glib/gio/glocalfilemonitor.c

935 lines
29 KiB
C
Raw Normal View History

/* GIO - GLib Input, Output and Streaming Library
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
*
* Copyright (C) 2006-2007 Red Hat, Inc.
*
* SPDX-License-Identifier: LGPL-2.1-or-later
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General
2014-01-23 12:58:29 +01:00
* Public License along with this library; if not, see <http://www.gnu.org/licenses/>.
*
* Author: Alexander Larsson <alexl@redhat.com>
*/
#include "config.h"
#include "gioenumtypes.h"
#include "glocalfilemonitor.h"
#include "giomodule-priv.h"
#include "gioerror.h"
#include "glibintl.h"
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
#include "glocalfile.h"
#include "glib-private.h"
#include <string.h>
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
#define DEFAULT_RATE_LIMIT 800 * G_TIME_SPAN_MILLISECOND
#define VIRTUAL_CHANGES_DONE_DELAY 2 * G_TIME_SPAN_SECOND
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
/* GFileMonitorSource is a GSource responsible for emitting the changed
* signals in the owner-context of the GFileMonitor.
*
* It contains functionality for cross-thread queuing of events. It
* also handles merging of CHANGED events and emission of CHANGES_DONE
* events.
*
* We use the "priv" pointer in the external struct to store it.
*/
struct _GFileMonitorSource {
GSource source;
GMutex lock;
GWeakRef instance_ref;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
GFileMonitorFlags flags;
gchar *dirname;
gchar *basename;
gchar *filename;
GSequence *pending_changes; /* sorted by ready time */
GHashTable *pending_changes_table;
GQueue event_queue;
gint64 rate_limit;
};
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
/* PendingChange is a struct to keep track of a file that needs to have
* (at least) a CHANGES_DONE_HINT event sent for it in the near future.
*
* If 'dirty' is TRUE then a CHANGED event also needs to be sent.
*
* last_emission is the last time a CHANGED event was emitted. It is
* used to calculate the time to send the next event.
*/
typedef struct {
gchar *child;
guint64 last_emission : 63;
guint64 dirty : 1;
} PendingChange;
/* QueuedEvent is a signal that will be sent immediately, as soon as the
* source gets a chance to dispatch. The existence of any queued event
* implies that the source is ready now.
*/
typedef struct
{
GFileMonitorEvent event_type;
GFile *child;
GFile *other;
} QueuedEvent;
static gint64
pending_change_get_ready_time (const PendingChange *change,
GFileMonitorSource *fms)
{
if (change->dirty)
return change->last_emission + fms->rate_limit;
else
return change->last_emission + VIRTUAL_CHANGES_DONE_DELAY;
}
static int
pending_change_compare_ready_time (gconstpointer a_p,
gconstpointer b_p,
gpointer user_data)
{
GFileMonitorSource *fms = user_data;
const PendingChange *a = a_p;
const PendingChange *b = b_p;
gint64 ready_time_a;
gint64 ready_time_b;
ready_time_a = pending_change_get_ready_time (a, fms);
ready_time_b = pending_change_get_ready_time (b, fms);
if (ready_time_a < ready_time_b)
return -1;
else
return ready_time_a > ready_time_b;
}
static void
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
pending_change_free (gpointer data)
{
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
PendingChange *change = data;
g_free (change->child);
g_slice_free (PendingChange, change);
}
static void
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
queued_event_free (QueuedEvent *event)
{
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_object_unref (event->child);
if (event->other)
g_object_unref (event->other);
g_slice_free (QueuedEvent, event);
}
static gint64
g_file_monitor_source_get_ready_time (GFileMonitorSource *fms)
{
GSequenceIter *iter;
if (fms->event_queue.length)
return 0;
iter = g_sequence_get_begin_iter (fms->pending_changes);
if (g_sequence_iter_is_end (iter))
return -1;
return pending_change_get_ready_time (g_sequence_get (iter), fms);
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
static void
g_file_monitor_source_update_ready_time (GFileMonitorSource *fms)
{
g_source_set_ready_time ((GSource *) fms, g_file_monitor_source_get_ready_time (fms));
}
static GSequenceIter *
