Timeline
05/22/97:
- 14:20 Changeset [10066] by
- getservbyname returns a service, not a host, so don't use it. fix error messages to include `lert' document -q option
05/21/97:
- 22:47 Changeset [10065] by
- Remove unused old copy of Console configuration and some unused code in the Imakefile
- 22:45 Changeset [10064] by
- Make console window 86 characters wide to fit 80-character-wide messages plus the timestamp
- 15:31 Changeset [10063] by
- We need a kernel update on the SGI since rtsymtab was updated. NEWOS doesn't do anything on the SGI, so punt. (Why are we changing the 8.1.0 update script after 8.1.0 is out? Well, anyone who took the update and didn't get their kernel rebuilt isn't going to be able to take an update now, so changing it here and releasing the volume is enough.)
- 07:03 Changeset [10062] by
- Take update from internic (k.root-servers.net moved to LINX).
05/20/97:
- 23:43 Changeset [10061] by
- Oops, correct a typo.
- 22:10 Changeset [10060] by
- We have an interesting problem that latex can't find fonts with the double / in the default tfm path. Since we don't need that double slash, eliminate it, and while we're here, remove a bunch of other stuff we don't need.
- 20:49 Changeset [10059] by
- Modify the bibtex paths to fit what we want. (The kpathsea distribution really should define these as variables in Makefile.in, but it doesn't.) Using texpooldir to get /usr/athena/lib/tex is pretty gross.
05/19/97:
- 21:12 Changeset [10058] by
- Nuke a double / in the fontdir.
- 21:11 Changeset [10057] by
- It turns out kpathsea can get built without the make arguments from the top-level Makefile (it can be built laterally from other directories). So we have to set up the paths here as well as in the top level.
05/18/97:
- 13:38 Changeset [10056] by
- Add 8.1.1 version script on Athena 8.1 release branch.
- 13:35 Changeset [10055]
- This commit was manufactured by cvs2svn to create branch 'athena-8_1'.
- 13:35 Changeset [10054] by
- Move /tmp onto /var partition on Suns. Refresh /etc/aliases on SGI; it was stomped when tracking /install.
05/17/97:
- 05:17 Changeset [10053] by
- Revert the change in rev. 1.38.2.1; it is no longer necessary.
- 04:51 Changeset [10052] by
- Pull up change from rev. 1.11 to Athena 8.1 release branch.
- 04:46 Changeset [10051] by
- Don't track platform directories at all, even as symlinks, to avoid trashing the appropriate one (which gets copied local by the update or install script) when tracking /os.
- 04:45 Changeset [10050] by
- Pull up change from rev. 1.88 to Athena 8.1 release branch.
- 04:44 Changeset [10049] by
- Calculate a deterministic and unique id for the utmp entry.
05/16/97:
- 11:55 Changeset [10048] by
- Pull up change from rev. 1.11 to Athena 8.1 release branch.
- 11:00 Changeset [10047] by
- Fix typo (insufficIent) and make "cannot" consistent (bugs[15138])
05/15/97:
- 22:31 Changeset [10046] by
- Pull up change from rev. 1.3 to Athena 8.1 release branch.
- 22:30 Changeset [10045] by
- Run with umaks 022 so that /var/athena/rc.conf.sync is not mode 666.
- 22:29 Changeset [10044] by
- Delete on 8.1 release branch.
- 22:28 Changeset [10043] by
- Pull up change from rev. 1.2 to Athena 8.1 release branch.
- 22:27 Changeset [10042] by
- var.cpio is no longer necessary.
- 22:26 Changeset [10041] by
- Pull up change in rev. 1.5 to Athena 8.1 release branch.
- 22:26 Changeset [10040] by
- The olc_browser filename has an underscore, not a dash.
- 22:25 Changeset [10039] by
- Pull up change from rev. 1.12 to Athena 8.1 release branch.
- 22:23 Changeset [10038] by
- Copy /os/var to /root/var using tar; punt the var.cpio file. Do the copy after tracking the OS but before tracking the srvd or copying config files, so the srvd and the configuration files override anything in there. Make an empty /var/adm/sulog, since we log to it.
- 16:21 Changeset [10037] by
- Pull up change from rev. 1.25 to Athena 8.1 release branch.
