summaryrefslogtreecommitdiff
path: root/ChangeLog
Commit message (Collapse)AuthorAgeFilesLines
* Added Tamil translation by Dinesh.Fatih Demir2001-10-071-0/+4
|
* user PAM_ESTABLISH_CRED for pam_setcredGeorge Lebl2001-10-051-0/+4
| | | | | | Fri Oct 05 13:58:41 2001 George Lebl <jirka@5z.com> * daemon/verify-pam.c: user PAM_ESTABLISH_CRED for pam_setcred
* Release 2.2.4.2GDM_2_2_4_2George Lebl2001-10-041-0/+8
| | | | | | | | | | Thu Oct 04 13:19:50 2001 George Lebl <jirka@5z.com> * Release 2.2.4.2 Thu Oct 04 13:18:30 2001 George Lebl <jirka@5z.com> * daemon/misc.c: the mode for sticky 01777 not 02777. Whoops.
* also check stale lock files when looking for empty display namesGeorge Lebl2001-10-041-2/+3
| | | | | | | Thu Oct 04 00:12:05 2001 George Lebl <jirka@5z.com> * daemon/misc.c, gui/gdmXnestserver.c: also check stale lock files when looking for empty display names
* Release 2.2.4.2George Lebl2001-10-041-0/+4
| | | | | | Wed Oct 03 23:55:32 2001 George Lebl <jirka@5z.com> * Release 2.2.4.2
* fix segfault on send_got_manged_forwardGeorge Lebl2001-10-041-2/+2
| | | | | | Wed Oct 03 23:54:49 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.c: fix segfault on send_got_manged_forward
* Release 2.2.4.2George Lebl2001-10-041-0/+4
| | | | | | Wed Oct 03 23:12:00 2001 George Lebl <jirka@5z.com> * Release 2.2.4.2
* ensure the /tmp/.ICE-unix directory. Hopefully we're correct.George Lebl2001-10-041-0/+9
| | | | | | | | | | | Wed Oct 03 20:26:27 2001 George Lebl <jirka@5z.com> * daemon/misc.[ch], daemon/slave.[ch]: ensure the /tmp/.ICE-unix directory. Hopefully we're correct. * daemon/server.c: when running server as root, unset suplementary groups and set group to 0. So that the files X creates it creates with gid 0 as well rather then with the 'gdm' group name.
* deal with the case where MSG_NOSIGNAL is not defined by temporairlyGeorge Lebl2001-09-271-0/+6
| | | | | | | | Wed Sep 26 23:54:11 2001 George Lebl <jirka@5z.com> * daemon/gdm-net.c, gui/gdmflexiserver.c: deal with the case where MSG_NOSIGNAL is not defined by temporairly ignoring the SIGPIPE signal. Fixes #60586
* Add a failsafe question dialog similar in operation to the failsafe errorGeorge Lebl2001-09-271-0/+17
| | | | | | | | | | | | | | | | | | | Wed Sep 26 21:01:38 2001 George Lebl <jirka@5z.com> * daemon/verify.h, daemon/verify-*.c, daemon/slave.c, daemon/errorgui.[ch], daemon/gdm.c: Add a failsafe question dialog similar in operation to the failsafe error box, and use it in the standalone pam conv function. Kill all instances of PAM_SILENT since we can now always converse with the user. * daemon/verify-pam.c: use a global pamh again but this time a bit smarter. Make sure if we use a global one that it has all the right things set. Also set PAM_RHOST to "localhost" or the remote host if not a "console" login. Set PAM_RUSER to "gdm" (or whatever the gdm user is). I'm sure I'm fucking something up again but I can't find enough docs to verify that what I'm doing is 100% correct. Not to mention that there seem to be some braindead modules out there to begin with.
* fix crash when quiver would happen while iconizedGeorge Lebl2001-09-171-0/+10
| | | | | | | | | | | | Mon Sep 17 07:17:32 2001 George Lebl <jirka@5z.com> * gui/gdmlogin.c: fix crash when quiver would happen while iconized * daemon/misc.c: include signal.h * gui/gdmlogin.c: don't go nuts on just minor version change, only tell the user to piss of when the greeter protocol version changed.