g_file_monitor_source_find_pending_change (GFileMonitorSource *fms,
const gchar *child)
{
return g_hash_table_lookup (fms->pending_changes_table, child);
}
static void
g_file_monitor_source_add_pending_change (GFileMonitorSource *fms,
const gchar *child,
gint64 now)
{
PendingChange *change;
GSequenceIter *iter;
change = g_slice_new (PendingChange);
change->child = g_strdup (child);
change->last_emission = now;
change->dirty = FALSE;
iter = g_sequence_insert_sorted (fms->pending_changes, change, pending_change_compare_ready_time, fms);
g_hash_table_insert (fms->pending_changes_table, change->child, iter);
}
static gboolean
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_set_pending_change_dirty (GFileMonitorSource *fms,
GSequenceIter *iter)
{
PendingChange *change;
change = g_sequence_get (iter);
/* if it was already dirty then this change is 'uninteresting' */
if (change->dirty)
return FALSE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
change->dirty = TRUE;
g_sequence_sort_changed (iter, pending_change_compare_ready_time, fms);
return TRUE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
static gboolean
g_file_monitor_source_get_pending_change_dirty (GFileMonitorSource *fms,
GSequenceIter *iter)
{
PendingChange *change;
change = g_sequence_get (iter);
return change->dirty;
}
static void
g_file_monitor_source_remove_pending_change (GFileMonitorSource *fms,
GSequenceIter *iter,
const gchar *child)
{
/* must remove the hash entry first -- its key is owned by the data
* which will be freed when removing the sequence iter
*/
g_hash_table_remove (fms->pending_changes_table, child);
g_sequence_remove (iter);
}
static void
g_file_monitor_source_queue_event (GFileMonitorSource *fms,
GFileMonitorEvent event_type,
const gchar *child,
GFile *other)
{
QueuedEvent *event;
event = g_slice_new (QueuedEvent);
event->event_type = event_type;
if (child != NULL && fms->dirname != NULL)
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
event->child = g_local_file_new_from_dirname_and_basename (fms->dirname, child);
else if (child != NULL)
{
gchar *dirname = g_path_get_dirname (fms->filename);
event->child = g_local_file_new_from_dirname_and_basename (dirname, child);
g_free (dirname);
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
else if (fms->dirname)
event->child = _g_local_file_new (fms->dirname);
else if (fms->filename)
event->child = _g_local_file_new (fms->filename);
event->other = other;
if (other)
g_object_ref (other);
g_queue_push_tail (&fms->event_queue, event);
}
static gboolean
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_file_changed (GFileMonitorSource *fms,
const gchar *child,
gint64 now)
{
GSequenceIter *pending;
gboolean interesting;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
pending = g_file_monitor_source_find_pending_change (fms, child);
/* If there is no pending change, emit one and create a record,
* else: just mark the existing record as dirty.
*/
if (!pending)
{
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGED, child, NULL);
g_file_monitor_source_add_pending_change (fms, child, now);
interesting = TRUE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
else
interesting = g_file_monitor_source_set_pending_change_dirty (fms, pending);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_update_ready_time (fms);
return interesting;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
static void
g_file_monitor_source_file_changes_done (GFileMonitorSource *fms,
const gchar *child)
{
GSequenceIter *pending;
pending = g_file_monitor_source_find_pending_change (fms, child);
if (pending)
{
/* If it is dirty, make sure we push out the last CHANGED event */
if (g_file_monitor_source_get_pending_change_dirty (fms, pending))
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGED, child, NULL);
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT, child, NULL);
g_file_monitor_source_remove_pending_change (fms, pending, child);
}
}
static void
g_file_monitor_source_file_created (GFileMonitorSource *fms,
const gchar *child,
gint64 event_time)
{
/* Unlikely, but if we have pending changes for this filename, make
* sure we flush those out first, before creating the new ones.
