 20:58 Changeset [13669] by
- Merge local changes. The prototypes changes are now standard, so only the srvtab file change remains.
- 20:53 Changeset [13668] by
- Revert to vendor branch state; krb5 1.1 has our change.
- 20:52 Changeset [13667] by
- Merge local changes.
- 20:03 Changeset [13666] by
- Merge local changes.
- 19:21 Changeset [13665] by
- Merge local changes.
- 18:42 Changeset [13664] by
- Merge local changes.
- 16:02 Changeset [13663] by
- Merge local changes.
- 15:46 Changeset [13662] by
- Merge local changes. The v4 rcmd support is no longer a local mod, but most of our bugfixes and the libal support are still local changes.
- 15:44 Changeset [13661] by
- Revert to vendor branch state; local changes no longer necessary.
- 00:08 Changeset [13660] by
- Merge local changes. This merge was done by hand, since krb5 1.1 reindents a lot of the code. Some changes were no longer applicable: * The krb5 ticket options will have to be set elsewhere, because of the new get-initial-creds interface. * The error-reporting when we can't verify a host ticket appears to have been fixed a different way from the way we did it. * The AFS stuff is already conditionalized on SIGSYS in krb5 1.1. * krb5 1.1 sets a SIGHUP handler in dofork(). We still have to nuke the code for getting rid of the controlling tty, though.
10/05/99:
- 14:05 Changeset [13659] by
- Merge local changes.
- 13:47 Changeset [13658] by
- Merge change from configure/configure.in: don't build in appl. (That gets built as a separate package, since stuff in appl depends on liblocker which depends on Kerberos.)
- 13:46 Changeset [13657] by
- Return to vendor-branch state; the subdir list is in Makefile.in now.
- 13:10 Changeset [13656] by
- explode rkinit
- 13:10 Changeset [13655] by
- explode athena/bin/rkinit into athena/bin/rkinit, athena/lib/rkinit, and athena/etc/rkinitd
- 12:55 Changeset [13654] by
- This commit was generated by cvs2svn to compensate for changes in r13653, which included commits to RCS files with non-trunk default branches.
- 12:55 Changeset [13653] by
- Went away in krb5 1.1.
- 12:53 Changeset [13652] by
- athena indent style
- 12:18 Changeset [13651]
- This commit was manufactured by cvs2svn to create tag 'krb5-1_1'.
- 12:18 Changeset [13650] by
- This commit was generated by cvs2svn to compensate for changes in r13649, which included commits to RCS files with non-trunk default branches.
- 12:18 Changeset [13649] by
- Import of Kerberos 5 release 1.1
10/04/99:
- 14:45 Changeset [13648] by
- Try to deal with a trashed CellServDB.
- 14:39 Changeset [13647] by
- Try to recover from trashed passwd and group files.
10/01/99:
- 09:46 Changeset [13646] by
- Only install a new CellServDB if it differs from the current version; sync if we do install a new version.
- 09:43 Changeset [13645] by
- Only copy passwd, shadow, and group if they differ from the .local file; sync after copying them.
09/30/99:
- 12:21 Changeset [13644] by
- further commit-o's.
- 12:20 Changeset [13643] by
- fix typo (we were setting sbindir to @bindir@)
09/29/99:
- 16:40 Changeset [13642] by
- add (empty) check: target
- 10:38 Changeset [13641] by
- Oops. This is a generated file.
09/28/99:
- 18:16 Changeset [13640] by
- fix the paths for several commented out services Digital doesn't sell us sun hardware, nuke that part of the comment
- 18:16 Changeset [13639] by
- change (int) casts to (unsigned char) in ctype calls
- 18:15 Changeset [13638] by
- update for neos explosion
- 18:14 Changeset [13637] by
- split into athena/lib/fxcl, athena/bin/neos, and athena/etc/fxserver
- 18:13 Changeset [13636] by
- NEOS server, from athena/lib/neos/server
- 18:11 Changeset [13635] by
- NEOS clients, from athena/lib/neos/clients
- 18:10 Changeset [13634] by
- kill the Athena deamons using their pid files on all platforms also kill the ssh dameon (but not ssh connections) if it exists.
- 18:07 Changeset [13633] by
- NEOS library, from athena/lib/neos/{lib,ets,include,protocol}
- 17:56 Changeset [13632] by
- make "do.sh" check work for configure-based systems (but not bomb out if there's no check target) and Makefile-based systems)
09/27/99:
- 12:14 Changeset [13631] by
- remove a & on the argument to locker_check_host to make it actually work
09/26/99:
- 09:58 Changeset [13630] by
- Fix warnings on Linux. (printf formats for uid_t, signal handler prototype.)
09/25/99:
- 18:32 Changeset [13629] by
- fix someone's thinko. autoconf is not imake.
Note: See TracTimeline
for information about the timeline view.