This suppressing the warning generated when running offlineimap on
python2 with -3 switch:
offlineimap/folder/Base.py:29: DeprecationWarning: Overriding __eq__
blocks inheritance of __hash__ in 3.x
Since this object is mutable it should not be hashable.
From Python documentation[1]:
A class that overrides __eq__() and does not define __hash__() will
have its __hash__() implicitly set to None.
Therefore old behaviour is preserved.
[1] https://docs.python.org/3/reference/datamodel.html
Signed-off-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
OfflineIMAP frequently delivers mail files to the Maildir consisting exclusively
of a single ASCII digit in IDLE mode. IMAPFolder.getmessage expects 'data' to be
of the form
[(fetch-info, message-body)]
However, the imapobj.uid call in getmessage returns a list of *all* pending
untagged FETCH responses. If any message flags were changed in the selected
IMAP folder since the last command (by another client or another thread in
OfflineIMAP itself), the IMAP server will issue unsolicited FETCH responses
indicating these flag changes (RFC3501, section 7). When this happens, 'data'
will look like, for example
['1231 (FLAGS (\\Seen) UID 5300)',
'1238 (FLAGS (\\Seen) UID 5318)',
('1242 (UID 5325 BODY[] {7976}', message-body)]
Unfortunately, getmessage retrieves the message body as data[0][1], which in
this example is just the string "2", and this is what gets stored in the mail
file.
Multi-threaded OfflineIMAP with IDLE or holdconnectionopen is particularly
susceptible to this problem because flag changes synced back to the IMAP server
on one thread will appear as unsolicited FETCH responses on another thread if it
happens to have the same folder selected. This can also happen without IDLE or
holdconnectionopen or even in single-threaded OfflineIMAP with concurrent access
from other IMAP clients (webmail clients, etc.), though the window for the bug
is much smaller.
Ideally, either imaplib2 or getmessage would parse the fetch responses to find
the response for the requested UID. However, since IMAP only specifies
unilateral FETCH responses for flag changes, it's almost certainly safe to
simply find the element of 'data' that is a tuple (perhaps aborting if there is
more than one tuple) and use that.
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/162
Based-on-patch-by: Austin Clements <amdragon@MIT.EDU>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
1. There is one database per folder and sqlite requires to serialize the
writings. Instead of locking at LocalStatusSQLiteFolder object level, introduce
a new DatabaseFileLock object which is shared across threads. This fixes the
concurrent writes issues that some users might experience by duplications or
flags restored to the previous state.
2. Close the database only when we are sure no other threads will use the
connection on a *per-file* basis. Previous fix 677afb8d8f is wrong
because the same lock is shared for all the databases.
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/350
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
1. There is one database per folder and sqlite requires to serialize the
writings. Instead of locking at LocalStatusSQLiteFolder instance level,
introduce a new DatabaseFileLock object which is shared across threads. This
fixes the concurrent writes issues that some users might experience by
duplications or flags restored to the previous state.
2. Close the database only when we are sure no other threads will use the
connection on a *per-file* basis. Previous fix 677afb8d8f is wrong
because the same lock is shared for all the database files.
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/350
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
This did not work and is not usefull since the purge() method was introduced in
1410a391bc. Actually, the purge() does what deletemessagelist() was supposed to
achieve.
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
It's required to have more than one connection to the database for the
maxconnections configuration option to work with threads. However,
connection.close() is closing all the connections. Only close the connection
when no more thread need it.
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/350
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
It's required to have more than one connection to the database for the
maxconnections configuration option to work with threads. However,
connection.close() is closing all the connections. Only close the connection
when no more thread need it.
Backported-from: 856b74407bd7f634cae5a8c2d9b84e13d14c12d2
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/350
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
Output a warning so that we can dig into this. For now, the patch doesn't fix
the root cause. If the server returns UID 0 as valid UID number, this must be
ignored as soon as possible.
Github-ref: https://github.com/OfflineIMAP/offlineimap/issues/336
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
When a maildir is removed it must be considered new for the sync. However, the
local cache of the folder remains. This means the sync of the folder removes all
the missing emails.
Avoid loosing of data for users not aware of the local cache by removing any
pre-existing status cache of a folder when we actually want to create the
database.
Improve style.
Github-fix: https://github.com/OfflineIMAP/offlineimap/issues/333
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
Backported from 8e995a69bfa003ab822b55731429d84b3bc5626f.
We currently close the database as soon as possible while we handle the status
backend but this is still too early because autorefresh induces looping on this
code block.
Instead of delaying the closing outside of the loop, it's easier to delay the
opening as late as possible (inside the loop). The downside is that the database
is opened/closed more than once when autorefresh is enabled. The good news is
that this make the code much easier.
Fixes regression introduces by 6fb5700.
Reported-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
The virtual imaplib2 first try to import imaplib2 when provided by the system.
If not provided or if the version is not supported, fallback on the bundled
imaplib2 version.
Distribution maintainers can now easily remove the bundled imaplib2 version if
they want to get it packaged outside of offlineimap.
We still want to provide imaplib2 by default because:
- this library is neither in Python core nor packaged by a lot of distributions;
- users expect to be able to run offlineimap by just downloading the tarball or
after a git clone.
In order to avoid unexpected (too old) versions of imaplib2, we restrict the
supported versions of this librabry.
Reviewed-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
Some servers are broken: they return zero as valid UID and respond "BAD invalid
parameter: 0" on the FETCH command.
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
There is no other way to make Python2 and Python3 happy, because syntax
raise E, V, T is incompatible with the latter.
Signed-off-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
Python3 accepts binary input for md5 function.
This patch is known to break setups using folder names not strictly conforming
the IMAP UTF-7 encoding. We always made it clear that such setup is unsupported
and might be broken at some point in time. See documentation about
'decodefoldernames' in the provided configuration file. This is why this patch
is considered introducing no regression for this use case.
Patches to support both Python 3 and Python 2 by re-encoding the MD5 in the
filenames are welcome. This likely requires a new CLI option to allow
backporting the feature for users downgrading or changing of Python environment.
Signed-off-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
We currently close the database as soon as possible while we handle the status
backend but this is still too early because autorefresh induces looping on this
code block.
Instead of delaying the closing outside of the loop, it's easier to delay the
opening as late as possible (inside the loop). The downside is that the database
is opened/closed more than once when autorefresh is enabled. The good news is
that this make the code much easier.
Fixes regression introduces by 6fb5700.
Reported-by: Łukasz Żarnowiecki <dolohow@outlook.com>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