*/
g_file_monitor_source_file_changes_done (fms, child);
/* Emit CREATE and add a pending changes record */
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CREATED, child, NULL);
g_file_monitor_source_add_pending_change (fms, child, event_time);
}
static void
g_file_monitor_source_send_event (GFileMonitorSource *fms,
GFileMonitorEvent event_type,
const gchar *child,
GFile *other)
{
/* always flush any pending changes before we queue a new event */
g_file_monitor_source_file_changes_done (fms, child);
g_file_monitor_source_queue_event (fms, event_type, child, other);
}
static void
g_file_monitor_source_send_synthetic_created (GFileMonitorSource *fms,
const gchar *child)
{
g_file_monitor_source_file_changes_done (fms, child);
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CREATED, child, NULL);
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT, child, NULL);
}
#ifndef G_DISABLE_ASSERT
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
static gboolean
is_basename (const gchar *name)
{
if (name[0] == '.' && ((name[1] == '.' && name[2] == '\0') || name[1] == '\0'))
return FALSE;
return !strchr (name, '/');
}
#endif /* !G_DISABLE_ASSERT */
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
gboolean
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_handle_event (GFileMonitorSource *fms,
GFileMonitorEvent event_type,
const gchar *child,
const gchar *rename_to,
GFile *other,
gint64 event_time)
{
gboolean interesting = TRUE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_assert (!child || is_basename (child));
g_assert (!rename_to || is_basename (rename_to));
if (fms->basename && (!child || !g_str_equal (child, fms->basename))
&& (!rename_to || !g_str_equal (rename_to, fms->basename)))
return TRUE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_mutex_lock (&fms->lock);
/* NOTE:
*
* We process events even if the file monitor has already been disposed.
* The reason is that we must not take a reference to the instance here as
* destroying it from the event handling thread will lead to a deadlock when
* taking the lock in _ih_sub_cancel.
*
* This results in seemingly-unbounded growth of the `event_queue` with the
* calls to `g_file_monitor_source_queue_event()`. However, each of those sets
* the ready time on the #GSource, which means that it will be dispatched in
* a subsequent iteration of the #GMainContext its attached to. At that
* point, `g_file_monitor_source_dispatch()` will return %FALSE, and this will
* trigger finalisation of the source. That will clear the `event_queue`.
*
* If the source is no longer attached, this will return early to prevent
* unbounded queueing.
glocalfilemonitor: Avoid file monitor destruction from event thread Taking a reference to the GFileMonitor when handling events may cause object destruction from th worker thread that calls the function. This condition happens if the surrounding code drops the otherwise last reference ot the GFileMonitor. The series of events causes destruction from an unrelated worker thread and also triggers g_file_monitor_cancel to be called from g_file_monitor_source_handle_event. For the inotify backend, this results in a deadlock as cancellation needs to take a lock that protects data structures from being modified while events are dispatched. One alternative to this approach might be to add an RCU (release, copy, update) approach to the lists contained in the wd_dir_hash and wd_file_hash hash tables. Fixes: #1941 An example stack trace of this happening is: Thread 2 (Thread 0x7fea68b1d640 (LWP 260961) "gmain"): #0 syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38 #1 0x00007fea692215dc in g_mutex_lock_slowpath (mutex=mutex@entry=0x7fea6911e148 <g.inotify_lock_lock>) at ../glib/gthread-posix.c:1493 #2 0x00007fea69222062 in g_mutex_lock (mutex=mutex@entry=0x7fea6911e148 <g.inotify_lock_lock>) at ../glib/gthread-posix.c:1517 #3 0x00007fea6908025a in _ih_sub_cancel (sub=0x1492620) at ../gio/inotify/inotify-helper.c:131 #4 0x00007fea6907f9da in g_inotify_file_monitor_cancel (monitor=0x14a3550) at ../gio/inotify/ginotifyfilemonitor.c:75 #5 0x00007fea68fae959 in g_file_monitor_cancel (monitor=0x14a3550) at ../gio/gfilemonitor.c:241 #6 0x00007fea68fae9dc in g_file_monitor_dispose (object=0x14a3550) at ../gio/gfilemonitor.c:123 #7 0x00007fea69139341 in g_object_unref (_object=<optimized out>) at ../gobject/gobject.c:3636 #8 g_object_unref (_object=0x14a3550) at ../gobject/gobject.c:3553 #9 0x00007fea6907507a in g_file_monitor_source_handle_event (fms=0x14c3560, event_type=<optimized out>, child=0x7fea64001460 "spawned-1", rename_to=rename_to@entry=0x0, other=other@entry=0x0, event_time=<optimized out>) at ../gio/glocalfilemonitor.c:457 #10 0x00007fea6907fe0e in ih_event_callback (event=0x7fea64001420, sub=0x1492620, file_event=<optimized out>) at ../gio/inotify/inotify-helper.c:218 #11 0x00007fea6908075c in ip_event_dispatch (dir_list=dir_list@entry=0x14c14c0, file_list=0x0, event=event@entry=0x7fea64001420) at ../gio/inotify/inotify-path.c:493 #12 0x00007fea6908094e in ip_event_dispatch (event=0x7fea64001420, file_list=<optimized out>, dir_list=0x14c14c0) at ../gio/inotify/inotify-path.c:448 #13 ip_event_callback (event=0x7fea64001420) at ../gio/inotify/inotify-path.c:548 #14 ip_event_callback (event=0x7fea64001420) at ../gio/inotify/inotify-path.c:530 #15 0x00007fea69081391 in ik_source_dispatch (source=0x14a2bf0, func=0x7fea69080890 <ip_event_callback>, user_data=<optimized out>) at ../gio/inotify/inotify-kernel.c:327 #16 0x00007fea691d0824 in g_main_dispatch (context=0x14a2cc0) at ../glib/gmain.c:3417 #17 g_main_context_dispatch (context=0x14a2cc0) at ../glib/gmain.c:4135 #18 0x00007fea691d0b88 in g_main_context_iterate (context=context@entry=0x14a2cc0, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at ../glib/gmain.c:4211 #19 0x00007fea691d0c2f in g_main_context_iteration (context=0x14a2cc0, may_block=may_block@entry=1) at ../glib/gmain.c:4276 #20 0x00007fea691d0c81 in glib_worker_main (data=<optimized out>) at ../glib/gmain.c:6176 #21 0x00007fea691f9c2d in g_thread_proxy (data=0x1487cc0) at ../glib/gthread.c:827 #22 0x00007fea68d93b1a in start_thread (arg=<optimized out>) at pthread_create.c:443 #23 0x00007fea68e18650 in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81
2022-03-18 12:16:12 +01:00
*/
if (g_source_is_destroyed ((GSource *) fms))
{
g_mutex_unlock (&fms->lock);
return TRUE;
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
switch (event_type)
{
case G_FILE_MONITOR_EVENT_CREATED:
g_assert (!other && !rename_to);
g_file_monitor_source_file_created (fms, child, event_time);
break;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
case G_FILE_MONITOR_EVENT_CHANGED:
g_assert (!other && !rename_to);
interesting = g_file_monitor_source_file_changed (fms, child, event_time);
break;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
case G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT:
g_assert (!other && !rename_to);
g_file_monitor_source_file_changes_done (fms, child);
break;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
case G_FILE_MONITOR_EVENT_MOVED_IN:
g_assert (!rename_to);
if (fms->flags & G_FILE_MONITOR_WATCH_MOVES)
g_file_monitor_source_send_event (fms, G_FILE_MONITOR_EVENT_MOVED_IN, child, other);
else
g_file_monitor_source_send_synthetic_created (fms, child);
break;
case G_FILE_MONITOR_EVENT_MOVED_OUT:
g_assert (!rename_to);
if (fms->flags & G_FILE_MONITOR_WATCH_MOVES)
g_file_monitor_source_send_event (fms, G_FILE_MONITOR_EVENT_MOVED_OUT, child, other);
else if (other && (fms->flags & G_FILE_MONITOR_SEND_MOVED))
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_send_event (fms, G_FILE_MONITOR_EVENT_MOVED, child, other);
else
g_file_monitor_source_send_event (fms, G_FILE_MONITOR_EVENT_DELETED, child, NULL);
break;
case G_FILE_MONITOR_EVENT_RENAMED:
g_assert (!other && rename_to);
if (fms->flags & (G_FILE_MONITOR_WATCH_MOVES | G_FILE_MONITOR_SEND_MOVED))
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
{
GFile *other_file;
const gchar *dirname;
gchar *allocated_dirname = NULL;
GFileMonitorEvent event;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
event = (fms->flags & G_FILE_MONITOR_WATCH_MOVES) ? G_FILE_MONITOR_EVENT_RENAMED : G_FILE_MONITOR_EVENT_MOVED;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
