2004-09-05 12:13  Stephan Kulow <coolo@kde.org>

	* client/local.cpp, client/remote.cpp, daemon/main.cpp,
	  daemon/serve.cpp: don't make up exit codes

2004-09-04 22:26  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: less debug output

2004-09-04 22:16  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: avoid a race between select and wait - loosing
	  data on the road

2004-09-04 19:18  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/local.cpp, client/main.cpp,
	  client/remote.cpp, mon/monitor.cc, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: tell the scheduler about
	  local jobs in a child proces that is started right away to avoid
	  any kind of delay (there is nothing the client waits on anyway)

	  This speeds up calling configure in kdelibs by around 20% in my
	  test system Said differently: This makes the icecream overhead
	  for configure go from 50% to 11% (for a ~100% for teambuilder)

2004-09-04 17:30  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, client/remote.cpp, daemon/main.cpp, mon/job.h,
	  mon/joblistview.cc, mon/monitor.cc, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: transfer min_flt too and
	  summarize it in the scheduler log so I can figure what memory
	  limits gcc fails on

2004-09-04 10:14  Stephan Kulow <coolo@kde.org>

	* client/local.cpp: another error not caused by the user but by the
	  network

2004-09-03 10:55  Stephan Kulow <coolo@kde.org>

	* TODO, client/main.cpp, client/remote.cpp, daemon/main.cpp,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp: one
	  class less to destruct

2004-09-02 11:07  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, client/remote.cpp, services/comm.cpp,
	  services/comm.h: for whatever reason these unblocking connects
	  generate more problems than they are worth on most parts of the
	  code - so use them only for the parts we need them

2004-09-02 08:11  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: these short time out are causing a lot of
	  trouble. And as soon as the trouble starts, it's getting worse
	  and worse ;(

2004-09-01 18:28  David Faure <faure@kde.org>

	* daemon/environment.cpp: More informative than just the assert.
	  The problem was solved by "rm -rf /tmp/icecc-envs" btw.

2004-09-01 17:54  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: --enable-final rules the world

2004-09-01 11:57  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: do not sleep just because we're overloaded

2004-09-01 10:14  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: I think I know now who eats my child pids
	  (never hack in keynotes :) No child pids -> no job done messages

2004-08-31 18:54  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/logging.h: trying to make sense of the
	  log files

2004-08-31 18:23  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: mean, mean regression: don't hang forever (and
	  eat client connections) While I looked at the code I noticed a
	  TODO to do

2004-08-31 10:41  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: * ignore the load of the submitter (it's
	  all his fault) * increase grace interval

2004-08-31 09:52  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: don't use log_error for things caused by the
	  network not by the user

2004-08-31 01:58  Thomas Capricelli <orzel@kde.org>

	* services/logging.h: compile fix

2004-08-30 22:09  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: prefer servers with fewer load

2004-08-30 22:01  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: less debug output

2004-08-30 22:00  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp, daemon/load.cpp, daemon/load.h,
	  daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: try to calculate the load a bit more
	  accurate so we take jobs that run niced in account and only
	  substract our own load.

2004-08-30 17:16  Stephan Kulow <coolo@kde.org>

	* daemon/: Makefile.am, findmyself.cpp, findmyself.h, main.cpp:
	  don't support this strange -w feature any longer (doesn't work
	  with NFS, etc, etc)

2004-08-30 16:06  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: if the client exited, we give the job a
	  different state. I see lately the effect, that the daemon misses
	  to send the jobdone message, so this will help to indicate the
	  problem

2004-08-30 15:56  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: put less output in the default debug

2004-08-30 12:16  Stephan Kulow <coolo@kde.org>

	* client/cpp.cpp, client/remote.cpp, client/util.cpp,
	  daemon/environment.cpp, daemon/findmyself.cpp, daemon/main.cpp,
	  daemon/serve.cpp, daemon/workit.cpp, services/Makefile.am,
	  services/comm.cpp, services/job.cpp, services/receiver.cpp,
	  services/scheduler.cpp, services/submitter.cpp: use log_perror
	  everywhere

2004-08-30 11:01  Stephan Kulow <coolo@kde.org>

	* suse/sysconfig.icecream: fixing comments

2004-08-30 10:15  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: the daemon already did a DNS lookup, so spare
	  this for the client and return the IP right away

2004-08-30 10:03  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: before I forget

2004-08-30 09:50  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: tricky. If you make -j, it goes in the client
	  dir to create icecc from there while another make is busy doing
	  it from the toplevel -> ugly file corruptions

2004-08-29 22:27  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: really remove debug output

2004-08-29 22:25  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: removing debug output as I verified the
	  10s timeout works

2004-08-29 22:19  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, logging.h: trying to support shorter time
	  outs in connect

2004-08-29 21:12  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: firewalls are mean and crash the client on
	  timeout

2004-08-29 18:22  Stephan Kulow <coolo@kde.org>

	* suse/init.icecream: this doesn't seem to work

2004-08-29 18:10  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, services/scheduler.cpp: improving debug output

2004-08-29 17:32  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: fixing mem leak on error case

2004-08-29 17:25  Stephan Kulow <coolo@kde.org>

	* TODO: run as user is dead (maybe we will later require a setuid
	  wrapper to be installed, but that's another story) adding API
	  cleanup as release critical

2004-08-29 17:23  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: fixing (rather unlikely) memory corruption in
	  the client found by Andreas Schwab 

2004-08-29 16:12  Stephan Kulow <coolo@kde.org>

	* mon/: joblistview.cc, starview.cc: fixing compile (with final)

2004-08-29 14:34  Stephan Kulow <coolo@kde.org>

	* suse/: init.icecream, sysconfig.icecream: create the log file as
	  user icecream before starting the scheduler as user

2004-08-29 14:01  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp, suse/init.icecream,
	  suse/sysconfig.icecream: support more features of the daemon
	  through sysconfig

2004-08-29 13:41  Stephan Kulow <coolo@kde.org>

	* client/create-env: clean up

2004-08-29 13:29  André Wöbbeking <Woebbeking@web.de>

	* mon/: detailedhostview.h, ganttstatusview.cc, ganttstatusview.h,
	  hostview.cc, hostview.h, mon-kde.cc, mon-kde.h, monitor.cc,
	  monitor.h, starview.cc, starview.h, statusview.cc, statusview.h,
	  summaryview.cc, summaryview.h: cleanedup the includes

2004-08-29 00:53  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: rename the option to -r for consistency

2004-08-28 19:05  André Wöbbeking <Woebbeking@web.de>

	* mon/mon-kde.cc: fixed accelerator clash

2004-08-28 19:01  André Wöbbeking <Woebbeking@web.de>

	* mon/joblistview.cc: sort by job id (descending)

2004-08-28 18:46  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Maybe this looks better.

2004-08-28 15:45  André Wöbbeking <Woebbeking@web.de>

	* mon/detailedhostview.cc: David is right, outgoing jobs can also
	  be incoming jobs

2004-08-28 13:37  André Wöbbeking <Woebbeking@web.de>

	* mon/starview.cc: skip hosts without name or which are offline

2004-08-28 12:02  André Wöbbeking <Woebbeking@web.de>

	* mon/joblistview.cc: JobListViewItem can be created for already
	  finished jobs so expire them too

2004-08-28 11:53  André Wöbbeking <Woebbeking@web.de>

	* mon/: Makefile.am, detailedhostview.cc, detailedhostview.h,
	  joblistview.cc, joblistview.h, listview.cc, listview.h: factored
	  out the list view from ListStatusView into JobListView

2004-08-28 10:50  André Wöbbeking <Woebbeking@web.de>

	* mon/: hostlistview.cc, hostlistview.h: Reimplemented clear() to
	  get rid of dangling pointers. Even if it's not called yet.

2004-08-28 09:55  Stephan Kulow <coolo@kde.org>

	* TODO, daemon/environment.cpp, daemon/environment.h,
	  daemon/main.cpp, daemon/serve.cpp, daemon/serve.h: making it
	  _way_ harder to accuse me of having caused a rootkit: * changed
	  the -u option to a -r option (still useful for homogenous
	  networks) * added -u <nobody_uid> (defaults to 65534) that it
	  runs all processes under   that handle any client data.

	  So now the compilations runs as nobody in a chroot environment
	  together with the "rootkit", this should make a sniffing attack
	  quite complex, but of course it still allows DoS (even though the
	  memory usage is limited and the CPU usage is niced).

2004-08-27 23:37  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: More commands.

2004-08-27 21:50  André Wöbbeking <Woebbeking@web.de>

	* mon/: listview.cc, listview.h: reimplement clear() to get rid of
	  dangling pointers

2004-08-27 21:42  André Wöbbeking <Woebbeking@web.de>

	* mon/listview.cc: only start the timer if it's not running

2004-08-27 20:09  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Ui.  Ugly bug.

2004-08-27 19:56  André Wöbbeking <Woebbeking@web.de>

	* mon/: detailedhostview.cc, listview.cc, listview.h: don't remove
	  finished jobs at once.

	  don't know if it's bug free because icecream's scheduler is down
	  again :-(

2004-08-27 18:34  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: add some consistency check

2004-08-27 17:24  André Wöbbeking <Woebbeking@web.de>

	* mon/: detailedhostview.cc, listview.cc, listview.h: hide useless
	  columns

2004-08-27 16:59  André Wöbbeking <Woebbeking@web.de>

	* mon/: listview.cc, listview.h: don't show the complete file path

2004-08-27 16:40  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: fixing layout

2004-08-27 16:32  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: seems, monitors aren't removed cleanly
	  from fd2chan

2004-08-27 14:39  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: work correctly

2004-08-27 14:35  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying to improve ways to debug

2004-08-27 13:24  André Wöbbeking <Woebbeking@web.de>

	* mon/: hostinfo.cc, hostinfo.h, hostlistview.cc: coolo wants it
	  coolo gets it: the load of the hosts

2004-08-27 13:05  André Wöbbeking <Woebbeking@web.de>

	* mon/hostlistview.cc: the state column is useless (a host is
	  online or not shown)

2004-08-27 12:26  André Wöbbeking <Woebbeking@web.de>

	* mon/: hostlistview.cc, hostlistview.h, listview.cc, listview.h:
	  hey, it's a KDE app so use KDE classes

2004-08-27 12:15  André Wöbbeking <Woebbeking@web.de>

	* mon/detailedhostview.cc: some labels added

2004-08-26 23:34  André Wöbbeking <Woebbeking@web.de>

	* mon/: detailedhostview.cc, detailedhostview.h, hostlistview.cc,
	  hostlistview.h, icemonui.rc, listview.cc, listview.h, mon-kde.cc,
	  mon-kde.h, Makefile.am: new view: DetailedHostView. Cleanups will
	  follow tomorrow.

2004-08-26 20:24  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: yes, the network is flacky and icecream can
	  handle it

2004-08-26 20:24  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: give better hint

2004-08-26 15:02  Simon Hausmann <hausmann@kde.org>

	* mon/doc/: .cvsignore, icemon/.cvsignore: - I admit, I'm building
	  icecream with srcdir==builddir :)

2004-08-26 14:57  André Wöbbeking <Woebbeking@web.de>

	* mon/listview.cc: - use MB, KB for size in/out - right align all
	  numbers - setAllColumnsShowFocus(true)

2004-08-26 11:07  Simon Hausmann <hausmann@kde.org>

	* mon/mon-kde.cc: - a slightly prettier window title :)

2004-08-26 03:07  Daniel Molkentin <molkentin@kde.org>

	* mon/doc/icemon/index.docbook: use entities, plus I wasn't aware
	  that frerich works for suse ;)

2004-08-26 02:20  Cornelius Schumacher <schumacher@kde.org>

	* mon/: Makefile.am, doc/.cvsignore, doc/Makefile.am,
	  doc/icemon/.cvsignore, doc/icemon/Makefile.am,
	  doc/icemon/index.docbook: Add icemon manual.

2004-08-26 02:18  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Make halos slightly bigger.