- 16:20 Changeset [10036] by
- Don't redirect input to be from /dev/null for a manual update.
05/14/97:
- 23:44 Changeset [10035] by
- Pull up change from rev. 1.5 to Athena 8.1 release branch.
- 23:43 Changeset [10034] by
- On Solaris, run named, since it isn't normally running.
- 23:41 Changeset [10033] by
- Pull up change from rev. 1.11 to Athena 8.1 release branch.
- 23:40 Changeset [10032] by
- Copy in the appropriate /platform directory after tracking the OS. Copy in the crontab file from the srvd after the cpio for /var; this is not the best answer but will do for now.
- 23:33 Changeset [10031] by
- Pull up change from rev. 1.3 to Athena 8.1 release branch.
- 23:32 Changeset [10030] by
- Install access_off owned by root, since it's a setuid binary.
- 15:11 Changeset [10029] by
- Prefer fcntl locking to lockf locking in all three places, not just the first one.
- 06:00 Changeset [10028] by
- Pull up change in rev 1.8 to Athena 8.1 release branch.
- 05:58 Changeset [10027] by
- Propagate old 8.0 key repeat change.
05/13/97:
- 20:13 Changeset [10026] by
- Fix the tag name for the release branch. Add a sanity check for partition sizes, since we lost on that during beta for Solaris in 8.1.
- 16:29 Changeset [10025] by
- Emergency temporary change: don't track the /os to avoid nuking stuff under /platform. I'll work out a better way of dealing with this soon. (I hate pressing problems.)
- 16:06 Changeset [10024] by
- Pull changes from rev. 1.10 into 8.1 release branch.
- 16:05 Changeset [10023] by
- Pull changes from rev. 1.18 into 8.1 release branch.
- 16:01 Changeset [10022] by
- Bring the relevant /platform directory for the current architecture local during the update. Streamline the $platform handling a bit.
- 16:00 Changeset [10021] by
- Put a bunch of things in exceptions lists to save space. The relevant /platform directory for a given architecture will be copied local by the update script.
- 09:47 Changeset [10020] by
- Don't run aklog on `attach -h -n'. (speed hack)
05/12/97:
- 16:23 Changeset [10019] by
- DTRT if .anyone doesn't have a final newline
05/11/97:
- 13:54 Changeset [10018] by
- Default to finding sources in /mit/source-8.1
05/08/97:
- 17:23 Changeset [10017]
- This commit was manufactured by cvs2svn to create tag 'patch-2_1'.
- 17:23 Changeset [10016]
- This commit was manufactured by cvs2svn to create tag 'ctan-1997_02_28'.
- 17:23 Changeset [10015]
- This commit was manufactured by cvs2svn to create tag 'ctan-1996_10_21'.
- 17:23 Changeset [10014]
- This commit was manufactured by cvs2svn to create tag 'athena-8_1-beta'.
- 17:23 Changeset [10013]
- This commit was manufactured by cvs2svn to create tag 'afs-3_4a_patches_4'.
- 17:23 Changeset [10012]
- This commit was manufactured by cvs2svn to create branch 'athena-8_1'.
- 17:23 Changeset [10011] by
- Whoops, the keycode for the Alt key is 26, not 19 (there's an offset of 7 between the number in the keymaps and the X keycode, whee).
- 17:07 Changeset [10010] by
- Add a section on putting out patch releases.
- 03:35 Changeset [10009] by
- changes for the dialups (the dialup load balancing named requires these) including load average reporting for Solaris, and determining the existance of /etc/nologin. Changes by nathanw.
- 03:31 Changeset [10008] by
- From danw: use override flag to attach packs, since it's restricted.
- 03:31 Changeset [10007] by
- Use override flag to attach /srvd, since it's restricted.
- 03:30 Changeset [10006] by
- From danw: clean up, and don't allow mortals to attach on /srvd or /os.
05/07/97:
- 21:30 Changeset [10005] by
- Rename site-init.el to site-start.el so that "emacs -q --no-site-file" yields a completely clean emacs 19.34b.
- 20:54 Changeset [10004] by
- chgrp mailboxes to mail, necessary for mail delivery on Solaris 2.5.1.
- 20:54 Changeset [10003] by
- Support $ROOT again, since we get run during the install as well as during the update.