if (fms->dirname != NULL)
dirname = fms->dirname;
else
{
allocated_dirname = g_path_get_dirname (fms->filename);
dirname = allocated_dirname;
}
other_file = g_local_file_new_from_dirname_and_basename (dirname, rename_to);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_file_changes_done (fms, rename_to);
g_file_monitor_source_send_event (fms, event, child, other_file);
g_object_unref (other_file);
g_free (allocated_dirname);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
else
{
g_file_monitor_source_send_event (fms, G_FILE_MONITOR_EVENT_DELETED, child, NULL);
g_file_monitor_source_send_synthetic_created (fms, rename_to);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
break;
case G_FILE_MONITOR_EVENT_DELETED:
case G_FILE_MONITOR_EVENT_ATTRIBUTE_CHANGED:
case G_FILE_MONITOR_EVENT_PRE_UNMOUNT:
case G_FILE_MONITOR_EVENT_UNMOUNTED:
g_assert (!other && !rename_to);
g_file_monitor_source_send_event (fms, event_type, child, NULL);
break;
case G_FILE_MONITOR_EVENT_MOVED:
/* was never available in this API */
default:
g_assert_not_reached ();
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_update_ready_time (fms);
g_mutex_unlock (&fms->lock);
return interesting;
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
static gint64
g_file_monitor_source_get_rate_limit (GFileMonitorSource *fms)
{
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
gint64 rate_limit;
g_mutex_lock (&fms->lock);
rate_limit = fms->rate_limit;
g_mutex_unlock (&fms->lock);
return rate_limit;
}
static gboolean
g_file_monitor_source_set_rate_limit (GFileMonitorSource *fms,
gint64 rate_limit)
{
gboolean changed;
g_mutex_lock (&fms->lock);
if (rate_limit != fms->rate_limit)
{
fms->rate_limit = rate_limit;
g_sequence_sort (fms->pending_changes, pending_change_compare_ready_time, fms);
g_file_monitor_source_update_ready_time (fms);
changed = TRUE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
else
changed = FALSE;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_mutex_unlock (&fms->lock);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
return changed;
}
static gboolean
g_file_monitor_source_dispatch (GSource *source,
GSourceFunc callback,
gpointer user_data)
{
GFileMonitorSource *fms = (GFileMonitorSource *) source;
QueuedEvent *event;
GQueue event_queue;
gint64 now;
GFileMonitor *instance = NULL;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
/* make sure the monitor still exists */
instance = g_weak_ref_get (&fms->instance_ref);
if (instance == NULL)
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
return FALSE;
now = g_source_get_time (source);
/* Acquire the lock once and grab all events in one go, handling the
* queued events first. This avoids strange possibilities in cases of
* long delays, such as CHANGED events coming before CREATED events.
*
* We do this by converting the applicable pending changes into queued
* events (after the ones already queued) and then stealing the entire
* event queue in one go.
*/
g_mutex_lock (&fms->lock);
/* Create events for any pending changes that are due to fire */
while (!g_sequence_is_empty (fms->pending_changes))
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
{
GSequenceIter *iter = g_sequence_get_begin_iter (fms->pending_changes);
PendingChange *pending = g_sequence_get (iter);
/* We've gotten to a pending change that's not ready. Stop. */
if (pending_change_get_ready_time (pending, fms) > now)
break;
if (pending->dirty)
{
/* It's time to send another CHANGED and update the record */
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGED, pending->child, NULL);
pending->last_emission = now;
pending->dirty = FALSE;
g_sequence_sort_changed (iter, pending_change_compare_ready_time, fms);
}
else
{
/* It's time to send CHANGES_DONE and remove the pending record */
g_file_monitor_source_queue_event (fms, G_FILE_MONITOR_EVENT_CHANGES_DONE_HINT, pending->child, NULL);
g_file_monitor_source_remove_pending_change (fms, iter, pending->child);
}
}
/* Steal the queue */
memcpy (&event_queue, &fms->event_queue, sizeof event_queue);
memset (&fms->event_queue, 0, sizeof fms->event_queue);