2004-08-25 22:01  Stephan Binner <binner@kde.org>

	* mon/: ganttstatusview.cc, hostview.cc, mon-kde.cc, starview.cc:
	  CVS_SILENT i18n style guide fixes

2004-08-25 20:07  Simon Hausmann <hausmann@kde.org>

	* TODO: - this one is done

2004-08-25 19:58  Stephan Kulow <coolo@kde.org>

	* mon/monitor.cc: forget jobs after some time - the list view would
	  need to do some code on it's own, but at least you can clean up
	  the history now by changing views

2004-08-25 18:17  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: try some times more if bind fails - not forever
	  though

2004-08-25 18:06  Simon Hausmann <hausmann@kde.org>

	* client/local.cpp: - when ICECC_VERISON is not set we want to pick
	  the same compiler as	 create-env, the one in /usr/bin

2004-08-25 14:37  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying to fix client fairness again

2004-08-24 23:27  Cornelius Schumacher <schumacher@kde.org>

	* mon/: job.h, starview.cc, starview.h: Color halo in client color.
	  Needs more testing with multi-job machines.

2004-08-24 22:19  Cornelius Schumacher <schumacher@kde.org>

	* mon/mon-kde.cc: Correctly remember view and view menu. Patch by
	  bruggie.

2004-08-24 21:25  Simon Hausmann <hausmann@kde.org>

	* client/: client.h, cpp.cpp, create-env: - when calling the
	  preprocessor find gcc/g++ from $PATH - added support for
	  --respect-path to create-env, which picks gcc/g++ from   $PATH
	  instead of /usr/bin, allowing the use of icecream with custom
	  gcc installations

2004-08-24 10:05  Stephan Kulow <coolo@kde.org>

	* TODO: I already said that I'm not sure if it's worth it, no?

2004-08-24 07:53  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: another protection against stupid KDE
	  developers

2004-08-24 04:20  Nikolas Zimmermann <wildfox@kde.org>

	* client/: main.cpp, remote.cpp: Use 127.0.0.1 instead of localhost
	  - we sat nearly two hours in front of Max Howell's laptop
	  trying to figure out why the heck icecream didn't work..  -> the
	  hosts file somehow contained an entry with a different    ip for
	  'localhost'; this would avoid such a problem in future...

2004-08-24 00:06  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-08-23 23:10  Stephan Kulow <coolo@kde.org>

	* client/create-env: the -a is of no importance now that we care
	  for symlinks

2004-08-23 22:30  Frerich Raabe <raabe@kde.org>

	* Makefile.cvs: - Trying to approximate 'cp -a' as it exists on
	  Linux, so that we can run   gmake -f Makefile.cvs here on FreeBSD
	  boxen, too.

2004-08-23 22:03  Frerich Raabe <raabe@kde.org>

	* mon/IDEAS.poolview: - So that I don't forget about it

2004-08-23 21:56  Stephan Kulow <coolo@kde.org>

	* client/create-env: avoid starting /

2004-08-23 21:15  Stephan Kulow <coolo@kde.org>

	* client/create-env: try to solve symlinks more globally

2004-08-23 16:35  Stephan Kulow <coolo@kde.org>

	* TODO: I was sure it's commited

2004-08-23 16:34  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: slashes - you can never have enough of it

2004-08-23 16:18  Stephan Kulow <coolo@kde.org>

	* client/create-env: trying to support gentoo and debian

2004-08-23 16:07  Frerich Raabe <raabe@kde.org>

	* mon/starview.cc: - Don't update the canvas for every single new
	  node item, but only once   after all items have been added.

2004-08-23 14:29  Stephan Kulow <coolo@kde.org>

	* TODO, configure.in.mid, daemon/environment.cpp: don't rely on
	  $PATH

2004-08-23 13:52  Stephan Kulow <coolo@kde.org>

	* TODO: investigated again what it would mean to build as non-root
	  and put missing things in categories

2004-08-23 13:33  Frerich Raabe <raabe@kde.org>

	* daemon/main.cpp: - Typo fix :-)

2004-08-23 11:52  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: * don't give out 32bit compilers to
	  hammers (kernel 2.6.8 rules ;( * don't bother me with chroot
	  problems when I said I allow them (old daemons)

2004-08-23 11:23  Stephan Kulow <coolo@kde.org>

	* daemon/: workit.cpp, workit.h: f*** - 64GB of free memory are
	  just too much for the little ints ;(

2004-08-22 17:47  Dirk Mueller <mueller@kde.org>

	* services/scheduler.cpp: d'oh

2004-08-22 17:37  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Don't allow non-chroot-able daemons to
	  connect without cmdline option.  And don't give them out as
	  compile hosts.

2004-08-22 17:20  Dirk Mueller <mueller@kde.org>

	* services/scheduler.cpp: prepare to be able to reject clients that
	  cannot do chroot - they will always compile on localhost. add
	  command line option that disables the behaviour for very
	  homogenous compiler setups

2004-08-22 16:58  Dirk Mueller <mueller@kde.org>

	* services/: comm.cpp, comm.h: add the information if the daemon is
	  running as root (for chroot/environment support) to the login
	  message, so that the scheduler can disconnect clients that do not
	  accept chrooted compiles

2004-08-22 16:34  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: this wasn't meant to go into CVS

2004-08-22 16:33  Michael Matz <matz@kde.org>

	* client/create-env: Gentoo has a funny wrapper.  Perhaps this
	  works for them.  Perhaps not OTOH.

2004-08-22 15:10  Dirk Mueller <mueller@kde.org>

	* client/safeguard.cpp: I do believe the safeguard code works - so
	  don't bother me with its debug output

2004-08-22 14:59  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Show dashed lines in correct color from the
	  beginning.

2004-08-22 14:50  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: More info.

2004-08-22 14:36  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Better.

2004-08-22 14:09  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: More debug possibilities.

2004-08-22 14:05  Stephan Kulow <coolo@kde.org>

	* mon/: hostinfo.cc, starview.cc: don't invent clients

2004-08-22 13:00  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: don't let Dirk f*** the system

2004-08-22 11:02  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Text based control
	  stuff.  Long live telnet.

2004-08-21 19:14  Michael Matz <matz@kde.org>

	* TODO: Done.

2004-08-21 19:11  Stephan Kulow <coolo@kde.org>

	* daemon/: environment.cpp, main.cpp: send stats also while
	  retrieving and unpacking environments to avoid falling out of the
	  network

2004-08-21 19:09  Michael Matz <matz@kde.org>

	* services/comm.cpp: Not really tested, but anyway.  The lunch is
	  starting soon.

2004-08-21 18:49  Michael Matz <matz@kde.org>

	* TODO: Bloody social event.

2004-08-21 18:15  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: don't bind to random port if there is already a
	  iceccd running. There are just too many conflicts

2004-08-21 17:06  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: don't crash

2004-08-21 16:51  Stephan Kulow <coolo@kde.org>

	* TODO: Martin's bug

2004-08-21 16:51  Frerich Raabe <raabe@kde.org>

	* mon/starview.cc: - comment out the assertion that the hostname is
	  empty. I have no idea why   this happens, and don't want to debug
	  this right now, but having icemon   bail out everytime i switch
	  to the star view is annoying.

2004-08-21 16:45  Michael Matz <matz@kde.org>

	* TODO: Chris is paranoid

2004-08-21 15:48  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: don't want to brag, but with 100 clients
	  it's quite common that there are undone jobs in the queue

2004-08-21 15:15  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: enough faking

2004-08-21 15:03  Frerich Raabe <raabe@kde.org>

	* mon/ganttstatusview.cc: - Draw the node names in bold letters so
	  it's easier to see which box has   which color. Now I know it's
	  the damn 'linux' node which has the same color   as mine. :-)

2004-08-21 13:41  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: finishing what Scott started ages ago
	  (and never tested :): give i386 environments out to i686 and
	  x86_64 hosts

2004-08-21 11:47  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: Success? Aja

2004-08-21 11:36  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: We seem to be unable to recover why this
	  test was there exactly, and when it was still valid.	Currently
	  it doesn't test a fishy condition, but is fishy itself.

2004-08-21 11:14  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: don't fool dirk - icecc++ is only my testing
	  symlink

2004-08-21 11:00  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: ah, I forgot to commit that - funny that
	  they got the problem just today :)

2004-08-21 10:58  Stephan Kulow <coolo@kde.org>

	* TODO: I guess, there are about 3 people wanting this feature

2004-08-21 00:50  Daniel Molkentin <molkentin@kde.org>

	* daemon/main.cpp: - allow for specifying an env basedir - disallow
	  running icecream as user by default. coolo said it's not
	  supported and it creates problem here at the cluster. The proper
	  fix would probably be to block non-chrooted daemons at the
	  scheduler.

2004-08-20 18:19  Stephan Kulow <coolo@kde.org>

	* TODO: Matz warms up old dreams

2004-08-19 14:48  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: my "I give up" commit was a bit too brute
	  force. Still the same reasoning leads to this commit: remove the
	  channel from all jobs before it's deleted

2004-08-19 09:24  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: avoid warnings from valgrind

2004-08-18 21:08  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: some more detailed debug output

2004-08-18 20:57  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: there are still cases where the client
	  exists before all jobs are done. This leads to a dangling
	  job->channel and it's just not worth the couple of cycles we
	  spare

2004-08-18 16:07  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: STL is just different - that's all

2004-08-18 12:16  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: try harder not to crash in strange
	  circumstances

2004-08-17 18:00  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Either remove the comment, or don't
	  remove the function call it is talking about.  But I like it
	  better this way.

2004-08-17 13:49  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: only fake with speeds when there is some
	  history

2004-08-17 13:46  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: handle --param and make sure we don't mix the
	  order of command line options in case we find a second input file

2004-08-17 01:06  Scott Wheeler <wheeler@kde.org>

	* services/scheduler.cpp: build

2004-08-16 21:29  Stephan Kulow <coolo@kde.org>

	* mon/monitor.cc: $USE_SCHEDULER doesn't need a netname

2004-08-16 18:41  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: if it doesn't have a speed yet, take
	  whatever you can get

2004-08-16 18:34  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying to improve the statistics

2004-08-16 18:12  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: if you can't access the ICECC environment, don't
	  bother the scheduler

2004-08-16 17:25  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Smooth out the speed measurement history.
	  It compiles.  No idea if it works as well as the perl script ;)

2004-08-16 16:07  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, logging.cpp, logging.h: if get_msg() times
	  out, it's quite interesting to know what it waited for: backtrace
	  might help

2004-08-16 16:06  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: -frepo creates local files

2004-08-16 16:05  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: give the preprocessor some time

2004-08-16 09:17  Stephan Kulow <coolo@kde.org>

	* TODO: noticed something ;)

2004-08-15 22:00  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, daemon/environment.cpp, daemon/environment.h,
	  daemon/main.cpp, services/comm.cpp, services/comm.h: only package
	  the native directory when it's requested by the client

2004-08-15 16:23  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, services/job.cpp, services/scheduler.cpp: some
	  more statistics fishing

2004-08-15 13:37  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/scheduler.cpp: didn't mean to commit
	  this debug

2004-08-15 13:33  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: I hope this cures the fd leak I added

2004-08-15 11:37  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/local.cpp, client/remote.cpp,
	  daemon/main.cpp, services/comm.h, services/scheduler.cpp: collect
	  speed infos also for local jobs. If you're only compiling from
	  your old laptop, you want the scheduler to know it's slow That
	  local jobs also count the preprocessor in it's used time, causes
	  some bias. But as the builtin preprocessor is damn quick, I doubt
	  it weights a lot

2004-08-15 09:19  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: don't write to stderr, but catch the error in
	  the log file

2004-08-14 20:01  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: better error message

2004-08-14 19:37  Stephan Kulow <coolo@kde.org>

	* daemon/: serve.cpp, workit.cpp: I guess, we only had one place
	  where a correct calculation of the time was necessary. Too bad
	  that one was bogus ;(

2004-08-14 19:29  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: catching some failing syscalls. I still get
	  some weired lock ups ;(

2004-08-14 18:26  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: Michael, that code was filling the log files
	  on ia64. Not sure it's the most elegant solution to fix it, but
	  the warning is gone

2004-08-14 13:51  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: this is bad - the real exit code was only in a
	  local variable, the return value was the status of the
	  preprocessor ;(

2004-08-13 18:13  Michael Matz <matz@kde.org>

	* client/cpp.cpp: More refactoring.

2004-08-13 17:58  Michael Matz <matz@kde.org>

	* client/remote.cpp: Make the multiple job case also recognize when
	  to build locally.  This all has many nearly duplicated code paths
	  handling slightly different conditions.  Factor it a bit at
	  least.

2004-08-13 17:16  Michael Matz <matz@kde.org>

	* client/: client.h, cpp.cpp, remote.cpp: Be a bit more paranoid
	  with file descriptors.  Perhaps it solves the funny hangs of cc1,
	  or makes them more reproducable.