- 20:53 Changeset [10002] by
- The alt key now gets a keysym of Meta_L, so add that to mod1 (to get both Meta_L keys) instead of Alt_L.
- 20:52 Changeset [10001] by
- Sync with Solaris 2.5.1
- 20:51 Changeset [10000] by
- Define CPUTYPE for use in driver install scripts. Replace hardcoded sun4m with `uname -m`, which is not quite as good as `uname -i` (which is documented), but which works in a 2.4 miniroot.
- 20:49 Changeset [9999] by
- Sync with Solaris 2.4 and Solaris 2.5.1.
- 20:48 Changeset [9998] by
- Sync with Solaris 2.5.1, commenting out new services
- 20:47 Changeset [9997] by
- Assign the Alt key a keysym of Meta_L, instead of overloading mod1 to contain both Meta_[LR] and Alt_L. This makes it possible for ICCCM- compliant applications to determine what modifier to use for Meta and what to use for Alt (nothing).
- 20:43 Changeset [9996] by
- Mention that named.local won't work on Solaris 2.5.1, and refer to the comment in named.boot.
- 20:42 Changeset [9995] by
- Add a comment describing how to do local records, since named.local won't work on Solaris 2.5.1.
- 10:15 Changeset [9994] by
- Add timed to the standard services not included by Sun.
05/06/97:
- 23:42 Changeset [9993] by
- Use the Solaris 2.5.1 services file instead of an Ultrix services file as a basis. We have two sections at the end, one for standard services not included by Sun (taken from the IRIX 5.3 services file) and a section for Athena additions.
- 22:09 Changeset [9992] by
- Update the build system documentation to reflect the source-8.1 locker and friends.
- 14:19 Changeset [9991] by
- Reconcile the table of contents with the first section heading.
05/05/97:
- 18:55 Changeset [9990] by
- Catalog all the modified operating system files in packs/config and packs/maint.
05/04/97:
- 01:02 Changeset [9989] by
- Add an option -r to update_ws which prevents the post-reboot shell prompt, for remote updates.
- 00:57 Changeset [9988] by
- Prevent questions from mkserv during an update, by making its input come from /dev/null.
- 00:48 Changeset [9987] by
- Use /sbin/machtype instead of /srvd/bin/athena/machtype, since the former is more likely to run if the packs were compiled against a different version of the OS than the miniroot is running. Remove some obsolete copies. Assume boot blocks are in sun4m; "uname -i" doesn't work until we have a 2.5.1 miniroot on the install server.
05/01/97:
- 20:55 Changeset [9986] by
- Revert the last change; auto_update still needs to know why it was invoked.
04/30/97:
- 19:03 Changeset [9985] by
- Garbage collect.
- 18:59 Changeset [9984] by
- Punt installing the Mwm app-defaults; we already handle it on a per-platform basis.
- 18:58 Changeset [9983] by
- Revert last change; we'll nuke Mwm from packs/config instead.
- 18:52 Changeset [9982] by
- We were installing mwm app-defaults both from here and from packs/config. Nuke it from here.
- 14:28 Changeset [9981] by
- Correct typo: s/singed/signed/
- 14:25 Changeset [9980] by
- Add a comment from bert.
- 14:19 Changeset [9979] by
- From bert: Unused; only Imake.tmpl is used.
- 14:19 Changeset [9978] by
- From bert: Configuration files for various client-like things.
- 14:18 Changeset [9977] by
- From bert: Moved assignments out of "if" statement tests.
- 14:18 Changeset [9976] by
- From bert: Use POSIX signal semantics (currently conditionalized) Use getrlimit(RLIMIT_NOFILE,...) instead of getdtablesize() Moved assignments out of "if" statement tests.
- 14:18 Changeset [9975] by
- From bert: We don't need libbsd, so we don't use it.
- 14:17 Changeset [9974] by
- From bert: Added install commands for configuration files.
- 14:17 Changeset [9973] by
- From bert: Made sure that the header body is included only once. Added a prototype for set_env_var() which was added to olc.
- 14:17 Changeset [9972] by
- From bert: Changed the prototype for edit_message to match definition. Removed unused extern declarations of int OLC, OLCR and OLTA.
- 14:17 Changeset [9971] by
- From bert: Made sure that the header body is included only once.
