2002-03-17 Havoc Pennington <hp@pobox.com>
* src/resizepopup.c: Add some total crackrock resize-grid
indication for windows that have width_inc/height_inc
so I can debug gnome-terminal sizing.
2002-03-17 Havoc Pennington <hp@pobox.com>
* src/session.c (set_clone_restart_commands): use proper property
name for SmDiscardCommand (instead of setting the clone command to
"rm"). Also fix typo that iterated over clonev not discardv to
fill in prop list, and NULL-terminate discardv. #74584 from Kang
Jeong-Hee.
2002-03-13 Havoc Pennington <hp@pobox.com>
* src/session.c (meta_session_init): don't save a file here, only
in response to SaveYourself. Change the code to properly use a
unique state file for each SaveYourself. Totally, totally
untested.
2002-03-11 Havoc Pennington <hp@pobox.com>
* src/keybindings.c: use new functions
* src/display.c (meta_display_get_tab_next):
(meta_display_get_tab_list): new tab order functions using
MRU list instead of map order
* src/window.c (meta_window_notify_focus): maintain focus MRU list
* src/display.h (struct _MetaDisplay): Keep an MRU list of
windows.
2002-03-10 Havoc Pennington <hp@pobox.com>
* src/display.c (event_callback): support _NET_NUMBER_OF_DESKTOPS
message so you can change number of desktops with the pager
* src/prefs.c (meta_prefs_set_num_workspaces): new function
* src/display.c (meta_spew_event): print stacking aspects of
configure request
2002-03-10 Havoc Pennington <hp@pobox.com>
* src/screen.c (set_supported_hint): we didn't claim to support
_NET_ACTIVE_WINDOW so gtk_window_present() didn't work. Mumble.
Only worked with tasklist because libwnck didn't check for
WM support.
* src/window.c (meta_window_free): clean off window state
when windows are withdrawn, to avoid sticking dialogs
to their initial desktop.
(meta_window_queue_calc_showing): return if window is withdrawn
2002-03-06 Havoc Pennington <hp@pobox.com>
* src/core.c (meta_core_get_grab_frame): add some assertions
* src/menu.c (meta_window_menu_new): make another warning
into a verbose
* src/display.c (meta_change_button_grab): use verbose rather than
warning to log failures to grab button, since these are typically
BadWindow from a destroyed window.
2002-03-06 Havoc Pennington <hp@redhat.com>
* src/frames.c (meta_frames_manage_window): use hash_table_replace
instead of g_hash_table_insert
* src/main.c (main): only enable verbose/debug if you set
METACITY_VERBOSE/METACITY_DEBUG
* src/util.c (ensure_logfile): only use a log file if
METACITY_USE_LOGFILE is set
* src/display.c (meta_display_for_x_display): add warning if
MetaDisplay isn't found
* src/window.c (meta_window_free): add an assertion that we
successfully cleared the grab window
2002-03-05 Havoc Pennington <hp@pobox.com>
Work on opaque animations more, still suck too much
to turn on. Not sure how to make them good.
* src/effects.c (meta_effects_draw_box_animation):
add a slide-up mode for shading
* src/ui.c (meta_image_window_set): change image window to work by
setting back pixmap on the GtkWindow, instead of using GtkImage.
2002-03-04 Havoc Pennington <hp@pobox.com>
* src/main.c (main): try ignoring SIGXFSZ, though I'm not
sure what that does exactly. I'm hoping it gives me EFBIG.
* src/util.c (ensure_logfile): log to a file in /tmp instead
of to ~/metacity.log.
2002-03-02 Havoc Pennington <hp@pobox.com>
* src/window.c (recalc_window_features): disable resize etc. if
we're fullscreen
(constrain_size): fix size constraints when fullscreen
* src/display.c (meta_display_open): fix missing comma that
ended up concatenating two of the properties breaking
FULLSCREEN state and PING protocol
2002-03-02 Havoc Pennington <hp@pobox.com>
* src/display.c: Add hacking to fix the problem that we made our
XGrabPointer() during Alt+Tab actually succeed, so on popping down
Alt+Tab we got an EnterNotify from the ungrab, which resulted in
focusing the window under the mouse. i.e. Alt+Tab didn't work with
sloppy focus.
2002-02-26 Havoc Pennington <hp@pobox.com>
Screw around with Anders's ping patch so he'll get plenty of CVS
conflicts. ;-)
* src/display.c (meta_display_ping_window): spew warnings
if we try to call this with CurrentTime
(meta_display_ping_timeout): remove ping from the pending pings
after it times out.