2004-08-13 15:34  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, scheduler.cpp: * reducing a bit of debug
	  (100M in a night is just too much for my little $HOME :) * after
	  doing some linear fitting I found some factors to not prefer -g
	  jobs	(hunting further factors)

2004-08-12 21:56  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, services/comm.cpp, services/comm.h,
	  services/job.cpp, services/job.h, services/scheduler.cpp:
	  transfer some important flags with the job, so that the scheduler
	  has a chance to split between non debug build and debug builds
	  (that generate way more bytes in the same time). Now we can start
	  to collect data ;)

2004-08-12 17:48  Michael Matz <matz@kde.org>

	* daemon/workit.cpp: Ensure an invariant which bit us.

2004-08-12 17:42  Michael Matz <matz@kde.org>

	* daemon/: main.cpp, serve.cpp: 0 is a valid fd.

2004-08-12 16:28  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: no comment

2004-08-12 15:38  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: don't create busy loops while waiting for local
	  jobs to finish

2004-08-12 14:47  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: as a matter of fact, now that the
	  compiler is given room for garbage collection, the overhead price
	  isn't as big anymore

2004-08-12 10:22  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: always assume the
	  local daemon got the environment - increased minimum protocol
	  version for that

2004-08-12 09:23  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: fixing grave bug: it's just too hot ;(

2004-08-12 09:01  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: another place that can take longer than 10s

2004-08-11 21:20  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: this was supposed to go out only shortly

2004-08-11 20:48  Stephan Kulow <coolo@kde.org>

	* README: this looks nicer

2004-08-11 20:39  Stephan Kulow <coolo@kde.org>

	* README: better have it there for the moment

2004-08-11 19:50  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: it shouldn't even be there

2004-08-11 16:55  David Faure <faure@kde.org>

	* daemon/workit.cpp: thou shalt compile

2004-08-11 16:41  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/workit.cpp, services/scheduler.cpp:
	  implementing gcc's heuristics for our compiler, so it works
	  quickly even if run in chroot. That chroot feature caused
	  unexpected slow downs because of missing /proc, now it should
	  have the same speed again

2004-08-11 14:51  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: interesting debug message

2004-08-10 23:10  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: neat experiment

2004-08-10 20:55  Stephan Kulow <coolo@kde.org>

	* README: for the scheduler it's of little importance and the
	  daemon is root only

2004-08-10 19:06  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp: waiting for a long time where
	  the other side compiles

2004-08-10 18:08  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Configurable timeout
	  for get_msg.	coolo: you need to update at least one of the
	  callers in the client.  The default is now only 10 seconds.

2004-08-10 18:00  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: be a good guy and use the API

2004-08-10 14:26  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: * cleanup * pessimism factor * nicer
	  debug output

2004-08-10 14:17  Cornelius Schumacher <schumacher@kde.org>

	* mon/: starview.cc, starview.h: Implement architecture filter.

2004-08-10 13:49  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Ugly bug in stats collection.

2004-08-10 12:04  Stephan Kulow <coolo@kde.org>

	* TODO: outdated infos

2004-08-10 11:58  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: doing some memory leak hunting

2004-08-10 10:54  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: this feature requires files that are stored
	  locally

2004-08-10 09:20  Stephan Kulow <coolo@kde.org>

	* TODO, services/scheduler.cpp: removing untested feature for now

2004-08-09 20:41  Stephan Kulow <coolo@kde.org>

	* .cvsignore: ignore

2004-08-09 20:40  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: I swear: I said, divide by 0.7 - what did
	  he do? Math on his own

2004-08-09 20:38  Stephan Kulow <coolo@kde.org>

	* client/local.cpp, client/remote.cpp, daemon/main.cpp: if I say
	  that I'm not proud off it, can I leave it in?

2004-08-09 18:10  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Prefer the submitter to compile a job.
	  coolo: About the environment thing we have to talk again.
	  Depending on the outcome we have to ignore it for the submitter
	  (then we have to rely on the client to use the exact same
	  compiler as what was used for preprocessing, in _every case_), or
	  we can't ignore it, namely when the short-cut is not implemented
	  for all cases.

2004-08-09 17:48  Michael Matz <matz@kde.org>

	* services/comm.h: Surely this was wanted.

2004-08-09 17:38  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/local.cpp, client/main.cpp,
	  client/remote.cpp, daemon/main.cpp, services/comm.cpp,
	  services/comm.h: reworked my saturday night commit - the daemon
	  no longer decides if the client should ask the scheduler, but in
	  case the scheduler decides to give the localhost to the client,
	  it will use a short path within the daemon and compile itself.

2004-08-09 09:15  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: Matz is right, g++ test.c compiles as C++ not as
	  C

2004-08-07 23:12  Stephan Kulow <coolo@kde.org>

	* TODO, client/client.h, client/local.cpp, client/main.cpp,
	  daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: implementing a neat feature described in
	  TODO (danimo, should be interesting read for you) Michael: I left
	  some MATZ comments in the code

2004-08-05 17:36  Stephan Kulow <coolo@kde.org>

	* TODO: before I forget

2004-08-04 11:55  Stephan Kulow <coolo@kde.org>

	* mon/configure.in.in: this is needed for the monitor to work kind
	  of reliable.	Note: this might change your icecream installation
	  path from /usr/local to $KDEDIR

2004-08-03 09:59  Stephan Kulow <coolo@kde.org>

	* services/: comm.h, scheduler.cpp: Lubos is right, I was wrong

2004-08-02 21:07  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: you can talk to them all day, but they
	  don't listen

2004-08-02 17:00  Stephan Kulow <coolo@kde.org>

	* mon/hostinfo.cc, mon/hostinfo.h, mon/starview.cc,
	  services/scheduler.cpp: showing the speed the scheduler considers
	  in the starview

2004-08-02 15:59  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/load.cpp, daemon/main.cpp,
	  daemon/serve.cpp, daemon/serve.h, daemon/workit.cpp,
	  daemon/workit.h, services/comm.cpp, services/comm.h: implemented
	  some heuristic that prevents processes to take all memory. Niced
	  compile jobs are all fine, but if they take 2GB on a 512MB
	  machine, you notice them

	  In this case, the client will fall back to compiling locally
	  under this condition and will not treat it as hard error (of
	  course you need 2GB then)

2004-08-02 15:46  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Hmm, perhaps it might not be a good idea
	  to trigger interesting errors in the public version, as long as
	  they are still there.

2004-08-02 15:44  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: * Don't optimize cleanups with old
	  clients * ignore load of the submitter when farm is dense *
	  forced throwing out of daemon.  Triggers interesting errors.

2004-08-02 11:08  Stephan Kulow <coolo@kde.org>

	* mon/hostinfo.cc: even more colourful

2004-08-02 10:40  Stephan Kulow <coolo@kde.org>

	* mon/hostinfo.cc: trying to make the color distribution appearing
	  more random (for those with less than 70 nodes, it might not be
	  that obvious :)

2004-08-02 08:54  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-08-01 13:13  Stephan Kulow <coolo@kde.org>

	* README: give a little hint to those that are never satisfied

2004-07-30 16:48  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: these can walk like its older sister too

2004-07-30 16:04  Cornelius Schumacher <schumacher@kde.org>

	* mon/: hostinfo.cc, monitor.cc, starview.cc: Don't create empty
	  nodes for end of job message for offline hosts.

2004-07-30 15:01  Stephan Kulow <coolo@kde.org>

	* services/comm.h: matz needs a CURRENT

2004-07-30 14:55  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Forget all nodes when scheduler goes down.

2004-07-30 12:19  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Fix crash.

2004-07-30 12:06  Stephan Kulow <coolo@kde.org>

	* suse/init.icecream: do not start before you know your network
	  name - we had _tons_ of nodes named linux after a while

2004-07-30 09:02  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-29 15:01  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: throw an error in case something wild
	  happened. Otherwise the scheduler sees an EndMsg and expects a
	  sane processing (he only does the big cleanup for unexpected
	  errors)

2004-07-28 13:29  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: .c is a C file too

2004-07-28 08:38  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-26 16:00  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: This fixes some rare segfaults.

2004-07-18 20:36  Stephan Kulow <coolo@kde.org>

	* suse/init.icecream: that's nonsense

2004-07-15 11:11  Will Stephenson <lists@stevello.free-online.co.uk>

	* mon/hostinfo.cc: In a previous life I worked in an Eiscafe.

2004-07-11 09:07  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-06 08:33  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-04 08:15  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-03 09:02  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-07-02 13:43  Andrew Coles <andrew_coles@yahoo.co.uk>

	* mon/icemonapplet.desktop:

	  CVS_SILENT

	  Corrected typo.

2004-06-30 09:06  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-06-29 09:15  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: gcc allows -o foo -o bar and will output to bar

2004-06-27 09:30  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-06-19 18:10  Cornelius Schumacher <schumacher@kde.org>

	* daemon/main.cpp: Remove untrue message.

2004-06-16 16:46  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: always sleep when the scheduler goes away

2004-06-15 16:01  Cornelius Schumacher <schumacher@kde.org>

	* mon/: monitor.cc, monitor.h: Remember scheduler state when
	  switching views.

2004-06-14 11:50  Cornelius Schumacher <schumacher@kde.org>

	* mon/monitor.cc: Detect missing scheduler in all cases.

2004-06-11 13:48  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am: almost

2004-06-11 13:01  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: filter out ada and fortran extensions and set
	  unknown extensions to local jobs too

2004-06-07 21:35  Frerich Raabe <raabe@kde.org>

	* services/scheduler.cpp: - Simplify, and don't call
	  multimap<T>::upper_bound all that often

2004-06-07 19:49  Scott Wheeler <wheeler@kde.org>

	* services/scheduler.cpp: Add some basic support for compatible
	  platforms.

2004-06-07 19:49  Scott Wheeler <wheeler@kde.org>

	* client/remote.cpp: Make things a little more clear.

2004-06-07 12:02  David Faure <faure@kde.org>

	* configure.in.cvs: Don't bother me for uic, it's not used here.

2004-06-07 10:20  Stephan Kulow <coolo@kde.org>

	* Makefile.cvs, mon/Makefile.am: trying to fix automake build

2004-06-04 19:59  Scott Wheeler <wheeler@kde.org>

	* client/main.cpp: Repeating one out of every 5 jobs 3 times is way
	  too high for the common case.  Default to 1 in 100 and make it
	  possible to override this with an environment variable.

2004-06-04 15:16  Stephan Kulow <coolo@kde.org>

	* Makefile.am, Makefile.am.in: trying to avoid problems with
	  toplevel kdenonbeta

2004-06-03 12:48  Frerich Raabe <raabe@kde.org>

	* mon/: hostview.cc, hostview.h: - Walking past this without doing
	  anything would've made me feel bad.

2004-06-03 09:09  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-06-02 13:43  Cornelius Schumacher <schumacher@kde.org>

	* mon/: hostview.cc, hostview.h: Show state of scheduler in host
	  view.

2004-06-02 13:39  Cornelius Schumacher <schumacher@kde.org>

	* README: Add hint how to get debug output.

2004-06-02 13:38  Cornelius Schumacher <schumacher@kde.org>

	* mon/: mon-kde.cc, monitor.cc, starview.cc, starview.h,
	  statusview.cc, statusview.h: Show state of scheduler in star
	  view.

2004-05-31 08:53  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-05-30 09:14  Script Kiddy <sysadmin@office.kde.org>

	* mon/icemonapplet.desktop: CVS_SILENT made messages

2004-05-29 11:27  Stephan Kulow <coolo@kde.org>

	* .cvsignore: icecream works with unsermake finally :)

2004-05-29 10:57  Stephan Kulow <coolo@kde.org>

	* daemon/Makefile.am, minilzo/Makefile.am, mon/Makefile.am,
	  mon/configure.in.in, services/Makefile.am: trying to be good guys

2004-05-28 15:52  Cornelius Schumacher <schumacher@kde.org>

	* mon/: Makefile.am, hostview.cc, hostview.h, icemonapplet.cpp,
	  icemonapplet.desktop, icemonapplet.h, icemonui.rc, job.h,
	  mon-kde.cc, mon-kde.h, monitor.cc, monitor.h, starview.cc,
	  starview.h, statusview.cc, statusview.h: - Add view for
	  displaying status of single host.  - Add kicker applet making use
	  of this view.