- 14:16 Changeset [9970] by
- From bert: Solaris may be a little braindead, but it doesn't need to include <string.h> twice. =) Changes to support config files (to make olc*, olt* and owl* one binary): * Added #include <olc/incarnate.h> * Added OLC_CONFIG_PATH (used as default if $OLXX_CONFIG is unset) and OLC_CONFIG_EXT. * Replaced (client-dependent) OLC_PROMPT and OLCR_PROMPT with OLC_FALLBACK_PROMPT (which is used only if the configuration file is missing). * Replaced (client-dependent) DEFAULT_CONSULTANT_TITLE with OLC_FALLBACK_TITLE, which is "consultant" for lack of a better option (and is used only if configuration is missing). Removed hardcoded definition of server name (OLC_SERVER) which should have been client-dependent but wasn't. * Removed unused OLCD_PORT. * Removed OLC{,R}_HELP_*, which must now be specified in the config file. * Removed OLC_SERVICE_NAME, which must now be specified in the config file. Incremented the minor version number for the client. (This is only visible to the user; protocol version is separate.) Changed the mailhub from "athena.mit.edu" to "mit.edu".
- 14:15 Changeset [9969] by
- From bert: Header file for incarnate.c, which deals with the "incarnation" of various clients from the same binary.
- 14:14 Changeset [9968] by
- Header file for libcfgfile.a
- 14:14 Changeset [9967] by
- From bert: Moved src/browser/curses/globals.h here.
- 14:14 Changeset [9966] by
- From bert: Merged src/browser/curses/cref.h and src/browser/motif/cref.h and moved them here.
- 14:13 Changeset [9965] by
- From bert: Solaris has vendor Motif, ANSI includes and vfork().
- 14:13 Changeset [9964] by
- From bert: Added support for using the wcl locker to get WCL if we really want to. (This is not the default.) Otherwise, development is harder because non-build machines don't have WCL locally.
- 14:12 Changeset [9963] by
- From bert: Don't explicitly link with -lhesiod. (This will break things if -L/usr/athena/lib didn't happen to be specified by some other library, and in any case, Imakefiles should link with Hesiod explicitly if they need it.) SGIs have vendor Motif libraries.
- 14:11 Changeset [9962] by
- From bert: Added support for "imake -DIsOlta" and "imake -DIsOwl". (This now only makes a difference when compiling the server.) Support linking with $(CONFIGLIB) [libcfgfile.a] Set $(CFLAGS)
- 14:10 Changeset [9961] by
- From bert: Use $(STD_CPP_DEFINES) for compiling too. A dependency on Imake.template should have been for Imake.tmpl
- 14:10 Changeset [9960] by
- From bert: Changes to make the build break when a dependency in a subdirectory fails and make wasn't invoked with "-i" or "-k". (Support for this existed, but didn't work; make would just go on an return with exit status 0.)
- 14:09 Changeset [9959] by
- From bert: Include <unistd.h> even if not on Solaris. Replaced a ludicrously outdated lseek() flag [L_XTND -> SEEK_END]
- 14:09 Changeset [9958] by
- From bert: Removed ancient High C compatibility junk (hardcoded for IBM RT's) Moved assignments out of "if" statement tests.
- 14:09 Changeset [9957] by
- From bert: Removed some hardcoded system external declarations and macros. No, Solaris doesn't need to explicitly include <sys/select.h>. Moved assignments out of "if" statement tests. Removed useless casts to "unsigned" for argument to malloc().
- 14:08 Changeset [9956] by
- From bert: Replaced <sys/time.h> with <time.h> which declares time() on Solaris _POSIX_SOURCE is the wrong answer to any question. (Removed.) Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Also use this to determine if the client is a consulting ("*r") client. Improved an error message to reflect that we may be using Hesiod in addition to local configuration files. Replaced "kinit" with "renew" in error messages. edit_message() now admits to propagating the return from the editor. Initialize random numbers on first call to happy_message() Moved assignments out of "if" statement tests.
- 14:07 Changeset [9955] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Moved assignments out of "if" statement tests.
- 14:06 Changeset [9954] by
- From bert: Removed hardcoded declaration of ioctl() and #include'd unistd.h instead Moved assignments out of "if" statement tests. Replaced a 0 with a NULL in char* assignment.