g_file_monitor_source_update_ready_time (fms);
g_mutex_unlock (&fms->lock);
g_clear_object (&instance);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
/* We now have our list of events to deliver */
while ((event = g_queue_pop_head (&event_queue)))
{
/* an event handler could destroy 'instance', so check each time */
instance = g_weak_ref_get (&fms->instance_ref);
if (instance != NULL)
g_file_monitor_emit_event (instance, event->child, event->other, event->event_type);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_clear_object (&instance);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
queued_event_free (event);
}
return TRUE;
}
static void
g_file_monitor_source_dispose (GFileMonitorSource *fms)
{
GHashTableIter iter;
gpointer seqiter;
QueuedEvent *event;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_mutex_lock (&fms->lock);
g_hash_table_iter_init (&iter, fms->pending_changes_table);
while (g_hash_table_iter_next (&iter, NULL, &seqiter))
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
{
g_hash_table_iter_remove (&iter);
g_sequence_remove (seqiter);
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
while ((event = g_queue_pop_head (&fms->event_queue)))
queued_event_free (event);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_assert (g_sequence_is_empty (fms->pending_changes));
g_assert (g_hash_table_size (fms->pending_changes_table) == 0);
g_assert (fms->event_queue.length == 0);
g_weak_ref_set (&fms->instance_ref, NULL);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_file_monitor_source_update_ready_time (fms);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_source_destroy ((GSource *) fms);
g_mutex_unlock (&fms->lock);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
}
static void
g_file_monitor_source_finalize (GSource *source)
{
GFileMonitorSource *fms = (GFileMonitorSource *) source;
/* should already have been cleared in dispose of the monitor */
g_assert (g_weak_ref_get (&fms->instance_ref) == NULL);
g_weak_ref_clear (&fms->instance_ref);
g_assert (g_sequence_is_empty (fms->pending_changes));
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_assert (g_hash_table_size (fms->pending_changes_table) == 0);
g_assert (fms->event_queue.length == 0);
g_hash_table_unref (fms->pending_changes_table);
g_sequence_free (fms->pending_changes);
g_free (fms->dirname);
g_free (fms->basename);
g_free (fms->filename);
g_mutex_clear (&fms->lock);
}
static guint
str_hash0 (gconstpointer str)
{
return str ? g_str_hash (str) : 0;
}
static gboolean
str_equal0 (gconstpointer a,
gconstpointer b)
{
return g_strcmp0 (a, b) == 0;
}
static GFileMonitorSource *
g_file_monitor_source_new (gpointer instance,
const gchar *filename,
gboolean is_directory,
GFileMonitorFlags flags)
{
static GSourceFuncs source_funcs = {
NULL, NULL,
g_file_monitor_source_dispatch,
g_file_monitor_source_finalize,
NULL, NULL
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
};
GFileMonitorSource *fms;
GSource *source;
source = g_source_new (&source_funcs, sizeof (GFileMonitorSource));
fms = (GFileMonitorSource *) source;
g_source_set_static_name (source, "GFileMonitorSource");
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_mutex_init (&fms->lock);
g_weak_ref_init (&fms->instance_ref, instance);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
fms->pending_changes = g_sequence_new (pending_change_free);
fms->pending_changes_table = g_hash_table_new (str_hash0, str_equal0);
fms->rate_limit = DEFAULT_RATE_LIMIT;
fms->flags = flags;
if (is_directory)
{
fms->dirname = g_strdup (filename);
fms->basename = NULL;
fms->filename = NULL;
}
else if (flags & G_FILE_MONITOR_WATCH_HARD_LINKS)
{
fms->dirname = NULL;
fms->basename = NULL;
fms->filename = g_strdup (filename);
}
else
{
fms->dirname = g_path_get_dirname (filename);
fms->basename = g_path_get_basename (filename);
fms->filename = NULL;
}
return fms;
}
G_DEFINE_ABSTRACT_TYPE (GLocalFileMonitor, g_local_file_monitor, G_TYPE_FILE_MONITOR)
enum {
PROP_0,
PROP_RATE_LIMIT,
};
static void
g_local_file_monitor_get_property (GObject *object, guint prop_id,
GValue *value, GParamSpec *pspec)
{