* Apply patch from Ali Akcaagac <ali.akcaagac@stud.fh-wilhelmshaven.de> toGeorge Lebl2001-09-131-0/+6
| | | | | | | | Thu Sep 13 00:20:46 2001 George Lebl <jirka@5z.com> * gui/gdmlogin.c: Apply patch from Ali Akcaagac <ali.akcaagac@stud.fh-wilhelmshaven.de> to fix up the 24hr time display to not show the %p
* Apply and finish a patch from Terje Rosten <terjeros@phys.ntnu.no> to addGeorge Lebl2001-09-111-0/+7
| | | | | | | | | Tue Sep 11 02:34:47 2001 George Lebl <jirka@5z.com> * daemon/gdm.h, config/gdm.conf.in, gui/gdmconfig.(c|glade), gui/gdmlogin.c, docs/C/gdm.sgml: Apply and finish a patch from Terje Rosten <terjeros@phys.ntnu.no> to add an option to force 24 hour clock on the greeter
* use the safe fork functionGeorge Lebl2001-09-111-0/+10
| | | | | | | | | | | | Tue Sep 11 01:37:42 2001 George Lebl <jirka@5z.com> * daemon/errorgui.c: use the safe fork function * daemon/gdm.c: a doh kind of error, waitpid returns 0 to mean 'no child', not -1 as I assumed. * gui/gdmflexiserver.c: another doh!, actually whack the screen portion of the display number
* raise versionGeorge Lebl2001-09-111-0/+23
| | | | | | | | | | | | | | | | | | | | | | | | | Mon Sep 10 20:20:59 2001 George Lebl <jirka@5z.com> * configure.in: raise version * daemon/gdm.[ch], daemon/auth.c, daemon/misc.[ch], daemon/display.c, daemon/server.c, daemon/slave.c, gui/gdmflexiserver.c: Implement a new scheme for the FLEXI_XNEST command. Old implementations will get ERROR 100 (I doubt there are any). This solves some security issues with the Xnest stuff. The Xnest server is now run as the user who owns the .Xauthority file, but only if that file is readable ONLY by the user and only if the user can access keys inside it. An MIT-MAGIC-COOKIE-1 must exist for the current display on which Xnest will run. This is of course true for gdm created sessions. * daemon/misc.[ch], daemon/gdm.c, daemon/server.c, daemon/slave.c, daemon/display.c: Implement a race safe fork, and improve the signal handlers to handle some more corner cases and avoid some races. Also remove the old last_killed_pid hack since it's now handled nicer. * daemon/gdm.[ch]: add a protocol
* Got the bug number wrong darnit, it should be #59977George Lebl2001-09-081-1/+1
| | | | -George
* OK, PAM documentation is officially on crack. Some places (the kernel.orgGeorge Lebl2001-09-081-0/+7
| | | | | | | | | Sat Sep 08 02:35:51 2001 George Lebl <jirka@5z.com> * daemon/verify-pam.c: OK, PAM documentation is officially on crack. Some places (the kernel.org web manual) say order should be set_cred -> open_session, but it seems that the order is open_session -> set_cred. Should fix #51605
* use pclose on popen'ed FILE *George Lebl2001-09-081-0/+4
| | | | | | Sat Sep 08 02:06:44 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.c: use pclose on popen'ed FILE *
* handle error from fork when daemonizing wellGeorge Lebl2001-09-071-0/+4
| | | | | | Thu Sep 06 20:17:31 2001 George Lebl <jirka@5z.com> * daemon/gdm.c: handle error from fork when daemonizing well
* "fix" for the toolbar deadlock problem. apparently gnome-dock code isGeorge Lebl2001-09-061-0/+9
| | | | | | | | | | | Thu Sep 06 15:55:20 2001 George Lebl <jirka@5z.com> * gui/gdmwm.c: "fix" for the toolbar deadlock problem. apparently gnome-dock code is pretty braindead and uses the bulldog technique for grabbing the pointer, grab in a tight loop until we succeed. But the it will fail until the main loop is hit so a deadlock results. Fix this by defining that function again in gdmwm with a more "maybe" correct semantics.