- 14:06 Changeset [9953] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Also use this to determine if the client has "hours" and whether it's a consulting ("*r") client. Moved assignments out of "if" statement tests.
- 14:05 Changeset [9952] by
- From bert: Moved assignments out of "if" statement tests.
- 14:05 Changeset [9951] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Also use this to determine if the client is a consulting ("*r") client.
- 14:04 Changeset [9950] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values.
- 14:01 Changeset [9949] by
- From bert: Use client "incarnation" data from incarnate.c to determine if the client is a consulting ("*r") client.
- 14:00 Changeset [9948] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Also use this to determine if the client has "hours" and whether it's a consulting ("*r") client. Removed extern declaration for HELP_FILE, HELP_DIR, and HELP_EXT, which are no longer used. Moved assignments out of "if" statement tests.
- 13:59 Changeset [9947] by
- From bert: Removed ancient High C compatibility junk (hardcoded for IBM RT's) Made a function local variable whose address was being returned (oops!) be static, so it's guaranteed to still be there when the execution isn't.
- 13:59 Changeset [9946] by
- Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Also use this to determine if the client has "hours" and whether it's a consulting ("*r") client.
- 13:58 Changeset [9945] by
- From bert: Changes to support config files (to make olc*, olt* and owl* one binary): * Downcased the default instance name, so it could be used as a configuration file name. * Added code to support "incarnating" a particular client from a configuration file, found in path given by OLXX_PATH or cmd line. [currently using "olc" (or "olta" or "owl") config files.] * Added new command line option ("-C") to deal with specifying the search path for the config files. Decreased ugliness: * Added comments to clarify meanings of command-line options. * Documented all existing options in the usage message. * Changed a hardcoded "SLOC" to the appropriate symbol from olc/olc.h Added code to fall back and use a default if Hesiod is down. Moved assignments out of "if" statement tests. (!!! NOTE: This code still needs much debugging etc before becoming functional. !!!)
- 13:57 Changeset [9944] by
- From bert: Fixed the fact that the Imakefile was trying to install oreplay instead of osend... =)
- 13:56 Changeset [9943] by
- From bert: Replaced <strings.h> with <string.h> Changes to support config files (to make olc*, olt* and owl* one binary): * Downcased the default instance name, so it could be used as a configuration file name. * Added code to support "incarnating" a particular client from a configuration file, found in path given by OLXX_PATH or cmd line. [currently using "olc" (or "olta" or "owl") config files.] * Added new command line option ("-C") to deal with specifying the search path for the config files. Decreased ugliness: * Added comments to clarify meanings of command-line options. * Documented all existing options in the usage message. * Changed a hardcoded "SLOC" to the appropriate symbol from olc/olc.h Added a flag ("-p") to make oreplay think it's oshow (analogous to "-l" for olist). Added code to fall back and use a default if Hesiod is down. Moved assignments out of "if" statement tests.
- 13:55 Changeset [9942] by
- From bert: Support for configuration files (through libcfgfile). Changed rules for "olist" and "oshow" to depend on "oreplay", so they will get re-hardlinked in the build tree when oreplay is updated.
- 13:50 Changeset [9941] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Removed an "if" with a condition which is necessarily true at that point in the code.
- 13:49 Changeset [9940] by
- From bert: Changes to support config files (to make olc*, olt* and owl* one binary): * Replaced the OLC and OLCR command table initializers with a single "template" table which can accomodate differences between various olc-like client binaries. * Removed extern declarations for int OLC, HELP_FILE, HELP_DIR, HELP_EXT. In main(): * Removed "magic" comparison to determine from argv[0] if this is supposed to be a user or a consulting client. * Removed hardcoding of HELP_FILE, HELP_DIR and HELP_EXT. Added code to support "incarnating" a particular client from its configuration file, found in path given by OLXX_PATH or cmd line. * Added new command line options to deal with specifying client name without argv[0] hacking ("-name") specifying the search path for the config file ("-config"). In build_command_table(): * This function was added to build a command table at run-time, based on client "incarnation" parameters. In olc_init(): * Updated the copyright message. * OLC (in an irrelevant context) changed to 0 * Replaced an #if to determine if "answers" and "hours" are available with a run-time check using client_has_answers() or client_has_hours(). Decreased ugliness: Changed all putenv()/setenv() alternatives to use a single function call instead (to set_env_var(), which was also added). Removed a bug which made flags with no arguments cause unusual behavior or segmentation faults.