2004-05-28 15:09  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am: WORKSFORME

2004-05-28 15:03  Stephan Kulow <coolo@kde.org>

	* Makefile.cvs: remove it before link

2004-05-17 09:27  Stephan Kulow <coolo@kde.org>

	* Makefile.cvs, configure.in.cvs, configure.in.in: you want tricks?
	  You can have them

2004-05-14 12:09  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.in.in: needed

2004-05-14 10:43  Stephan Kulow <coolo@kde.org>

	* .cvsignore, Makefile.cvs, configure.ac, configure.in.in: using
	  ../admin

2004-05-13 02:12  Otto Bruggeman <ottobruggeman@kde.nl>

	* client/remote.cpp: Last commit to fix this, i swear :)

2004-05-12 10:24  Cornelius Schumacher <schumacher@kde.org>

	* mon/: mon-kde.cc, starview.cc: Use table for tooltip formatting,
	  update star view when jobs are finished, fix about data.

2004-05-12 09:17  Otto Bruggeman <ottobruggeman@kde.nl>

	* client/remote.cpp: This makes the logic better. Sorry, should not
	  do this when just waking up.

2004-05-12 09:15  Otto Bruggeman <ottobruggeman@kde.nl>

	* client/remote.cpp: Ok now gcc >= 4 should work as well when it
	  comes. Thanks Helge for noticing.

2004-05-12 00:24  Otto Bruggeman <ottobruggeman@kde.nl>

	* client/remote.cpp: Need this on our Itanium cluster (for the rest
	  it works just fine there, great work guys!) since gcc 3.2.2 does
	  not understand this option either.

2004-05-11 15:20  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: fix compilation of several C files in one
	  compiler call (missed one)

2004-05-09 14:18  Scott Wheeler <wheeler@kde.org>

	* daemon/environment.cpp, services/job.cpp: build

2004-05-07 19:27  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.h: Yuck.  I had two things labeled "parent".

2004-05-04 20:13  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: compile g++ -c *.c as local job instead of
	  aborting

2004-05-04 14:20  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: some more fairness

2004-05-04 14:18  Stephan Kulow <coolo@kde.org>

	* mon/: hostinfo.cc, hostinfo.h, starview.cc: showing the platform

2004-05-04 13:35  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: excellent interface micha

2004-05-03 13:11  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: make sure not to compare apple and pears

2004-05-03 11:54  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: everything that we can ignore, is a warning.
	  So if you find your farm pretty slow and unused, check
	  ICECC_DEBUG=warnings to find out if the client does something at
	  all :)

2004-05-03 11:52  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: avoid net storms

2004-05-01 09:58  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Add config dialog
	  for gantt view.

2004-05-01 09:42  Cornelius Schumacher <schumacher@kde.org>

	* mon/ganttstatusview.cc: Use white background.

2004-05-01 09:34  Cornelius Schumacher <schumacher@kde.org>

	* mon/ganttstatusview.cc: Put text in the center of the bar.

2004-05-01 09:31  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, starview.cc, statusview.cc,
	  statusview.h: Use same text color in gantt view as in star view.

2004-05-01 08:50  Cornelius Schumacher <schumacher@kde.org>

	* mon/mon-kde.cc: Compile without Qt STL support.

2004-04-30 21:14  Stephan Kulow <coolo@kde.org>

	* TODO: I forgot

2004-04-30 16:35  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: or that

2004-04-30 14:42  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp, daemon/serve.cpp,
	  services/comm.cpp, services/comm.h, services/job.cpp,
	  services/job.h, services/scheduler.cpp: great, I cross compiled!
	  :) (you need to update all of your farm, I bumped the minimum
	  protocol)

2004-04-30 13:56  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Use center of items as reference.

2004-04-30 13:32  Cornelius Schumacher <schumacher@kde.org>

	* mon/: icemonui.rc, mon-kde.cc, mon-kde.h, starview.cc,
	  starview.h, statusview.h: Show nodes in more than one ring if
	  there are lots of hosts.  Add configuration dialog for star view.

2004-04-30 11:59  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/environment.cpp, daemon/environment.h,
	  daemon/main.cpp, daemon/serve.cpp, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: further preparation for
	  cross compiling

2004-04-30 10:07  Stephan Kulow <coolo@kde.org>

	* TODO, daemon/environment.cpp, daemon/environment.h,
	  daemon/main.cpp: package the local environment for users without
	  ICECC_VERSION - convenient feature (and another nail in the
	  coffin for "let the daemon run as user")

2004-04-29 17:50  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/main.cpp, client/remote.cpp,
	  daemon/environment.cpp, daemon/environment.h, daemon/main.cpp,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp:
	  preparation for cross-compiler support - from now on
	  $ICECC_VERSION got a bit different format   (mine is
	  /work/users/coolo/9.1-i386.tar.bz2,x86_64:/work/users/coolo/9.1-x86_64-cross-i386.tar.bz2)

	  this is just the first step. Until Michael finished reading his
	  mail, the scheduler will ignore all but the environments from the
	  target platform (the one without actual prefix in $ICECC_VERSION)

2004-04-29 17:00  Stephan Kulow <coolo@kde.org>

	* mon/: job.h, mon-kde.cc: no one cares for the version

2004-04-29 11:05  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, statusview.cc,
	  statusview.h: Add vertical scrollbar to ganttstatusview.

2004-04-28 17:00  Stephan Kulow <coolo@kde.org>

	* client/create-env: that's interesting. 2.4.0 breaks lib64 support
	  :)

2004-04-28 12:01  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: as I said

2004-04-28 11:37  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: oops

2004-04-28 11:31  Stephan Kulow <coolo@kde.org>

	* Makefile.cvs, TODO, client/main.cpp, client/remote.cpp,
	  daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/job.cpp, services/job.h, services/scheduler.cpp: support
	  for differing platforms (by uname's .machine) - require
	  ICECC_VERSION to be set for now (read the TODO file if you're
	  annoyed :)   - which in return means iceccd requires being
	  started as root feedback wanted - especially from those that
	  share i686 and i586 build hosts (those poor souls :)

2004-04-27 18:03  Cornelius Schumacher <schumacher@kde.org>

	* mon/: hostinfo.cc, hostinfo.h, mon-kde.cc, starview.cc,
	  starview.h, statusview.h: - Tell the status view if a node has
	  gone offline.  - Remove offline nodes in star view.

2004-04-27 16:35  Cornelius Schumacher <schumacher@kde.org>

	* mon/: hostinfo.cc, hostinfo.h, starview.cc, starview.h:
	  Immediately create all known hosts.

2004-04-27 11:29  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Show lines not only for the server but also for
	  the client.

2004-04-27 09:53  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Use full width of window.

2004-04-26 17:22  Stephan Kulow <coolo@kde.org>

	* client/: Makefile.am, client.h, cpp.cpp, local.cpp, main.cpp,
	  safeguard.cpp: trying to protect against recursive calling -
	  Scott, please test

2004-04-26 16:31  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: building binutils with icecream is no fun - tons
	  of bugs :)

2004-04-26 15:58  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: don't annoy users that do not know
	  ICECC_VERSION

2004-04-26 15:03  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: empty files are absolute too - by definition
	  :)

2004-04-26 11:02  Stephan Kulow <coolo@kde.org>

	* suse/: icecream.spec, init.icecream, sysconfig.icecream: for
	  whoever finds it interesting to have packages

2004-04-26 10:34  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/local.cpp, client/remote.cpp,
	  daemon/environment.cpp, daemon/environment.h,
	  daemon/findmyself.h, daemon/load.h, daemon/main.cpp,
	  daemon/ncpus.h, daemon/serve.cpp, daemon/serve.h,
	  daemon/workit.cpp, daemon/workit.h, services/comm.cpp,
	  services/comm.h, services/job.cpp, services/job.h,
	  services/logging.cpp, services/logging.h, services/receiver.cpp,
	  services/scheduler.cpp, services/submitter.cpp,
	  services/tempfile.h: adding copyright headers to all files

2004-04-26 10:22  Stephan Kulow <coolo@kde.org>

	* client/client.h, client/local.cpp, mon/mon-kde.cc,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp:
	  submit the output filename for local jobs too

2004-04-26 01:01  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: Remove machines from the
	  monitor as they drop out of the cluster.

2004-04-25 23:01  Scott Wheeler <wheeler@kde.org>

	* client/main.cpp: This is no good.  If the user does "icecc foo.c"
	  this currently just prints the usage message and exits.  Only
	  print the usage stuff if it's asked for.

2004-04-25 21:02  Scott Wheeler <wheeler@kde.org>

	* client/remote.cpp: -frandome-seed doesn't exist on GCC < 3.

2004-04-25 12:44  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, services/job.h: found the "make gcc
	  compilation reproducible" option - which gives comparing md5sum
	  of the output kind of a sense for some more advanced usage of C++
	  :)

2004-04-24 09:26  Stephan Kulow <coolo@kde.org>

	* mon/hostinfo.cc: a fresh one :)

2004-04-23 23:46  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/serve.cpp: another case where the farm
	  locked ;(

2004-04-23 21:17  Stephan Kulow <coolo@kde.org>

	* mon/hostinfo.cc: minor optimizations

2004-04-23 18:38  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp, services/comm.cpp: I was hunting this bug for
	  quite a while - which shows what UNIX novice I am ;(

2004-04-23 14:21  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: I guess, relying on the preprocessor to create
	  always the same output is naiv

2004-04-23 12:12  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: don't let the number of current childs go below
	  0 just because the scheduler is unavailable for a moment

2004-04-23 09:42  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: too much information even for debug

2004-04-22 11:22  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: jobs started at the same time tend to have the
	  same random - that wasn't my intention :)

2004-04-22 11:19  Stephan Kulow <coolo@kde.org>

	* TODO: cleaning up a bit

2004-04-22 09:57  Stephan Kulow <coolo@kde.org>

	* TODO, client/client.h, client/main.cpp, client/remote.cpp,
	  services/scheduler.cpp: finally finished my failsafe feature.
	  every ~5th job is compiled three times and then checked if the
	  resulting object files are all the same.

	  Don't worry, I'll make that configurable through ICECC_DEBUG as
	  soon as I know it works :)

2004-04-22 09:32  Stephan Kulow <coolo@kde.org>

	* mon/: hostinfo.cc, hostinfo.h: make it easier to have the color
	  in kcolorchooser CCMAIL: wimer@suse.de

2004-04-22 01:20  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.cc: Don't get confused by failed / canceled jobs.

2004-04-21 17:06  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am, client/md5.c, client/md5.h,
	  client/remote.cpp, daemon/Makefile.am, daemon/serve.cpp,
	  daemon/tempfile.c, daemon/tempfile.h, daemon/workit.cpp,
	  services/Makefile.am, services/tempfile.c, services/tempfile.h:
	  getting somewhere with my failsafe mode

2004-04-21 14:29  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: a little less conversation...

2004-04-21 13:50  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: is that part of "off by one"?

2004-04-21 13:30  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, mon/mon-kde.cc, services/comm.cpp,
	  services/scheduler.cpp: fixing the handling of protocol out of
	  rage

2004-04-21 11:51  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/scheduler.cpp: no one can spell
	  --no-detach - default to foreground and make the switch -d,
	  --daemonize

2004-04-21 11:32  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, comm.h: making the protocol version hand
	  shake more safe

2004-04-21 10:44  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: * introducing at least a
	  concept of different protocol versions talking with	each other
	  in trying to get the minimal version and then chat with that
	  (requires increased protocol version :) * let the client request
	  x jobs per file and compile it - still experimental

2004-04-21 01:41  Scott Wheeler <wheeler@kde.org>

	* mon/: hostinfo.cc, hostinfo.h: make methods const where
	  appropriate

2004-04-21 01:28  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.cc: Add color coding.

2004-04-20 23:19  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.cc: Add a bit of spacing between items if the
	  system can handle multiple jobs.

2004-04-20 20:44  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: Support for multiple CPUs.

	  CCMAIL:staikos@kde.org

2004-04-20 19:04  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: minor layouting fixes

2004-04-20 18:56  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.cc: Set the minimum height and avoid a weird
	  resizing thing that happens when less than is used when the
	  heights are equal.

2004-04-20 18:50  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: Make it scrollable.

2004-04-20 17:19  Scott Wheeler <wheeler@kde.org>

	* mon/summaryview.cc: Just show the file name -- not the full path.

2004-04-20 17:12  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: Make sure that the columns
	  are balanced.