* Updated ChangeLogTrevor Curtis2001-09-041-0/+4
|
* Rename all traces of zh_TW.Big5 to zh_TW.Kwok-Koon Cheung2001-09-041-0/+4
|
* When we run a new flexi server lock the current screen with xscreensaver.George Lebl2001-09-031-0/+8
| | | | | | | | | | Mon Sep 03 00:38:23 2001 George Lebl <jirka@5z.com> * gui/gdmflexiserver.c: When we run a new flexi server lock the current screen with xscreensaver. We need a nicer locker though and one that can also run gdmflexiserver. And this even for normal gnome locking. It would be nice if xscreensaver allowed a custom locker dialog.
* Updated ChangeLogTrevor Curtis2001-09-031-0/+4
|
* Implement a GOT_MANAGED_FORWARD opcode. Also send MANAGED_FORWARD 3 timesGeorge Lebl2001-09-021-0/+7
| | | | | | | | | Sun Sep 02 16:44:28 2001 George Lebl <jirka@5z.com> * daemon/gdm.h, daemon/xdmcp.c: Implement a GOT_MANAGED_FORWARD opcode. Also send MANAGED_FORWARD 3 times (1.5 sec spaced apart) or until we get a GOT_MANAGED_FORWARD. This is a compatible change to the gdm xdmcp protocol so version wasn't raised.
* Updated trevors emailGeorge Lebl2001-09-021-0/+4
| | | | | | Sat Sep 01 14:53:22 2001 George Lebl <jirka@5z.com> * AUTHORS, docs/gdmconfig/C/gdmconfig.sgml: Updated trevors email
* the decline messages were switched still. I've been up too lateGeorge Lebl2001-09-011-0/+5
| | | | | | | Sat Sep 01 07:00:52 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.c: the decline messages were switched still. I've been up too late
* Release 2.2.4.1George Lebl2001-09-011-0/+4
| | | | | | Sat Sep 01 05:31:07 2001 George Lebl <jirka@5z.com> * Release 2.2.4.1
* Fix decline strings, and make the MANAGED_FORWARD work if the originGeorge Lebl2001-09-011-0/+5
| | | | | | | Sat Sep 01 05:29:31 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.[ch], daemon/choose.c: Fix decline strings, and make the MANAGED_FORWARD work if the origin address is local.
* Change the gdm extension protocol, the MANAGED_FORWARD passes address ofGeorge Lebl2001-09-011-0/+12
| | | | | | | | | | | | | | Sat Sep 01 04:15:38 2001 George Lebl <jirka@5z.com> * daemon/gdm.h, daemon/xdmcp.c, daemon/choose.[ch]: Change the gdm extension protocol, the MANAGED_FORWARD passes address of origin rather then hostname of self. And we now handle the message correctly. Also on DECLINE, send a descriptive english string to aid in debugging XDMCP setup. * configure.in: raise version * config/gdm.conf.in: add the DisplaysPerHost default
* Updating ChangeLog, I forgot an entryGeorge Lebl2001-09-011-0/+4
| | | | -George
* Implement DisplaysPerHost which was documented so it should have worked inGeorge Lebl2001-09-011-0/+22
| | | | | | | | | | | | | | | | | | | | | | | | Sat Sep 01 00:47:36 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.c, daemon/gdm.[ch], daemon/server.c: Implement DisplaysPerHost which was documented so it should have worked in the first place * daemon/xdmcp.c, daemon/server.c, daemon/auth.c, daemon/gdm.c, daemon/choose.c, daemon/cookie.c, daemon/display.c, daemon/slave.c, gui/gdmlogin.c: fix some memleaks, none really dangerous * daemon/xdmcp.c: messages such as status information shouldn't be translated. Because they go over a wire with no specification of charset or language. * gui/gdmlogin.c: Fix broken "session don't exist wanna save new one" message * daemon/gdm.c: Make sure static display numbers are unique. and fix a possible crash in flexible server code * docs/C/gdm.sgml: a small update
* Update for all the server def stuff.George Lebl2001-09-011-0/+4
| | | | | | Fri Aug 31 20:22:33 2001 George Lebl <jirka@5z.com> * docs/C/gdm.sgml: Update for all the server def stuff.
* Just update the changelog with stuff I forgot. Mainly that theGeorge Lebl2001-09-011-1/+4
| | | | Willing patch was sent by Matt Forrest <mforrest@scs.ryerson.ca>
* run the Willing script at most once every 3 seconds. That way we canGeorge Lebl2001-09-011-0/+5
| | | | | | | Fri Aug 31 18:21:46 2001 George Lebl <jirka@5z.com> * daemon/xdmcp.c: run the Willing script at most once every 3 seconds. That way we can prevent some DOS attacks.
