| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
We're missing a lot of dbus_message_unref after sending
messages, add them.
|
|
|
|
|
|
|
|
|
|
|
| |
There is a race in g_vfs_job_close_read_finalize vs command_read_cb
where if these run in different threads (the command_read_cb in the main
thread and the close in another thread), such as in the smb case then
we may free the channel early.
We fix this by having the command reading propely ref the channel and
add a cancellable that allows us to cancel the command reading when the
channel is closed.
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=614554
|
|
|
|
|
|
|
| |
We were moving and renaming files, instead of just renaming them,
as the display_name wasn't properly rooted to the existing directory.
https://bugzilla.gnome.org/show_bug.cgi?id=638191
|
|
|
|
|
|
|
|
|
| |
This is actually a workaround for inacessible root of the share.
By setting default location to a user-specified path Nautilus will
go to this folder by default, leaving errors from inaccessible
root behind.
https://bugzilla.gnome.org/show_bug.cgi?id=586129
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=641856
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is the first step in making extraction faster. Some GIO-based
file managers like Nautilus sort files by inode number and making
copy/extraction faster (more linear read on disk, linear read from
archive).
For the archiving backend, let's give each file a number corresponding
with the position the file is stored in the archive. We can't use
native inode number stored in some types of archives as long as there
are no guarantees it reflects actual position in the file.
This change will have no effect until we make do_open_for_read() not
to close archive after reading and continue from the last position.
I'm afraid this concept would conflict with multiple accesses at one
time though.
|
|
|
|
| |
Happens with some archive types, such as rpm (libarchive 2.8.4).
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=631483
|
|
|
|
|
|
|
| |
This will fix an abort when archive is to be mounted for the second time:
ERROR:gvfsbackendarchive.c:240:g_vfs_backend_archive_finalize: assertion failed: (archive->file == NULL)
https://bugzilla.redhat.com/show_bug.cgi?id=645630
|
| |
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=416221
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=416221
|
|
|
|
|
|
|
| |
Support for it will come in later patches.
The patch also detects chgrp support, but we're not intending to use it
yet.
|
|
|
|
|
|
| |
To access the Documents/ folder within applications.
https://bugzilla.gnome.org/show_bug.cgi?id=636132
|
| |
|
|
|
|
|
|
| |
In error cases, we would get the refcounting wrong, which would lead to
a warning when finalizing the backend, because we carried around an
invalid pointer to an unreffed address for the whole time.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Prevent fd leakage and possible selinux alert.
|
|
|
|
|
|
|
|
|
|
|
|
| |
When switching from select() to poll() in commit
"sftp: Use poll() to cope with openssh-5.6 changes" (c6be45c8934)
the difference in seconds vs milliseconds for select/poll timeout
argument was missed.
SFTP_READ_TIMEOUT is defined in seconds, so multiply it with 1000
when using it with poll().
https://bugzilla.gnome.org/show_bug.cgi?id=631169
|
|
|
|
|
| |
Patch by Otto Allmendinger
See bug 629184 for details.
|
|
|
|
| |
Solve the deadlock described in bug 628889
|
|
|
|
|
|
|
| |
And iOS2. This also adds support for preview icon for movies
captured on the devices themselves.
https://bugzilla.gnome.org/show_bug.cgi?id=627193
|
|
|
|
|
| |
xgettext only extracts comments in po files if the comment is on the
line just before the translatable string.
|
|
|
|
|
|
|
| |
Adds host and username information when mounting a SSH share, instead of
just "Enter password".
https://bugzilla.gnome.org/show_bug.cgi?id=621583
|
|
|
|
|
|
|
|
|
|
|
| |
gvfsftpdircache.c: In function 'g_vfs_ftp_dir_cache_funcs_process':
gvfsftpdircache.c:589: error: 'unix' undeclared (first use in this function)
...caught when building on Mac OS X.
Apparently using "unix" works for some magic reason on gnu/linux and friends.
https://bugzilla.gnome.org/show_bug.cgi?id=622159
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=618019
|
|
|
|
| |
This makes gvfs build with recent glib again, see bug #619537
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=614401
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=615039
|
|
|
|
|
|
|
|
| |
The jailbroken file system access isn't supported by
libgpod, so it's useless for us to say that it's a
media-player.
https://bugzilla.gnome.org/show_bug.cgi?id=614615
|
|
|
|
|
|
|
| |
AFC_E_INVALID_ARGUMENT was renamed to AFC_E_INVALID_ARG
in http://github.com/MattColyer/libiphone/commit/36048ded8efda588a20b5cf284670a984f7cc650
https://bugzilla.gnome.org/show_bug.cgi?id=613588
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=612536
|
|
|
|
|
|
| |
NULL
https://bugzilla.gnome.org/show_bug.cgi?id=612536
|
|
|
|
|
|
|
|
|
| |
Previous code was assuming that archive_errno() would report the last
error that happened or SUCCESS if none. This is not true, however as the
return value of archive_errno() is undefined if no error happened.
This patch changes it to only call archive_errno() after errors.
https://bugzilla.gnome.org/show_bug.cgi?id=611057
|
|
|
|
|
| |
Closing the data condition works unconditionally, the close function
checks itself.
|
|
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=610259
Signed-off-by: David Zeuthen <davidz@redhat.com>
|
|
|
|
|
|
|
|
|
| |
After the delete op, cache was invalidated and an attempt was made to retrieve
file size in do_pull() while data connection was already opened. This patch
only moves all necessary operations before setting up actual data stream for
download.
Fixes bug 610395
|
|
|
|
|
|
| |
When username changes, we need to respawn the underlying process and
go through handle_login() again. We need to save the GPasswordSave value
temporarily to have the password actually saved if desired.
|
|
|
|
| |
Close the active read_reply_async() channel waiting for input from the sftp process.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As PASV and PORT are older, there's a higher chance that they are
supported. As only IPv4 connections can use these commands, IPv6 will
still prefer EPSV/EPRT.
The order of methods now looks like this (assuming EPSV and EPRT are
supported):
IPv4: PASV EPSV PORT EPRT
IPv6: EPSV EPRT PASV
Note that PASV for IPv6 will try the returned port and the known remote
address since the IP is not reliable (That's the fallback part in IPv4).
|
| |
|
| |
|
| |
|