2004-04-20 14:47  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, services/job.h: preparing the client in
	  general to compile all jobs several times (experimental, still
	  disabled)

2004-04-20 13:12  Frerich Raabe <raabe@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/comm.h: - Make it
	  possible to tell the daemon what host the scheduler runs on via
	  the commandline.

2004-04-19 21:41  Stephan Kulow <coolo@kde.org>

	* mon/starview.cc: init variables

2004-04-19 21:06  Scott Wheeler <wheeler@kde.org>

	* services/scheduler.cpp, daemon/main.cpp: Mention the --no-detach
	  option.

2004-04-19 16:40  Cornelius Schumacher <schumacher@kde.org>

	* mon/: hostinfo.cc, hostinfo.h, starview.cc: Show IP address in
	  tooltip.

2004-04-19 16:08  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: give the IP too

2004-04-19 16:00  Cornelius Schumacher <schumacher@kde.org>

	* mon/: Makefile.am, ganttstatusview.cc, ganttstatusview.h,
	  hostinfo.cc, hostinfo.h, job.cc, job.h, listview.cc, listview.h,
	  mon-kde.cc, mon-kde.h, starview.cc, starview.h, statusview.cc,
	  statusview.h, summaryview.cc, summaryview.h: - Hold host
	  information in own objects independent of the views.	- Assign
	  fixed colors to hosts based on host names.  - Move Job class two
	  own source file.

2004-04-18 23:43  Cornelius Schumacher <schumacher@kde.org>

	* mon/: starview.cc, starview.h: Show all available nodes in star
	  view.

2004-04-18 19:17  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: as long as the summary view doesn't scale, using
	  it as default is bad ;(

2004-04-18 10:24  Stephan Kulow <coolo@kde.org>

	* mon/statusview.cc, services/scheduler.cpp: tell the monitor when
	  one node went away - currently unused by most views

2004-04-16 16:54  Frerich Raabe <raabe@kde.org>

	* services/scheduler.cpp: - Mention --log-file in usage
	  instructions

2004-04-16 16:37  Stephan Kulow <coolo@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, mon-kde.cc,
	  statusview.cc, statusview.h, summaryview.cc, summaryview.h:
	  finding the number of maximum jobs

2004-04-16 16:25  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, statusview.cc, statusview.h: inheriting the
	  host names is the minimum you expect when switching views

2004-04-16 16:20  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: find the server for the local job - we
	  used to compare by IP, but this is no longer good enough

2004-04-16 15:56  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: IDs are expensive, spare them for servers

2004-04-16 15:11  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: conditional casts are stupid ;(

2004-04-16 14:44  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: don't rely on STL support in Qt

2004-04-16 14:30  Stephan Kulow <coolo@kde.org>

	* mon/: Makefile.am, mon-kde.cc, starview.cc, starview.h: it
	  compiles and adds nodes

2004-04-16 14:22  Stephan Kulow <coolo@kde.org>

	* mon/: Makefile.am, mon-kde.cc, summaryview.cc, summaryview.h:
	  summary view is in again too

2004-04-16 14:17  Stephan Kulow <coolo@kde.org>

	* mon/: Makefile.am, ganttstatusview.cc, ganttstatusview.h,
	  mon-kde.cc, statusview.cc, summaryview.cc, summaryview.h:
	  ganttview is ported too

2004-04-16 14:03  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am, mon/ganttstatusview.cc, mon/ganttstatusview.h,
	  mon/listview.cc, mon/mon-kde.cc, mon/starview.cc,
	  mon/statusview.cc, mon/statusview.h, services/comm.cpp,
	  services/scheduler.cpp: porting monitor to new protocol - only
	  list view activated

2004-04-16 12:00  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: raising protocol version for * daemons
	  submit there local node name * the statistics message from
	  scheduler to monitor is free text allowing me   to change the
	  format and the content without raising protocol version (does not
	  compile right now - work in progress)

2004-04-16 09:38  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, serve.cpp, serve.h: make nice level
	  configurable (Lubos, your fault)

2004-04-15 17:24  Stephan Kulow <coolo@kde.org>

	* mon/starview.cc: just a proof of concept - combining starview and
	  summary view

2004-04-15 09:12  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: on scheduler disconnect, forget about the jobs

2004-04-14 22:21  Scott Wheeler <wheeler@kde.org>

	* services/logging.h: build

2004-04-14 20:29  Stephan Kulow <coolo@kde.org>

	* mon/statusview.h: compile

2004-04-14 17:20  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Adapt text color to host color.

2004-04-14 17:04  Cornelius Schumacher <schumacher@kde.org>

	* mon/starview.cc: Draw star view lines in color of client.

2004-04-14 16:51  Cornelius Schumacher <schumacher@kde.org>

	* mon/: starview.cc, starview.h: Remove finished local jobs from
	  star view.

2004-04-14 15:38  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, starview.cc: Show multiple jobs on
	  compile hosts in star view.

2004-04-14 14:24  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, starview.cc,
	  starview.h, statusview.cc, statusview.h: Show host names and
	  colors in star view.

2004-04-14 13:24  Cornelius Schumacher <schumacher@kde.org>

	* mon/: starview.cc, starview.h: Make star view show something more
	  useful.

2004-04-14 11:38  Cornelius Schumacher <schumacher@kde.org>

	* mon/: Makefile.am, ganttstatusview.cc, mon-kde.cc, starview.cc,
	  starview.h: Make star view compile and at least do something.

2004-04-14 10:54  Cornelius Schumacher <schumacher@kde.org>

	* mon/: Makefile.am, ganttstatusview.cc, ganttstatusview.h,
	  listview.cc, listview.h, mon-kde.cc, mon-kde.h, starview.cc,
	  starview.h, statusview.cc, statusview.h, summaryview.h: - Moved
	  view classes to own files.  - Added copyright headers.  -
	  Remember current view.

2004-04-14 09:51  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Show host names
	  instead of ip addresses.

2004-04-13 21:07  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: no warnings by default

2004-04-13 20:43  Scott Wheeler <wheeler@kde.org>

	* mon/: summaryview.cc, summaryview.h: Notice which machines are on
	  the network...

2004-04-13 14:01  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, comm.h: new protocol version - with working
	  hand shake :)

2004-04-13 14:00  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: just a warning

2004-04-13 12:04  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: give feedback for the protocol version -
	  trying to avoid hanging clients

2004-04-13 11:40  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, services/scheduler.cpp: trying harder to give
	  any host in the farm a chance to proof he's fastest (if the farm
	  is big enough, we used to stick to the same set of hosts)

2004-04-13 10:58  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: no cout

2004-04-13 10:48  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: if the scheduler is down, reconnect

2004-04-12 22:52  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: trying harder not to loose track of
	  daemons - ping them and if they still don't answer, kill them.

2004-04-12 17:13  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: oops

2004-04-12 16:30  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: don't printf

2004-04-12 14:11  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/scheduler.cpp: call iceccd as
	  --no-detach -vvv to get the old behaviour

2004-04-12 13:44  Stephan Kulow <coolo@kde.org>

	* TODO, daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: first try to get the protocol version

2004-04-12 12:07  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: the monitor calls trace() indirectly too

2004-04-12 12:01  Stephan Kulow <coolo@kde.org>

	* daemon/load.cpp: trace() the load (to /dev/null)

2004-04-12 12:01  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: more trace() debugs

2004-04-12 12:00  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, daemon/main.cpp, services/Makefile.am,
	  services/logging.cpp, services/logging.h, services/scheduler.cpp:
	  log to a file and filter out trace() etc per default

2004-04-09 23:03  Scott Wheeler <wheeler@kde.org>

	* mon/mon-kde.cc: i18n-ifying these strings...

2004-04-09 21:54  Scott Wheeler <wheeler@kde.org>

	* mon/: Makefile.am, icemonnode.png, icemonui.rc, mon-kde.cc,
	  mon-kde.h, summaryview.cc, summaryview.h: Adding a new view mode
	  (and by coolo's request making it the default).

	  This is a more visually attrative way of viewing things than the
	  listview and doesn't take near as much CPU as the Gantt View.

	  There are still several things left to be done like having it
	  search for nodes on startup and turning it into a scrollview for
	  build farms with lots of nodes, but that will come later...

2004-03-10 10:25  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, icemonui.rc,
	  mon-kde.cc, mon-kde.h: Remove dead nodes.

2004-03-03 22:17  Cornelius Schumacher <schumacher@kde.org>

	* README: Fix typo.

2004-03-03 11:19  Cornelius Schumacher <schumacher@kde.org>

	* README: Add some docs about ICECC_VERSION.

2004-03-03 10:44  Stephan Kulow <coolo@kde.org>

	* client/create-env: back

2004-03-03 09:54  Stephan Kulow <coolo@kde.org>

	* client/create-env: this is now the third time I tell someone he
	  needs to patch PATH before calling this script

2004-03-02 17:28  Stephan Kulow <coolo@kde.org>

	* daemon/: load.cpp, main.cpp: be more conservative on what hosts
	  you label as free

2004-03-02 16:24  Stephan Kulow <coolo@kde.org>

	* client/: main.cpp, remote.cpp: don't send an EndMsg to the
	  scheduler if something failed. Just throw an exception, shut down
	  all connections and build local without telling anyone

2004-03-02 14:49  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: take canceled jobs out of the monitors

2004-03-01 16:01  Luboš Luňák <l.lunak@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Don't repaint the
	  filename in the monitor all the time (all job 50 times a second),
	  instead cache the drawn text in a pixmap and redo it only when
	  necessary. Greatly reduces CPU usage.

2004-03-01 15:57  Luboš Luňák <l.lunak@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Use normal struct
	  for data in the list instead of QPair.

2004-03-01 07:48  George Staikos <staikos@kde.org>

	* TODO: more to do :)

2004-02-27 22:29  David Faure <faure@kde.org>

	* services/scheduler.cpp: Don't crash

2004-02-27 08:10  Stephan Kulow <coolo@kde.org>

	* client/util.cpp: less debug

2004-02-23 11:16  Stephan Kulow <coolo@kde.org>

	* daemon/environment.cpp: allow names with .

2004-02-20 20:03  Stephan Kulow <coolo@kde.org>

	* TODO, client/arg.cpp, client/local.cpp, client/util.cpp,
	  client/util.h, services/comm.cpp: don't kill the machine if the
	  scheduler can't be reached, but lock the machine and run one job
	  at a time

2004-02-20 18:38  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: This should fix it.

2004-02-20 15:21  Stephan Kulow <coolo@kde.org>

	* TODO: Cornelius was annoyed by this very simple thing we added,
	  but at least he had an idea we like :)

2004-02-20 10:08  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: don't even listen for clients when we're not
	  connected to a scheduler. Otherwise we listen but don't talk -
	  which gives hanging clients

2004-02-20 09:54  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: reconnect to the scheduler

2004-02-20 09:07  Stephan Kulow <coolo@kde.org>

	* Makefile.cvs, services/comm.h, services/scheduler.cpp: some more
	  debug

2004-02-19 19:14  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: I overlooked one path, where this is
	  indeed an login message.  Like, ...  erm ..., linking.

2004-02-18 21:39  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Make toanswer a list, not queue.
	  Correctly send end messages for deleted jobs.  Otherwise if a
	  daemon dies all clients currently waiting on that machine would
	  hang.

2004-02-18 18:12  Michael Matz <matz@kde.org>

	* TODO: Would be nice.

2004-02-18 17:56  Michael Matz <matz@kde.org>

	* client/remote.cpp: This is done later unconditionally.

2004-02-18 17:55  Michael Matz <matz@kde.org>

	* client/remote.cpp: No need to two local 100k arrays.

2004-02-18 17:47  Michael Matz <matz@kde.org>

	* client/Makefile.am: Installing it is a good idea.

2004-02-18 17:44  Michael Matz <matz@kde.org>

	* client/main.cpp: Done by a script now.

2004-02-18 15:35  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: If the client end normally, don't do
	  cleanups.  (This will break if the daemon exits after sending the
	  .o file, but before sending the job_end message).

2004-02-18 15:27  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: avoid bogus warning

2004-02-18 15:16  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.h, services/scheduler.cpp:
	  reannounce environments and ignore meanwhile the servers with
	  installing environments

2004-02-18 15:05  Michael Matz <matz@kde.org>

	* client/create-env: I didn't want to check in the use of gzip.