* Add an xdmcp/Willing script configuration option. To run a script when theGeorge Lebl2001-09-011-0/+14
| | | | | | | | | | | | | | | | Fri Aug 31 18:12:05 2001 George Lebl <jirka@5z.com> * config/gdm.conf.in, daemon/gdm.[ch], daemon/xdmcp.c: Add an xdmcp/Willing script configuration option. To run a script when the system gets the QUERY command. * daemon/gdm.c: Fix checks for the chooser and greeter binaries * gui/gdmconfig.{glade,c}: Add configuration for the Willing thing * docs/C/gdm.sgml: Document some of the new configuration keys. Not everything is documented yet.
* Fix reading and changing servers with respect to server numbers. AllGeorge Lebl2001-09-011-0/+7
| | | | | | | | | Fri Aug 31 17:04:02 2001 George Lebl <jirka@5z.com> * gui/gdmconfig.c: Fix reading and changing servers with respect to server numbers. All server numbers are now normalized and they equal the actual row (such assumptions were made before, but were not enforced). Also fix some memory leaks.
* fix writing of servers to config fileGeorge Lebl2001-08-311-0/+4
| | | | | | Fri Aug 31 16:34:58 2001 George Lebl <jirka@5z.com> * gui/gdmconfig.c: fix writing of servers to config file
* Rename zh_TW.Big5 -> zh_TW, to prevent future encoding naming problem.Kwok-Koon Cheung2001-08-311-0/+4
|
* s/login/username/ one more time, and in XDMCP mode call it "Disconnect"George Lebl2001-08-291-0/+6
| | | | | | | | Tue Aug 28 23:17:41 2001 George Lebl <jirka@5z.com> * gui/gdmlogin.c: s/login/username/ one more time, and in XDMCP mode call it "Disconnect" not "Quit", but in flexi mode, it should still be "Quit".
* redo restart logic to be a lot smarter and not hang on leftover fifo'sGeorge Lebl2001-08-291-0/+9
| | | | | | | | | | | Tue Aug 28 19:23:34 2001 George Lebl <jirka@5z.com> * gdm.spec.in: redo restart logic to be a lot smarter and not hang on leftover fifo's Tue Aug 28 17:37:12 2001 George Lebl <jirka@5z.com> * Release 2.2.4.0
* Updated traditional Chinese translation, and cleared up ambiguity of aKwok-Koon Cheung2001-08-261-0/+5
| | | | string in gdm_file_check function.
* Minor cleanup and paranoiaGeorge Lebl2001-08-261-0/+22
| | | | | | | | | | | | | | | | | | | | | | | | Sun Aug 26 01:45:33 2001 George Lebl <jirka@5z.com> * daemon/auth.c, daemon/cookie.c: Minor cleanup and paranoia * daemon/gdm-net.[ch], daemon/gdm.[ch], daemon/server.c, daemon/slave.[ch]: The daemon now keeps track of the cookies in text form at least so that it can do authentication of it's own. FLEXI_XSERVER now requires an "authenticated" socket connection using AUTH_LOCAL (a new opcode). This way only those users that are logged in from gdm on a console can start new console/fullscreen flexi servers. I mean for remote people it wouldn't work anyway. * gui/gdmflexiserver.c: implement AUTH_LOCAL authentication scheme by reading from the .Xauthority file and looking for FamilyLocal entries from gdm. (or those looking like it). Tests all candidates and find the one that works for gdm. This could cause some gdm_info messages in the log, but I suppose that's ok, vast majority of the time there will only be one that is even tried. Also added a debug flag and get xauth filename from XauFileName rather then my own hand created method.
* don't use an extra printf, if this was user supplied it could have been aGeorge Lebl2001-08-251-0/+9
| | | | | | | | | | | Sat Aug 25 21:35:29 2001 George Lebl <jirka@5z.com> * daemon/display.c, daemon/gdm.c: don't use an extra printf, if this was user supplied it could have been a security hole, like this it was just one extra allocation. * gui/gdmXnestchooser.c, gui/gdmchooser.c: cast address argument to connect to avoid a warning.
