marco 1.29.1
Window manager
Loading...
Searching...
No Matches
Bug List
Struct _MetaDrawOpList
Do we really win anything from not using the equivalent GLib structures?
Struct _MetaFrameLayout
button_sizing isn't really necessary, because we could easily say that if button_aspect is zero, the height and width are fixed values. This would also mean that MetaButtonSizing didn't need to exist, and save code.
Global _MetaFrameLayout::title_border
(blue?!)
Struct _MetaTheme
It is rather useless to keep the metadata fields in core, I think.
Global _MetaTheme::filename
Kept lying around for no discernable reason.
Global main (int argc, char **argv)
It's a bit long. It would be good to split it out into separate functions.
Global meta_bell_init (MetaDisplay *display)
There is a line of code that's never run that tells XKB to reset the bell status after we quit. Bill H said (http://bugzilla.gnome.org/show_bug.cgi?id=99886#c12) that XFree86's implementation is broken so we shouldn't call it, but that was in 2002. Is it working now?
Global meta_bell_shutdown (MetaDisplay *display)
This is never called! If we had XkbSetAutoResetControls enabled in meta_bell_init(), this wouldn't be a problem, but we don't.
Global meta_display_ping_window (MetaDisplay *display, MetaWindow *window, guint32 timestamp, MetaWindowPingFunc ping_reply_func, MetaWindowPingFunc ping_timeout_func, gpointer user_data)
This should probably be a method on windows, rather than displays for one of their windows.
Global meta_display_window_has_pending_pings (MetaDisplay *display, MetaWindow *window)
This should probably be a method on windows, rather than displays for one of their windows.
Global meta_is_syncing (void)
This is only called by meta_display_open, but by that time we have already turned syncing on or off on startup, and we don't have any way to do so while Marco is running, so it's rather pointless.
Global meta_set_syncing (gboolean setting)
Of course there is only one display ever anyway, so this can be rather hugely simplified.
Global meta_stack_get_default_focus_window (MetaStack *stack, MetaWorkspace *workspace, MetaWindow *not_this_one)
Never called!
Global meta_stack_update_layer (MetaStack *stack, MetaWindow *window)
What's with the dummy parameter?
Global meta_stack_update_transient (MetaStack *stack, MetaWindow *window)
What's with the dummy parameter?
Global MetaButtonSizing
This could be done away with; see the comment at the top of MetaFrameLayout.
Global parse_number (const char *p, const char **end_return, PosToken *next, GError **err)

The "while (*start)..." part: what's wrong with strchr-ish things?

The name is wrong: it doesn't parse anything.

Global pos_eval (MetaDrawSpec *spec, const MetaPositionExprEnv *env, int *val_p, GError **err)

Shouldn't spec be const?

Global pos_eval_get_variable (PosToken *t, int *result, const MetaPositionExprEnv *env, GError **err)

shouldn't t be const?

we should perhaps consider some sort of lookup arrangement into an array; also, the duplication of code is unlovely; perhaps using glib string hashes instead of quarks would fix both problems?

Global pos_eval_helper (PosToken *tokens, int n_tokens, const MetaPositionExprEnv *env, PosExpr *result, GError **err)

Yes, we really do reparse the expression every time it's evaluated. We should keep the parse tree around all the time and just run the new values through it.

Struct PosExpr

operator is char; it should really be of PosOperatorType.

Global prefs_changed_callback (MetaPreference pref, gpointer data)
Why are these particular prefs handled in main.c and not others? Should they be?
File theme.c
This is a big file with lots of different subsystems, which might be better split out into separate files.
Module tokenizer
We could remove almost all this code by using GScanner instead, but we would also have to find every expression in every existing theme we could and make sure the parse trees were the same.