2004-02-18 14:15  Scott Wheeler <wheeler@kde.org>

	* configure.ac, client/main.cpp, client/remote.cpp,
	  daemon/Makefile.am, daemon/environment.cpp, daemon/serve.cpp,
	  daemon/workit.cpp, services/Makefile.am, services/comm.cpp,
	  services/getifaddrs.cpp, services/getifaddrs.h,
	  services/scheduler.cpp: Make icecream build on older systems
	  (glibc 2.2.5, gcc 2.95).  Approved by Matz...

2004-02-18 00:55  Michael Matz <matz@kde.org>

	* TODO: In retrospective just the scheduler is broken.

2004-02-18 00:53  Michael Matz <matz@kde.org>

	* TODO: Did something.

2004-02-17 23:50  Michael Matz <matz@kde.org>

	* client/main.cpp, client/remote.cpp, services/scheduler.cpp:
	  local-job messages to the scheduler were broken.  Client tried to
	  send them after END was already sent.  Not possible.	Working
	  around this exposed some other bugs.

2004-02-17 23:01  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Also removing it from daemons joblist
	  leaves the scheduler in a saner state.

2004-02-17 22:50  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Rewrite end of socket handling.  Fixes at
	  least one segfault.  Still not complete though.

2004-02-17 22:49  Michael Matz <matz@kde.org>

	* client/remote.cpp: Don't access deleted message.  Send END
	  message only when daemon got successfully contacted.

2004-02-17 13:48  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp, services/scheduler.cpp: do not fall back in the
	  daemon when the environment doesn't exist

2004-02-16 17:48  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Otherwise we have forgotten in two weeks.

2004-02-16 17:11  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: don't let linux 2.6.2 DoS us ;(

2004-02-16 15:54  Michael Matz <matz@kde.org>

	* client/create-env: I think a script is better.

2004-02-09 21:57  Stephan Kulow <coolo@kde.org>

	* TODO: one part is done

2004-02-09 21:51  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, client/remote.cpp, daemon/environment.cpp,
	  daemon/environment.h, daemon/main.cpp, daemon/serve.cpp,
	  daemon/serve.h, services/comm.cpp, services/comm.h: make support
	  for versions work. You still better do not update though :)

2004-02-09 19:33  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, scheduler.cpp: two fixes for the env
	  transfer

2004-02-08 22:27  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, client/remote.cpp, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: more work on compiler
	  environments

2004-02-07 23:31  Stephan Kulow <coolo@kde.org>

	* client/main.cpp, services/job.h: adding an option to create a tar
	  ball with the compile environment

2004-02-07 13:12  Stephan Kulow <coolo@kde.org>

	* services/: job.cpp, job.h: let the compiler decide if he can
	  inline

2004-02-07 13:06  Stephan Kulow <coolo@kde.org>

	* services/job.cpp: thanks Frerich, but the arguments once are fine

2004-02-06 23:50  Frerich Raabe <raabe@kde.org>

	* services/comm.h: - Whoever wanted to make this compile on FreeBSD
	  must've been using an editor	 window which was only three lines
	  in height.

2004-02-06 23:38  Frerich Raabe <raabe@kde.org>

	* services/scheduler.cpp: - fixed copy&paste error

2004-02-06 19:39  Daniel Molkentin <molkentin@kde.org>

	* services/scheduler.cpp: - allow for selecting a custom port
	  (pending: implement in daemon and monitor) - more verbose help

2004-02-06 16:30  Daniel Molkentin <molkentin@kde.org>

	* services/scheduler.cpp: compile

2004-02-06 16:23  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying harder to catch unexpected client
	  errors gracefully

2004-02-05 15:27  Luboš Luňák <l.lunak@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: When removing a
	  slot, I guess I forgot to remove it also from mJobMap.

2004-02-05 13:59  Frerich Raabe <raabe@kde.org>

	* services/: job.cpp, job.h: - Fix code duplication (and removed
	  funny appendList implementation, why on   earth is there a
	  dedicated method for such a task?)

2004-02-02 14:09  Luboš Luňák <l.lunak@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Show no slots for
	  gone nodes. They should be perhaps removed completely, but
	  QGridLayout can't be shrinked.

2004-02-02 14:09  Luboš Luňák <l.lunak@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, mon-kde.cc,
	  mon-kde.h: Limit the numbers of slots per node to the maximum
	  allowed processes, and drop excessive empty slots. It looks
	  stupid if a machine gets 10 slots forever just because it did 10
	  local link jobs due to make -j10.

2004-02-02 14:03  Luboš Luňák <l.lunak@kde.org>

	* client/local.cpp: Handle user break doing local compiles, in
	  order to always send JobLocalDoneMsg. They'd stay stuck in icemon
	  (at least).

2004-02-02 13:58  Luboš Luňák <l.lunak@kde.org>

	* client/remote.cpp: Maybe even better - create the data from
	  remote compilation first as a temporary file, rename only after
	  successfully receving it.

2004-02-02 13:57  Luboš Luňák <l.lunak@kde.org>

	* client/remote.cpp: Don't create empty .o files if remote compile
	  failed.

2004-02-02 13:56  Luboš Luňák <l.lunak@kde.org>

	* client/cpp.cpp: Fix usage when used together with ccache. Gcc has
	  the brai^H^H^Hinteresting feature that g++ -E preprocessed.ii
	  outputs big nothing. Detect .i/.ii files and use cat instead of
	  g++ -E for them.

2004-01-31 16:26  Frerich Raabe <raabe@kde.org>

	* daemon/load.cpp: - Trying to get /proc-dependant stuff ported to
	  FreeBSD

2004-01-30 13:23  Luboš Luňák <l.lunak@kde.org>

	* daemon/load.cpp: 1014 -> 1024 . Let's not risk load higher than
	  maximum ;).

2004-01-30 13:17  Michael Matz <matz@kde.org>

	* client/main.cpp: Too much copy-paste.  No sense in an own process
	  group for the client.

2004-01-29 20:36  Michael Matz <matz@kde.org>

	* client/arg.cpp: I prefer it this way.

2004-01-29 20:14  Luboš Luňák <l.lunak@kde.org>

	* client/arg.cpp: Nice cryptic error message. Either the type is
	  list, or the variable is list, but not both.

2004-01-29 16:05  Luboš Luňák <l.lunak@kde.org>

	* mon/Makefile.am: Add missing $(DESTDIR).

2004-01-26 22:07  George Staikos <staikos@kde.org>

	* daemon/environment.cpp, services/comm.h: compile

2004-01-26 20:04  George Staikos <staikos@kde.org>

	* client/remote.cpp: compile

2004-01-21 16:32  Michael Matz <matz@kde.org>

	* client/client.h: Forgot.

2004-01-21 16:32  Michael Matz <matz@kde.org>

	* client/: Makefile.am, arg.cpp, cpp.cpp, local.cpp, main.cpp,
	  remote.cpp, util.cpp, util.h: Clean up unused code.

2004-01-19 11:58  Cornelius Schumacher <schumacher@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: Add command line option for setting
	  icecream network name.

2004-01-19 11:42  Cornelius Schumacher <schumacher@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: Don't remeber all jobs when
	  switching to gantt view.

2004-01-18 23:14  Cornelius Schumacher <schumacher@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h: Make all progress
	  bars use the same height. That means machines taking more jobs
	  now get more space in the monitor.

2004-01-16 18:12  Michael Matz <matz@kde.org>

	* services/comm.cpp: Subtle.  Obviously people only compile correct
	  code, because with errors the client would segfault.

2004-01-09 10:19  Stephan Kulow <coolo@kde.org>

	* TODO: people are mean ;)

2004-01-09 09:49  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: do not use more jobs than CPUs (still suggested
	  to overrun machines that are idle most of the time - use -m for
	  that) -w is without option

2004-01-05 17:37  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: when a daemon exits, remove everything we
	  know about it to avoid crashes (this might not be the perfect
	  solution, better might be to keep it around for a while and mark
	  it offline)

2004-01-05 16:24  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: allowing 0 jobs is all fine, but it does it
	  really have to be the default?

2004-01-04 19:14  Daniel Molkentin <molkentin@kde.org>

	* daemon/main.cpp: As long as the scheduler is not smart enough to
	  spare old and slow PC's, allow 0 compile processes. Also allowing
	  negative -m's do not really make sense :)

2004-01-04 09:53  Daniel Molkentin <molkentin@kde.org>

	* client/main.cpp: fix copyright. Too bad there is no AUTHORS file
	  or any copyright headers.  Could somebody please add them?

2004-01-04 09:50  Daniel Molkentin <molkentin@kde.org>

	* mon/: ganttstatusview.cc, ganttstatusview.h, mon-kde.cc,
	  mon-kde.h: that was not meant to go in

2004-01-04 09:49  Daniel Molkentin <molkentin@kde.org>

	* README, mon/ganttstatusview.cc, mon/ganttstatusview.h,
	  mon/mon-kde.cc, mon/mon-kde.h: add warning about use in untrusted
	  networks

2004-01-04 01:33  Daniel Molkentin <molkentin@kde.org>

	* README: Add a small README :)

2004-01-04 01:01  Daniel Molkentin <molkentin@kde.org>

	* services/scheduler.cpp: ported to getopt

2004-01-04 00:53  Daniel Molkentin <molkentin@kde.org>

	* client/main.cpp: don't lie

2004-01-04 00:45  Daniel Molkentin <molkentin@kde.org>

	* daemon/main.cpp: port to getopt

2004-01-04 00:10  Daniel Molkentin <molkentin@kde.org>

	* daemon/main.cpp: Ok, everybody sick of hacking his own processor
	  limit in main.cpp: Enter -m (max_processes)

2004-01-02 19:05  Adriaan de Groot <groot@kde.org>

	* daemon/: main.cpp, ncpus.c, serve.cpp: More includes, and what
	  the heck is rs_log_error?

2004-01-02 18:48  Adriaan de Groot <groot@kde.org>

	* daemon/workit.cpp: Get the includes right

2004-01-02 18:20  Adriaan de Groot <groot@kde.org>

	* client/remote.cpp: Need other includes for sendfile; O_LARGEFILE
	  doesn't exist.

2004-01-02 18:17  Adriaan de Groot <groot@kde.org>

	* services/comm.h: socklen_t needs includes in BSD

2004-01-02 18:11  Adriaan de Groot <groot@kde.org>

	* services/comm.h: stdint.h is a linuxism

2003-12-31 15:49  Cornelius Schumacher <schumacher@kde.org>

	* mon/ganttstatusview.cc: Fix display of local jobs.

2003-12-30 17:55  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: At least some fairness. Rotate through
	  the clients when looking for a new job.

2003-12-30 00:28  Cornelius Schumacher <schumacher@kde.org>

	* mon/ganttstatusview.cc: Fix layout warning.

2003-12-29 23:06  Cornelius Schumacher <schumacher@kde.org>

	* mon/Makefile.am: Fix srcdir!=builddir.

2003-12-26 23:48  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: removing debug output always makes me
	  think it's stable :)

2003-12-26 22:11  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: removing some debug blah blah

2003-12-25 22:35  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am: reverting last commit - this configure doesn't
	  use KDE's build system

2003-12-25 22:34  Stephan Kulow <coolo@kde.org>

	* services/Makefile.am: relocatable

2003-12-25 22:30  Stephan Kulow <coolo@kde.org>

	* client/local.cpp, client/main.cpp, client/remote.cpp,
	  mon/ganttstatusview.cc, mon/mon-kde.cc, mon/mon-kde.h,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp: show
	  link jobs in the monitor. Triggers some bugs in ganttview and
	  shows nicely how urgently the libtool rewrite needs to be
	  finished :)

2003-12-23 22:58  Ingo Klöcker <kloecker@kde.org>

	* mon/Makefile.am: Make it compile if KDEDIR isn't set. I mean, why
	  do all_includes and all_libraries exist if they aren't used?

2003-12-23 21:38  Ingo Klöcker <kloecker@kde.org>

	* services/Makefile.am: make it compile

2003-12-18 12:08  cschum

	* mon/: ganttstatusview.cc, ganttstatusview.h, icemonui.rc,
	  mon-kde.cc, mon-kde.h: Show one progress bar for each slot on a
	  node. This now correctly shows all jobs including not showing
	  them anymore when they are ended.