* Fix wiping of pid file on "gdm already exists" messageGeorge Lebl2001-08-251-0/+13
| | | | | | | | | | | | | | | Sat Aug 25 10:06:39 2001 George Lebl <jirka@5z.com> * daemon/gdm.c, daemon/misc.c: Fix wiping of pid file on "gdm already exists" message * daemon/gdm.c, daemon/gdm-net.c: Don't unlink files on close. Instead only unlink things on final_cleanup. This fixes the disappearing socket trick. * gui/gdmflexiserver.c: When not local refuse to run. This is still not perfect. But I don't suppose we'll get much better. Unfortunately a new string. The translators will beat me up.
* update to newest from gnome-coreGeorge Lebl2001-08-241-0/+4
| | | | | | Fri Aug 24 19:32:42 2001 George Lebl <jirka@5z.com> * docs/sgmldocs.make: update to newest from gnome-core
* add gdmopen and gdkaskpass to the summaryGeorge Lebl2001-08-231-0/+4
| | | | | | Thu Aug 23 12:03:41 2001 George Lebl <jirka@5z.com> * configure.in: add gdmopen and gdkaskpass to the summary
* Update for the server stuffGeorge Lebl2001-08-231-0/+16
| | | | | | | | | | | | | | | | | | Thu Aug 23 10:10:33 2001 George Lebl <jirka@5z.com> * config/gdm.conf.in: Update for the server stuff * daemon/display.c, daemon/errorgui.c, daemon/misc.c, daemon/slave.c: close fds no longer needed. Close all needed things after forking, Some other related cleanup as well. * gui/gdmlogin.c: A minor UI cleanup, make the entry at least 250 pixels wide, in an attempt t make the window not change size all the time * gui/gdmphotosetup.c: Don't require the browser to be on. Just warn the user that the browser is off, but still let him pick the face.
* Also test for whiptail if we can't find dialog nor gdialogGeorge Lebl2001-08-221-0/+8
| | | | | | | | | | Wed Aug 22 23:23:21 2001 George Lebl <jirka@5z.com> * config/XKeepsCrashing, daemon/misc.c: Also test for whiptail if we can't find dialog nor gdialog * daemon/misc.[ch], daemon/gdm.c: Make sure things that run dialog (and could thus run gdialog) don't have DISPLAY set
* Add two new small sbin utils. gdmopen is like open (based on it) butGeorge Lebl2001-08-221-0/+44
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Wed Aug 22 23:01:59 2001 George Lebl <jirka@5z.com> * configure.in, Makefile.am, utils/gdmopen.c, utils/gdmaskpass.c: Add two new small sbin utils. gdmopen is like open (based on it) but better suited for gdm. Exitcode is returned so no temp files are needed. Process is cleaned up on a signal, the new vt is deallocated after use. And unlike openvt this thing actually waits for the child. Gdmaskpass just asks for the root password, for use in the XKeepsCrashing script. Both of these things will only run as root as I'm too paranoid and it doesn't make sense for them to run otherwise for gdm anyway. * configure.in: Fix sbin expansion. * daemon/gdm.c, daemon/misc.[ch], daemon/slave.c, daemon/server.c: Use gdmopen to run dialog and add helper functions for doing this. * daemon/gdm.h, daemon/slave.c, daemon/server.[ch]: When a server is busy, ask the user if we want to run it on a new display number or retry. * daemon/gdm.c, daemon/misc.c, daemon/slave.c: Make sure to whack any extra helper processes we launch when we are killed so that we don't leave things like 'dialog' processes around * daemon/gdm.c, config/XKeepsCrashing: Use gdmaskpass, ask if the user wants to view X logs if X is crashing. And use gdmopen in the script and correctly abort with exit 32. Also make it not use the ugly tempfile stuff (Though a tempfile is still given to the script for now) * gdm.spec.in: raise epoch to 1, people abuse this field, and if they do abuse it again and raise it beyond 1, I'll just max it out since it makes package upgrading not work. * daemon/slave.c, daemon/verify-(pam|crypt|shadow).c: add /sbin/nologin as a thing to test for for disabled logins in addition to /bin/false and /bin/true (it worked, but this adds a nicer message) * daemon/xdmcp.c: fix a warning. * gui/gdmflexiserver.desktop.in: kill TryExec here
View Lorry Controller Status