GLocalFileMonitor *monitor = G_LOCAL_FILE_MONITOR (object);
gint64 rate_limit;
g_assert (prop_id == PROP_RATE_LIMIT);
rate_limit = g_file_monitor_source_get_rate_limit (monitor->source);
rate_limit /= G_TIME_SPAN_MILLISECOND;
g_value_set_int (value, rate_limit);
}
static void
g_local_file_monitor_set_property (GObject *object, guint prop_id,
const GValue *value, GParamSpec *pspec)
{
GLocalFileMonitor *monitor = G_LOCAL_FILE_MONITOR (object);
gint64 rate_limit;
g_assert (prop_id == PROP_RATE_LIMIT);
rate_limit = g_value_get_int (value);
rate_limit *= G_TIME_SPAN_MILLISECOND;
if (g_file_monitor_source_set_rate_limit (monitor->source, rate_limit))
g_object_notify (object, "rate-limit");
}
#ifndef G_OS_WIN32
static void
g_local_file_monitor_mounts_changed (GUnixMountMonitor *mount_monitor,
gpointer user_data)
{
GLocalFileMonitor *local_monitor = user_data;
GUnixMountEntry *mount;
gboolean is_mounted;
GFile *file;
/* Emulate unmount detection */
mount = g_unix_mount_at (local_monitor->source->dirname, NULL);
is_mounted = mount != NULL;
if (mount)
g_unix_mount_free (mount);
if (local_monitor->was_mounted != is_mounted)
{
if (local_monitor->was_mounted && !is_mounted)
{
file = g_file_new_for_path (local_monitor->source->dirname);
g_file_monitor_emit_event (G_FILE_MONITOR (local_monitor), file, NULL, G_FILE_MONITOR_EVENT_UNMOUNTED);
g_object_unref (file);
}
local_monitor->was_mounted = is_mounted;
}
}
#endif
static void
g_local_file_monitor_start (GLocalFileMonitor *local_monitor,
const gchar *filename,
gboolean is_directory,
GFileMonitorFlags flags,
GMainContext *context)
{
GLocalFileMonitorClass *class = G_LOCAL_FILE_MONITOR_GET_CLASS (local_monitor);
GFileMonitorSource *source;
g_return_if_fail (G_IS_LOCAL_FILE_MONITOR (local_monitor));
g_assert (!local_monitor->source);
source = g_file_monitor_source_new (local_monitor, filename, is_directory, flags);
local_monitor->source = source; /* owns the ref */
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
if (is_directory && !class->mount_notify && (flags & G_FILE_MONITOR_WATCH_MOUNTS))
{
#ifdef G_OS_WIN32
/*claim everything was mounted */
local_monitor->was_mounted = TRUE;
#else
GUnixMountEntry *mount;
/* Emulate unmount detection */
mount = g_unix_mount_at (local_monitor->source->dirname, NULL);
local_monitor->was_mounted = mount != NULL;
if (mount)
g_unix_mount_free (mount);
local_monitor->mount_monitor = g_unix_mount_monitor_get ();
g_signal_connect_object (local_monitor->mount_monitor, "mounts-changed",
G_CALLBACK (g_local_file_monitor_mounts_changed), local_monitor,
G_CONNECT_DEFAULT);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
#endif
}
g_source_attach ((GSource *) source, context);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
G_LOCAL_FILE_MONITOR_GET_CLASS (local_monitor)->start (local_monitor,
source->dirname, source->basename, source->filename,
source);
}
static void
g_local_file_monitor_dispose (GObject *object)
{
GLocalFileMonitor *local_monitor = G_LOCAL_FILE_MONITOR (object);
g_file_monitor_source_dispose (local_monitor->source);
G_OBJECT_CLASS (g_local_file_monitor_parent_class)->dispose (object);
}
static void
g_local_file_monitor_finalize (GObject *object)
{
GLocalFileMonitor *local_monitor = G_LOCAL_FILE_MONITOR (object);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_source_unref ((GSource *) local_monitor->source);
G_OBJECT_CLASS (g_local_file_monitor_parent_class)->finalize (object);
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
static void
g_local_file_monitor_init (GLocalFileMonitor* local_monitor)
{
}
static void g_local_file_monitor_class_init (GLocalFileMonitorClass *class)
{
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
GObjectClass *gobject_class = G_OBJECT_CLASS (class);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
gobject_class->get_property = g_local_file_monitor_get_property;
gobject_class->set_property = g_local_file_monitor_set_property;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
gobject_class->dispose = g_local_file_monitor_dispose;
gobject_class->finalize = g_local_file_monitor_finalize;
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_object_class_override_property (gobject_class, PROP_RATE_LIMIT, "rate-limit");