2003-12-12 10:02  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, scheduler.cpp: uncommented some debug output
	  and do not abort on empty msgs until matz tells what it is :)

2003-12-11 14:04  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: transfer the max_kids
	  number per host to the mon

2003-12-11 13:51  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, scheduler.cpp: adding tons of debug output
	  to get closer to a problem that had a simple origin: select can
	  return with errno = EINTR without it being a fatal error

2003-12-11 09:31  Stephan Kulow <coolo@kde.org>

	* mon/ganttstatusview.cc: display short name - possibly remember
	  it?

2003-12-10 17:46  cschum

	* mon/: ganttstatusview.cc, ganttstatusview.h, mon-kde.cc,
	  mon-kde.h: - Show all known nodes, even when they haven't
	  compiled anything yet.  - Correctly position jobs on time scale.

2003-12-10 15:36  cschum

	* mon/: ganttstatusview.cc, ganttstatusview.h, mon-kde.cc,
	  mon-kde.h: - Color code hosts.  - Fix display for end of jobs
	  (seems not to cover all cases yet).  - Show host name labels.

2003-11-13 16:39  Stephan Kulow <coolo@kde.org>

	* TODO, client/arg.cpp: argument lists

2003-11-13 15:38  Stephan Kulow <coolo@kde.org>

	* client/main.cpp: start process group

2003-11-12 18:10  Michael Matz <matz@kde.org>

	* services/comm.cpp: Blocking write on a non-blocking socket needs
	  more that just some write calls (should fix some strange
	  behaviour of lingering cc1(plus)).

2003-11-10 17:08  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h: EOF is also a "message".

2003-11-10 16:24  Stephan Kulow <coolo@kde.org>

	* mon/: .cvsignore, Makefile.am, ganttstatusview.cc,
	  ganttstatusview.h, icemonui.rc, mon-kde.cc, mon-kde.h: ganttview

2003-10-31 23:42  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Make use of nonblocking reading.  Didn't
	  check if it's really nonblocking ;-) While doing this it turned
	  out that the lifetime of objects became quite complex and
	  unmanagable.

2003-10-31 02:15  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Backbone for
	  nonblocking MsgChannels.  Also has a (mostly) blocking interface
	  (by default) for easy use.

2003-10-28 20:21  Stephan Kulow <coolo@kde.org>

	* TODO: bug found

2003-10-28 16:16  Stephan Kulow <coolo@kde.org>

	* client/: arg.cpp, local.cpp: try to keep the arguments list
	  intact as much as possible

2003-10-28 16:15  Stephan Kulow <coolo@kde.org>

	* minilzo/Makefile.am: some fixes for make dist

2003-10-25 15:30  Stephan Kulow <coolo@kde.org>

	* TODO, client/arg.cpp, client/cpp.cpp, client/local.cpp,
	  daemon/workit.cpp, services/comm.cpp, services/job.cpp,
	  services/job.h: don't split the args in three list, but in one
	  flags list

2003-10-24 11:33  Stephan Kulow <coolo@kde.org>

	* TODO: ugly

2003-10-24 11:16  Stephan Kulow <coolo@kde.org>

	* TODO, client/local.cpp: local args last - as they contain the
	  link arguments

2003-10-23 01:34  Michael Matz <matz@kde.org>

	* services/comm.h: Ja ja, nicht wirklich viel.	Kann man zu dieser
	  Zeit auch nicht erwarten.

2003-10-22 15:13  Michael Matz <matz@kde.org>

	* mon/mon-kde.cc: Make it compile on 8.2

2003-10-22 14:50  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: for now it picks the first scheduler

2003-10-22 11:14  cschum

	* mon/mon-kde.cc: Fix sorting.

2003-10-22 09:36  Stephan Kulow <coolo@kde.org>

	* TODO: moved the file

2003-10-21 21:19  Stephan Kulow <coolo@kde.org>

	* TODO: note

2003-10-21 21:17  Stephan Kulow <coolo@kde.org>

	* daemon/: workit.cpp: return only the exit status

2003-10-21 21:06  Stephan Kulow <coolo@kde.org>

	* client/: main.cpp, remote.cpp: handle failures of preprocessor

2003-10-21 20:39  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: install symlinks

2003-10-21 19:37  Stephan Kulow <coolo@kde.org>

	* daemon/load.cpp, mon/.cvsignore, mon/mon-kde.cc, mon/mon-kde.h,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp:
	  distribute load to the monitors

2003-10-21 17:46  Michael Matz <matz@kde.org>

	* TODO: More stuff

2003-10-21 17:43  Michael Matz <matz@kde.org>

	* services/comm.cpp: Coolo had to go home, hence the bodies from
	  me.

2003-10-21 17:30  Stephan Kulow <coolo@kde.org>

	* TODO, daemon/load.cpp, services/comm.h: more load numbers

2003-10-21 17:27  Michael Matz <matz@kde.org>

	* daemon/main.cpp: Output all available netnames.

2003-10-21 17:27  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h: Factor code and a get_netnames()
	  method.

2003-10-21 17:14  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: removed debug output

2003-10-21 16:37  Stephan Kulow <coolo@kde.org>

	* client/: local.cpp, util.cpp, util.h: ignore tbcompiler links too

2003-10-21 16:23  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am: icemon!

2003-10-21 15:54  Stephan Kulow <coolo@kde.org>

	* TODO, daemon/main.cpp, services/comm.cpp, services/comm.h: fixing
	  two TODOs

2003-10-21 15:28  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: aetsch

2003-10-21 15:26  Stephan Kulow <coolo@kde.org>

	* TODO: checking

2003-10-21 15:25  Stephan Kulow <coolo@kde.org>

	* daemon/: Makefile.am, findmyself.cpp, findmyself.h, main.cpp:
	  finding the binary

2003-10-21 15:22  Michael Matz <matz@kde.org>

	* daemon/main.cpp: Add -n option for giving a netname.

2003-10-21 15:22  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Support searching for
	  a given netname.

2003-10-21 14:34  Michael Matz <matz@kde.org>

	* daemon/load.cpp: Count nice load as idle, so we ignore our own
	  jobs.

2003-10-21 13:42  Stephan Kulow <coolo@kde.org>

	* TODO: updated

2003-10-21 11:54  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc, mon/mon-kde.h, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: trying to catch failed
	  jobs

2003-10-21 08:47  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: don't update on every job change

2003-10-20 22:11  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc: fixing cols

2003-10-20 22:08  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: correct sorting

2003-10-20 21:39  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: don't flicker

2003-10-20 20:55  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: more infos

2003-10-20 20:37  Michael Matz <matz@kde.org>

	* daemon/main.cpp: Make the toplevel SIGCHLD handler not restart
	  some syscalls (select() in our case).

2003-10-20 20:31  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc, mon/mon-kde.h, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: the client is important
	  too

2003-10-20 20:17  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc, mon/mon-kde.h, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: finished

2003-10-20 20:05  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: it updates the GUI

2003-10-20 19:45  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: JobBegin

2003-10-20 19:34  Stephan Kulow <coolo@kde.org>

	* mon/mon-kde.cc, mon/mon-kde.h, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: GetCS message

2003-10-20 19:12  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: handle_ functions

2003-10-20 18:48  Stephan Kulow <coolo@kde.org>

	* mon/: mon-kde.cc, mon-kde.h: integrate the Msgs into the event
	  loop

2003-10-20 18:08  Michael Matz <matz@kde.org>

	* mon/: Makefile.am, mon-kde.cc: Dump some monitor messages.

2003-10-20 18:07  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Monitor messages.

2003-10-20 16:13  Stephan Kulow <coolo@kde.org>

	* daemon/: load.cpp, serve.cpp: nice 5, safer cut offs

2003-10-20 15:11  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Output the supported compiler version
	  when daemon connects.

2003-10-20 14:44  Stephan Kulow <coolo@kde.org>

	* daemon/load.cpp: total=

2003-10-20 14:41  Stephan Kulow <coolo@kde.org>

	* daemon/load.cpp: memory load

2003-10-20 14:08  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: collect and use
	  statistics.

2003-10-20 14:07  Michael Matz <matz@kde.org>

	* daemon/main.cpp: max_load is no more.

2003-10-20 13:23  Stephan Kulow <coolo@kde.org>

	* daemon/load.cpp: compile

2003-10-20 13:21  Stephan Kulow <coolo@kde.org>

	* daemon/Makefile.am, daemon/load.cpp, daemon/load.h,
	  daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: load is 0-1000 now

2003-10-20 11:07  Michael Matz <matz@kde.org>

	* daemon/main.cpp: First build, then commit.

2003-10-20 10:39  Michael Matz <matz@kde.org>

	* daemon/main.cpp: Transmit RSS in KB.

2003-10-15 08:59  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: let handle_* return bool, I always mix if
	  a 1 is good or bad :)

2003-10-14 15:09  Stephan Kulow <coolo@kde.org>

	* services/: comm.cpp, scheduler.cpp: checking for gnu indent style
	  :)

2003-10-13 21:13  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp: fixing get_absname

2003-10-13 20:37  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: don't try to div 0

2003-10-13 20:28  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: cooler debug output

2003-10-13 19:41  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/scheduler.cpp: removing some debug
	  output - fixing select in daemon

2003-10-13 17:50  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, services/comm.cpp,
	  services/comm.h: more infos in JobDoneMsg

2003-10-13 17:33  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, daemon/serve.h,
	  services/comm.cpp, services/comm.h: filling JobDoneMsg with flesh

2003-10-13 14:55  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: const

2003-10-13 14:09  Michael Matz <matz@kde.org>

	* services/: comm.cpp, scheduler.cpp: Make scheduler send a netname
	  (currently hardcoded to "ICECREAM").

2003-10-10 23:00  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: don't leak fds and channels

2003-10-10 22:38  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/scheduler.cpp: correctly spread jobs

2003-10-10 22:27  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, serve.cpp, tempfile.c, tempfile.h, workit.cpp:
	  trying to use the envs

2003-10-10 21:26  Stephan Kulow <coolo@kde.org>

	* TODO, client/main.cpp, client/remote.cpp, daemon/Makefile.am,
	  daemon/environment.cpp, daemon/environment.h, daemon/main.cpp,
	  daemon/serve.cpp, services/comm.cpp, services/comm.h,
	  services/scheduler.cpp: find and distribute the list of
	  environments

2003-10-10 17:52  Michael Matz <matz@kde.org>

	* mon/Makefile.am: srcdir != builddir

2003-10-10 16:54  Stephan Kulow <coolo@kde.org>

	* TODO: notes

2003-10-10 16:38  Stephan Kulow <coolo@kde.org>

	* TODO, client/main.cpp, client/remote.cpp, daemon/workit.cpp,
	  services/comm.cpp, services/comm.h, services/scheduler.cpp:
	  spread the version

2003-10-10 10:36  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying more gnu style :)

2003-10-09 22:29  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp: trying harder to broadcast,
	  but meanwhile I blame the firewall ;(

2003-10-09 21:38  Stephan Kulow <coolo@kde.org>

	* mon/Makefile.am: moc file

2003-10-09 21:27  Stephan Kulow <coolo@kde.org>

	* mon/.cvsignore: shut ushut upp

2003-10-09 21:27  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.ac, mon/.cvsignore, mon/Makefile.am,
	  mon/mon-kde.cc, mon/mon-kde.h: copyright Frerich Raabe

2003-10-09 20:24  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, serve.cpp, serve.h: don't fork off the
	  scheduler

2003-10-09 17:47  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, services/Makefile.am,
	  services/comm.cpp, services/scheduler.cpp: trying to cope better
	  with errors in the scheduler

2003-10-09 16:55  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, serve.cpp: trying to be a bit smarter in error
	  handling

2003-10-09 16:04  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp: don't wait too long for the
	  scheduler

2003-10-09 16:02  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp, daemon/serve.cpp,
	  services/comm.cpp, services/job.cpp, services/scheduler.cpp:
	  reconnect to scheduler

2003-10-09 14:07  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: better broadcast

2003-10-09 14:06  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Grmbl.