- 13:46 Changeset [9939] by
- From bert: Support for configuration files (through libcfgfile).
- 13:46 Changeset [9938] by
- From bert: Added <fcntl.h> Moved assignments out of "if" statement tests. Removed multiplication by sizeof(char).
- 13:46 Changeset [9937] by
- From bert: Removed type casts for signal handler assignments Fixed argument types for X (Motif) callbacks to match callback type Moved assignments out of "if" statement tests. Removed multiplication by sizeof(char).
- 13:45 Changeset [9936] by
- From bert: Changes to support config files (to make olc*, olt* and owl* one binary): * Removed extern declarations for int OLC and OLCR. In main(): * Removed "magic" comparison to determine from argv[0] if this is supposed to be a user or a consulting client. * Added code to support "incarnating" a particular client from its configuration file, found in path given by OLXX_PATH or cmd line. * Added new command line options to deal with: - specifying client name without argv[0] hacking ("-name") - specifying the search path for the config file ("-config") In olc_init(): * OLC (in an irrelevant context) changed to 0 [NOTE: xolc is still not capable of being an OLTA/OWL client, since things like "OLC" and "consultant" are hardcoded in too many places, but the libraries now require configuration file support in all clients.] Decreased ugliness: Changed all putenv()/setenv() alternatives to use a single function call instead (to set_env_var(), which was also added). Removed a bug which made flags with no arguments cause unusual behavior or segmentation faults.
- 13:42 Changeset [9935] by
- From bert: Fixed argument types for X (Motif) callbacks to match callback type
- 13:39 Changeset [9934] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values.
- 13:38 Changeset [9933] by
- From bert: Added <fcntl.h> Moved assignments out of "if" statement tests.
- 13:37 Changeset [9932] by
- From bert: Removed type casts for signal handler assignments Cleaned up a little Moved assignments out of "if" statement tests.
- 13:36 Changeset [9931] by
- From bert: Changed <sys/errno.h> to <errno.h> to get errno on Solaris Moved assignments out of "if" statement tests. Fixed some typos (error return test checked for 1, not -1)
- 13:35 Changeset [9930] by
- From bert: Removed explicit declaration of system functions/variables. Removed explicit declaration of DaemonHost, which is unnecessary. Removed some unnecessary typecasts. Fixed an accidentally found bug (instead of using Hesiod and falling back on /etc/services, we tried /etc/services twice). Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Moved assignments out of "if" statement tests.
- 13:35 Changeset [9929] by
- From bert: Use client "incarnation" data from incarnate.c in place of previously hardcoded values. Fixed an uninitialized-variable bug that broke "-server" flag Improved an error message Moved assignments out of "if" statement tests. eplaced a 0 with a NULL in char* comparison.
- 13:34 Changeset [9928] by
- From bert: Configuration support code for the OLC client library
- 13:34 Changeset [9927] by
- From bert: Moved assignments out of "if" statement tests.
- 13:33 Changeset [9926] by
- From bert: Added incarnate.c (which deals with olc/olta/owl/... init) to libclient.a
- 13:32 Changeset [9925] by
- From bert: Don't build the "osend" directory (which is horribly mangled).
- 13:32 Changeset [9924] by
- From bert: Code to read and parse config files
- 13:32 Changeset [9923] by
- From bert: Rules for making libcfgfile.a
- 13:32 Changeset [9922] by
- From bert: Moved cref.h to include/browser (and merged with the curses one) Moved assignments out of "if" statement tests.
- 13:31 Changeset [9921] by
- From bert: Moved cref.h to include/browser (and merged with the curses one). Some symbols changed names in the process. Fixed a missing #include for <Xm/MessageB.h> Moved assignments out of "if" statement tests. Grouped * with variables (not types) in variable declarations. Removed multiplication by sizeof(char).
- 13:30 Changeset [9920] by
- From bert: moved to include/browser
- 13:29 Changeset [9919] by
- From bert: Moved cref.h and globals.h to include/browser Moved assignments out of "if" statement tests.