}
static GLocalFileMonitor *
g_local_file_monitor_new (gboolean is_remote_fs,
gboolean is_directory,
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
GError **error)
{
GType type = G_TYPE_INVALID;
if (is_remote_fs)
type = _g_io_module_get_default_type (G_NFS_FILE_MONITOR_EXTENSION_POINT_NAME,
"GIO_USE_FILE_MONITOR",
G_STRUCT_OFFSET (GLocalFileMonitorClass, is_supported));
/* Fallback rather to poll file monitor for remote files, see gfile.c. */
if (type == G_TYPE_INVALID && (!is_remote_fs || is_directory))
type = _g_io_module_get_default_type (G_LOCAL_FILE_MONITOR_EXTENSION_POINT_NAME,
"GIO_USE_FILE_MONITOR",
G_STRUCT_OFFSET (GLocalFileMonitorClass, is_supported));
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
if (type == G_TYPE_INVALID)
{
g_set_error_literal (error, G_IO_ERROR, G_IO_ERROR_FAILED,
_("Unable to find default local file monitor type"));
return NULL;
}
return g_object_new (type, NULL);
}
GFileMonitor *
g_local_file_monitor_new_for_path (const gchar *pathname,
gboolean is_directory,
GFileMonitorFlags flags,
GError **error)
{
GLocalFileMonitor *monitor;
gboolean is_remote_fs;
is_remote_fs = g_local_file_is_nfs_home (pathname);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
monitor = g_local_file_monitor_new (is_remote_fs, is_directory, error);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
if (monitor)
g_local_file_monitor_start (monitor, pathname, is_directory, flags, g_main_context_get_thread_default ());
return G_FILE_MONITOR (monitor);
}
GFileMonitor *
g_local_file_monitor_new_in_worker (const gchar *pathname,
gboolean is_directory,
GFileMonitorFlags flags,
GFileMonitorCallback callback,
gpointer user_data,
GClosureNotify destroy_user_data,
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
GError **error)
{
GLocalFileMonitor *monitor;
gboolean is_remote_fs;
is_remote_fs = g_local_file_is_nfs_home (pathname);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
monitor = g_local_file_monitor_new (is_remote_fs, is_directory, error);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
if (monitor)
{
if (callback)
g_signal_connect_data (monitor, "changed", G_CALLBACK (callback),
user_data, destroy_user_data, G_CONNECT_DEFAULT);
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
g_local_file_monitor_start (monitor, pathname, is_directory, flags, GLIB_PRIVATE_CALL(g_get_worker_context) ());
}
substantially rework file monitors Remove all event merging and dispatch logic from GFileMonitor. The only implementation of GFileMonitor outside of glib is in gvfs and it already does these things properly. Get rid of GLocalDirectoryMonitor. We will use a single class, GLocalFileMonitor, for both directory and file monitoring. This will prevent every single backend from having to create two objects separately (eg: ginotifydirectorymonitor.c and ginotifyfilemonitor.c). Introduce GFileMonitorSource as a thread-safe cross-context dispatch mechanism. Put it in GLocalFileMonitor. All backends will be expected to dispatch via the source and not touch the GFileMonitor object at all from the worker thread. Remove all construct properties from GLocalFileMonitor and remove the "context" construct property from GFileMonitor. All backends must now get the information about what file to monitor from the ->start() call which is mandatory to implement. Remove the implementation of rate limiting in GFileMonitor and add an implementation in GLocalFileMonitor. gvfs never did anything with this anyway, but if it wanted to, it would have to implement it for itself. This was done in order to get the rate_limit field into the GFileMonitorSource so that it could be safely accessed from the worker thread. Expose g_local_file_is_remote() internally for NFS detection. With the "is_remote" functionality exposed, we can now move all functions for creating local file monitors to a proper location in glocalfilemonitor.c Port the inotify backend to adjust to the changes above. None of the other backends are ported yet. Those will come in future commits.
2015-01-12 20:59:35 +01:00
return G_FILE_MONITOR (monitor);
}