2003-10-08 22:13  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/scheduler.cpp: found
	  the problem: the scheduler queues getcs requests now.  I put a
	  limit of 100 unhandled requests in there, after that it stops
	  listening to the crowd :)

2003-10-08 18:25  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: trying to get cs request queueing work -
	  so far no luck

2003-10-08 17:37  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: removing wrong comment

2003-10-08 06:50  Michael Matz <matz@kde.org>

	* services/comm.h: cosmetic

2003-10-07 22:03  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: shuffling the code a bit

2003-10-07 21:45  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp, client/remote.cpp, daemon/main.cpp,
	  daemon/serve.cpp, daemon/serve.h, services/.cvsignore,
	  services/Makefile.am, services/comm.cpp, services/comm.h,
	  services/receiver.cpp, services/scheduler.cpp,
	  services/submitter.cpp: couple MsgChannel and Service much closer
	  together so you don't need to save two pointers for not leaking

2003-10-07 17:42  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, workit.cpp: cool progress

2003-10-07 17:09  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, workit.cpp: less debug output less leaking
	  sockets

2003-10-07 17:03  Stephan Kulow <coolo@kde.org>

	* client/local.cpp: even less debug output

2003-10-07 16:55  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: removing debug output

2003-10-07 16:52  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp, client/cpp.cpp, client/remote.cpp,
	  services/comm.cpp: taking out some client debug output

2003-10-07 16:40  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: make two iceccds on one
	  machine work

2003-10-07 15:00  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: Randomize compile host.

2003-10-07 14:59  Stephan Kulow <coolo@kde.org>

	* services/scheduler.cpp: debug output for stats

2003-10-07 14:51  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, services/comm.cpp, services/comm.h: every two
	  seconds the load

2003-10-07 13:42  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: no more teambuilder

2003-10-07 13:21  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: off by one

2003-10-07 11:06  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, daemon/serve.h,
	  services/comm.h, services/scheduler.cpp: implemented
	  Job{Begin,End}Msg

2003-10-07 10:16  Stephan Kulow <coolo@kde.org>

	* configure.ac, daemon/main.cpp, daemon/serve.cpp,
	  daemon/workit.cpp, minilzo/minilzo.c, services/comm.cpp,
	  services/comm.h: fixing memory leaks

2003-10-06 23:16  Michael Matz <matz@kde.org>

	* services/comm.cpp: close fds, and some timeout while looking for
	  a scheduler.

2003-10-06 23:01  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h: Deleting memory on stack is not very
	  segfault-avoiding, even if it removes leaks.

2003-10-06 22:53  Michael Matz <matz@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, daemon/serve.h,
	  services/comm.cpp, services/comm.h, services/receiver.cpp,
	  services/scheduler.cpp, services/submitter.cpp: broadcast asking
	  for scheduler

2003-10-06 22:15  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: memleak--

2003-10-06 22:14  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp, daemon/serve.cpp, services/comm.cpp,
	  services/comm.h: trying to fix mem leaks

2003-10-06 22:00  Stephan Kulow <coolo@kde.org>

	* services/Makefile.am: do not install

2003-10-06 22:00  Stephan Kulow <coolo@kde.org>

	* daemon/Makefile.am: install

2003-10-06 21:59  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/serve.cpp: bigger chunks

2003-10-06 21:39  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: not needed

2003-10-06 21:27  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.ac, client/cpp.cpp, client/remote.cpp,
	  daemon/serve.cpp, minilzo/.cvsignore, minilzo/Makefile.am,
	  minilzo/README.LZO, minilzo/lzoconf.h, minilzo/minilzo.c,
	  minilzo/minilzo.h, services/Makefile.am, services/comm.cpp:
	  compression using minilzo

2003-10-06 19:33  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: less debug

2003-10-06 19:01  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/serve.cpp, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: make it work again with
	  GET_SCHEDULER

2003-10-06 18:46  Stephan Kulow <coolo@kde.org>

	* services/.cvsignore: ignore

2003-10-06 18:45  Stephan Kulow <coolo@kde.org>

	* services/comm.cpp: compile (on 8.2)

2003-10-06 18:09  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp, daemon/serve.cpp,
	  daemon/serve.h, services/Makefile.am, services/comm.cpp,
	  services/comm.h, services/scheduler.cpp: trying to make
	  M_*_SCHEDULER work

2003-10-06 17:11  Stephan Kulow <coolo@kde.org>

	* daemon/main.cpp: login

2003-10-06 17:03  Stephan Kulow <coolo@kde.org>

	* configure.ac: nonsense

2003-10-06 16:58  Stephan Kulow <coolo@kde.org>

	* client/remote.cpp, daemon/main.cpp, daemon/serve.cpp,
	  daemon/serve.h, services/comm.cpp, services/comm.h,
	  services/job.cpp: it compiles again

2003-10-06 16:14  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, receiver.cpp, scheduler.cpp,
	  submitter.cpp: New msg type M_LOGIN.	scheduler is now select()
	  based and accepts multiple connections.

2003-10-06 15:09  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am, client/remote.cpp, daemon/main.cpp,
	  services/comm.cpp, services/comm.h: using MsgChannel

2003-10-06 14:27  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp, submitter.cpp: Add
	  port member to UseCSMsg (and Service).  Currently it's hardcoded
	  to 10245 (in the Message only).

2003-10-06 13:48  Stephan Kulow <coolo@kde.org>

	* configure.ac: disable static libs

2003-10-06 13:46  Stephan Kulow <coolo@kde.org>

	* services/.cvsignore: ignore

2003-10-06 13:26  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, receiver.cpp, submitter.cpp:
	  Factoring out the connection to a know host function (now it's a
	  Service constructor).

2003-10-06 12:11  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, receiver.cpp, submitter.cpp: They
	  talk.

2003-10-06 10:24  Stephan Kulow <coolo@kde.org>

	* services/: Makefile.am, comm.cpp, comm.h: removing unneeded
	  messages

2003-10-06 08:55  Stephan Kulow <coolo@kde.org>

	* daemon/.cvsignore: ignoring

2003-10-06 08:43  Stephan Kulow <coolo@kde.org>

	* client/: arg.cpp, cpp.cpp, main.cpp, util.cpp: fixincludes

2003-10-06 08:39  Stephan Kulow <coolo@kde.org>

	* client/util.cpp: removing unneeded stuff

2003-10-06 08:34  Stephan Kulow <coolo@kde.org>

	* client/cpp.cpp: not needed

2003-10-04 18:18  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am, daemon/main.cpp, daemon/serve.cpp,
	  daemon/workit.cpp: little debugging

2003-10-03 17:42  Stephan Kulow <coolo@kde.org>

	* Makefile.am, client/Makefile.am, client/arg.cpp, client/cpp.cpp,
	  client/local.cpp, client/main.cpp, client/remote.cpp,
	  daemon/Makefile.am, daemon/main.cpp, daemon/ncpus.c,
	  daemon/serve.cpp, daemon/tempfile.c, daemon/workit.cpp,
	  daemon/workit.h, services/Makefile.am, services/exitcode.h,
	  services/job.cpp, services/job.h, services/logging.h: making the
	  client smarter and the daemon dumber

2003-10-02 17:37  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp, submitter.cpp: I can
	  connect to it, and get an answer to use host "127.0.0.1"

2003-10-01 18:17  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp, submitter.cpp: Small
	  non-working test program.

2003-10-01 16:37  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Compile!	(plus a
	  non-tested main function for the scheduler)

2003-09-30 23:15  Stephan Kulow <coolo@kde.org>

	* client/: Makefile.am, cpp.cpp, main.cpp: some more cleanup

2003-09-30 23:15  Stephan Kulow <coolo@kde.org>

	* client/arg.cpp: some debug and some strip

2003-09-30 22:39  Stephan Kulow <coolo@kde.org>

	* client/: Makefile.am, arg.cpp, main.cpp: playing around with args

2003-09-30 21:59  Stephan Kulow <coolo@kde.org>

	* client/: Makefile.am, util.cpp, util.h: some more cleaning up

2003-09-30 21:29  Stephan Kulow <coolo@kde.org>

	* services/.cvsignore: shutup

2003-09-30 17:56  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: compile

2003-09-30 17:55  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.ac, client/Makefile.am, client/arg.cpp,
	  client/cpp.cpp, client/main.cpp, client/util.cpp, client/util.h,
	  daemon/Makefile.am, daemon/serve.cpp: moving common stuff into
	  client to sort it out there

2003-09-30 13:36  Stephan Kulow <coolo@kde.org>

	* services/.cvsignore, services/Makefile.am, configure.ac: giving
	  it a Makefile, even when it doesn't compile

2003-09-30 13:17  Michael Matz <matz@kde.org>

	* daemon/workit.cpp: Add volatile.

2003-09-30 11:55  Stephan Kulow <coolo@kde.org>

	* daemon/: serve.cpp, workit.cpp, workit.h: zombies ade

2003-09-29 23:34  Michael Matz <matz@kde.org>

	* services/: comm.cpp, scheduler.cpp: Some TODO, and JobDone
	  handling.

2003-09-29 23:13  Michael Matz <matz@kde.org>

	* services/: comm.cpp, comm.h, scheduler.cpp: Communication stuff
	  and some messages.

2003-09-29 22:14  Stephan Kulow <coolo@kde.org>

	* daemon/workit.cpp: I'm getting nowhere

2003-09-29 20:55  Stephan Kulow <coolo@kde.org>

	* configure.ac, daemon/serve.cpp, daemon/workit.cpp,
	  daemon/workit.h: first work on forking

2003-09-29 18:04  Michael Matz <matz@kde.org>

	* services/comm.h: Some more stuff

2003-09-29 18:04  Michael Matz <matz@kde.org>

	* services/comm.cpp: Seleton of the message sender/receivers.

2003-09-29 16:44  Michael Matz <matz@kde.org>

	* services/scheduler.cpp: skeleton of a scheduler

2003-09-29 16:44  Michael Matz <matz@kde.org>

	* services/comm.h: Some more messages I think I'll need.

2003-09-29 15:58  Stephan Kulow <coolo@kde.org>

	* configure.ac, daemon/Makefile.am, daemon/serve.cpp,
	  daemon/workit.cpp, daemon/workit.h, services/job.h: some more
	  work on CompileJob

2003-09-29 15:22  Stephan Kulow <coolo@kde.org>

	* services/job.h: some thinking

2003-09-28 21:17  Stephan Kulow <coolo@kde.org>

	* client/: .cvsignore, Makefile.am: ignoring

2003-09-26 17:17  Stephan Kulow <coolo@kde.org>

	* daemon/serve.cpp: remove verbose mode

2003-09-26 17:16  Stephan Kulow <coolo@kde.org>

	* daemon/.cvsignore: ignore

2003-09-26 17:16  Stephan Kulow <coolo@kde.org>

	* configure.ac, daemon/main.cpp, daemon/serve.cpp: debug

2003-09-26 16:52  Stephan Kulow <coolo@kde.org>

	* daemon/: Makefile.am, serve.cpp, tempfile.c, tempfile.h: new
	  files

2003-09-26 14:39  Stephan Kulow <coolo@kde.org>

	* daemon/: Makefile.am, serve.cpp: one step closer

2003-09-26 13:49  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am, daemon/.cvsignore, daemon/main.cpp,
	  daemon/serve.cpp: some more C<->CD proto work

2003-09-26 12:58  Stephan Kulow <coolo@kde.org>

	* daemon/: main.cpp, serve.cpp, serve.h: little

2003-09-26 10:25  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.ac: adding daemon

2003-09-26 10:25  Stephan Kulow <coolo@kde.org>

	* daemon/: .cvsignore, Makefile.am, main.cpp, ncpus.c, ncpus.h,
	  serve.cpp, serve.h: I better commit before it becomes serious

2003-09-25 22:04  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: provide a icecc++ too

2003-09-25 17:29  Stephan Kulow <coolo@kde.org>

	* client/Makefile.am: continue from here

2003-09-25 15:44  Michael Matz <matz@kde.org>

	* services/comm.h: tcpdump support

2003-09-25 15:21  Michael Matz <matz@kde.org>

	* services/comm.h: First thoughts.

2003-09-25 14:35  Stephan Kulow <coolo@kde.org>

	* .cvsignore: ignore all links

2003-09-25 14:26  Stephan Kulow <coolo@kde.org>

	* client/: .cvsignore, Makefile.am: renamed

2003-09-25 14:24  Stephan Kulow <coolo@kde.org>

	* Makefile.am, configure.ac, client/.cvsignore, client/Makefile.am:
	  more files to build

2003-09-25 13:46  Stephan Kulow <coolo@kde.org>

	* .cvsignore, Makefile.am, Makefile.cvs, configure.ac: framework

2003-09-25 12:58  Stephan Kulow <coolo@kde.org>

	* daemon/ncpus.c: Initial revision

2003-09-25 12:58  Stephan Kulow <coolo@kde.org>

	* daemon/ncpus.c: importing exploded distcc sources