- 13:29 Changeset [9918] by
- From bert: Moved cref.h and globals.h to include/browser Removed type casts for signal handler assignments Changed ancient Curses compatibility macro "crmode()" to "cbreak()", which is equivalent and generally better-known. Prompts are no longer hardcoded.
- 13:29 Changeset [9917] by
- From bert: Moved cref.h and globals.h to include/browser Changed init_globals() [which is still called before the command line is parsed], so we can initialize variables from the config file using read_config() *after* we read the command line.
- 13:28 Changeset [9916] by
- From bert: Moved cref.h and globals.h to include/browser Changed ancient Curses compatibility macro "crmode()" to "cbreak()", which is equivalent and generally better-known. Header messages are no longer hardcoded. Fixed a bug which made "more" very confused when running under SYSV curses, due to differences in tty modes. Moved assignments out of "if" statement tests.
- 13:27 Changeset [9915] by
- From bert: Moved cref.h and globals.h to include/browser Fixed an error in comments describing err_abort() and err_exit() Moved assignments out of "if" statement tests. Removed some "stylistically incorrect" spaces around parens.
- 13:26 Changeset [9914] by
- From bert: Moved cref.h and globals.h to include/browser Replaced *argv with argv[0], which is more understandable in context Added '-C' and '-N' flags for specifying a config file path/name '-f' flag is more convoluted if '-r' wasn't given, in order for the config files to DTRT Prompts are no longer hardcoded.
- 13:25 Changeset [9913] by
- From bert: Moved cref.h and globals.h to include/browser. Fixed a bug which made "more" very confused when running under SYSV curses, due to differences in tty modes. Moved assignments out of "if" statement tests. Re-indented a block which was 2 spaces from Gnu-style-correctness.
- 13:24 Changeset [9912] by
- Moved to include/browser
- 13:23 Changeset [9911] by
- From bert: Support for configuration files (through libcfgfile). Support for all o*_browser binaries being the same.
- 13:23 Changeset [9910] by
- From bert: Removed support for subdirectories which no longer exist (wcl and libMu). We can assume the release will build them elsewhere. Added cfgfile subdirectory. Use "$(MAKE)" instead of "make". Rewrote rules to make TAGS* files.
- 13:20 Changeset [9909] by
- From bert: A description of the layout of the server installation.
- 13:20 Changeset [9908] by
- From bert: Don't make links to TAGS files in the source tree.
- 05:00 Changeset [9907] by
- Check for space before update on SGI.
- 04:32 Changeset [9906] by
- Improve the awk code for reading /etc/athena/version, making it more readable and producing possibly more meaningful results. (Worth noting: the existing update scripts can't ever produce actual updates to the same version, so maybe the original intent was to count the number of times mkserv was run. In that case, further tweaking is in order.)
- 00:44 Changeset [9905] by
- auto_update doesn't need an argument any more.
- 00:42 Changeset [9904] by
- Garbage-collect some code that doesn't seem to have any effect any more (it waited 20 seconds under circumstances that don't happen any more).
04/29/97:
- 15:11 Changeset [9903] by
- Put Athena buttons in the middle of the window.
- 10:45 Changeset [9902] by
- Reconcile with Solaris 2.5.1 version. Again, this is just the stock version with an entry added for afs as 105.
- 10:43 Changeset [9901] by
- Athena changes: make all the audio devices mode 0666.
- 10:40 Changeset [9900] by
- Check in the stock 2.5.1 minor_perm for reference; a commit with the Athena changes will follow shortly. (The changes are complicated enough that we ought to be able to do a diff.)
- 10:38 Changeset [9899] by
- Reconcile with Solaris 2.5.1 version; again, this is just the stock version with two lines added at the end for the appletalk devices.
04/26/97:
- 01:06 Changeset [9898] by
- Fix typo.
- 01:05 Changeset [9897] by
- Fix dumb mistake. Sigh. (Install desync as desync, not as "rep".)
04/24/97:
- 23:56 Changeset [9896] by
- Increment ngroups earlier so we don't miss the last group.
- 18:19 Changeset [9895] by
- Run suncc instead of cc, since cc can't find itself correctly the way we run it.
04/23/97:
- 15:38 Changeset [9894] by
- Go back to setting require-final-newline.
Note: See TracTimeline
for information about the timeline view.