* src/util.h: added PING debug category
* src/display.c (remove_pending_pings_for_window): don't remove
"tmp" just before "tmp->next", don't break out of loop after
finding the first match
(meta_display_open): no trailing comma in array init
(event_callback): move the processing of ping replies into a
separate function
* src/screen.c (set_supported_hint): add _NET_WM_PING to supported
list
* src/display.h: change gpointer to void*
2002-02-26 Anders Carlsson <andersca@gnu.org>
* src/display.c: (ping_data_free),
(remove_pending_pings_for_window), (meta_display_open),
(event_callback), (meta_display_unregister_x_window),
(meta_display_ping_timeout), (meta_display_ping_window),
(meta_display_window_has_pending_pings):
Implement meta_display_ping_window, and filter out scroll wheel
events.
* src/display.h:
Add MetaWindowPingFunc, meta_display_ping_window and
meta_display_window_has_pending_pings.
2002-02-24 Havoc Pennington <hp@pobox.com>
* src/display.c (xcursor_for_op): switch on the op passed in, not
the active op. Gives us the right cursor during resizing, etc.
* src/errors.c: rearrange all the error stuff to adapt to the GDK
change a while back, so now we print our X errors again
* src/display.c (meta_display_begin_grab_op): remove KeyPressMask
and KeyReleaseMask from the XGrabPointer(), this caused BadValue
and kept the grab from ever succeeding. Fixes the problem with the
GTK resize grip - this is why you shouldn't break your X error
spew. ;-)
* src/display.c: debug spew tweaks
* src/window.c (meta_window_client_message): do some
s/verbose/topic/ stuff
2002-02-19 Havoc Pennington <hp@pobox.com>
* src/ui.c (meta_ui_init): put in hack to keep Pango from mangling
our server grab and locking up on startup. (hack doesn't work
but I want to fix it on my real computer not this laptop)
* src/window.c: Implement _NET_WM_STATE_FULLSCREEN
* src/display.c (meta_display_open): add atoms for
_NET_WM_STATE_FULLSCREEN
2002-02-14 Havoc Pennington <hp@pobox.com>
* src/theme-viewer.c: use the preview widget here
* src/preview-widget.h, src/preview-widget.c: make the theme
preview into a nice widget
* src/frames.c (meta_frames_ensure_layout): replace frame layout
if the frame style changes, this only ends up mattering if you
e.g. changed the font size for windows in a different state such
as maximized, which is crack, but the code may as well be correct
* src/theme.c (meta_theme_get_frame_style): new function so we can
detect an invalid cache of the PangoLayout in a frame
2002-02-13 Anders Carlsson <andersca@gnu.org>
* src/theme.c (meta_pango_font_desc_get_text_height): Use
pango_context_get_metrics instead of loading the font.
2002-02-12 Anders Carlsson <andersca@gnu.org>
* src/frames.c (meta_frames_manage_window): Set prelit_control
to META_FRAME_CONTROL_NONE.
(meta_frames_update_prelit_control): New function for setting
the prelit control.
(meta_frames_paint_to_drawable): Set prelight state.
(meta_frames_enter_notify_event): Update prelit control.
(meta_frames_leave_notify_event): Likewise.
(meta_frames_motion_notify_event): Likewise.
* src/frames.h (struct _MetaUIFrame): add prelit_control.
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/theme.c (meta_theme_set_current): add a newline to an error
message
* src/themes/Gorilla: add Gorilla theme by Jakub Steiner ported to
metacity by Kenneth Christiansen
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/theme.c (meta_draw_op_draw_with_env): implement wacky "tile"
draw op to lose some of the PNG files in Gorilla theme
* src/theme-parser.c: parse the tile primitive
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/window.c (update_icon): port to icon cache
* src/iconcache.c, src/iconcache.c: begin process of cleaning up
window.c by moving the icon-reading code in here, based on the
code in libwnck, which was in turn based on the earlier metacity
code
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/stack.c (meta_stack_sync_to_server): hmm, and don't set
last_window at all if we don't ++newp. Fixes even more obscure
stacking bug.
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/stack.c (meta_stack_sync_to_server): assign last_window
prior to ++newp, so we don't try to stack windows with respect to
themselves. Fixes some obscure stacking bugs.
