among other things, this contains a possible fix for
https://sourceforge.net/p/isync/bugs/22/ and a lot of related reports.
patch by Florian Lombard <f.lombard@montmirail.com>:
Common cfg section:
* Either skip or fix messages with lines more than xxx bytes
(typically no more than 9900 bytes with exchange)
MaxLineLength xxx (in bytes)
CutLongLines yes|no (fix or skip message)
* Allow to rescan all mails from a folder, ignoring the last sync
latest message pulled (usefull when playing with my new settings)
IgnoreMaxPulledUid yes|no
* Skip messages with raw binary content (bytes < 0x20 except CR/LF/TAB)
SkipBinaryContent yes|no
* Allow to delete non empty folders on slave (when you are sure about
what you're doing)
DeleteNonEmpty yes|no
Drivers cfg section (imap only):
* Suppress Keyword not supported warnings
IgnoreKeywordWarnings yes|no
The only missing part is long lines cutting when there's CR/LF
convertion (I don't use maildir++)
============
my response:
> Common cfg section:
>
> * Either skip or fix messages with lines more than xxx bytes
> (typically no more than 9900 bytes with exchange)
> MaxLineLength xxx (in bytes)
> CutLongLines yes|no (fix or skip message)
>
as mentioned before, i'm concerned about the "sledge hammer" approach of
hard-cutting the lines, because that falsifies the messages' content,
which may very well render them unreadable (if it's not plain text).
meanwhile i found that this should at least not invalidate possibly
present signatures, simply because the respective standards require
complete normalization of the contents before signing - specifically to
avoid the problem.
still, a cleaner approach would be encapsulating the message in a MIME
structure. i found in the imapsync FAQ that "reformime -r7" would do
that (i'm not suggesting to use that, but it should serve as a good
example).
i'd be interested in samples of such messages with excessively long
lines to assess what the "target audience" actually is. i would expect
that messages which already are MIME-encoded would not have this
problem. but then, a sloppily encoded multipart text+html mail could
very well be broken as well.
> * Allow to rescan all mails from a folder, ignoring the last sync
> latest message pulled (usefull when playing with my new settings)
> IgnoreMaxPulledUid yes|no
>
that seems to be overkill to me given that it's a workaround and can be
easily achieved by hacking the sync state files, for example by sed'ing
them.
i suppose you implemented this to resume syncing after implementing the
line length workaround?
> * Skip messages with raw binary content (bytes < 0x20 except CR/LF/TAB)
> SkipBinaryContent yes|no
>
i know that i suggested that this might be a problem, but i don't
remember whether you reported actual instances of that.
anyway, the treatment should be the same as for messages with excesively
long lines - MIME-encoding (presumably as quoted-printable).
> * Allow to delete non empty folders on slave (when you are sure about
> what you're doing)
> DeleteNonEmpty yes|no
>
i'll consider this.
my biggest concern is that some transient error would falsify the
mailbox list and thus cause the folders to be nuked. similary, a
permanent change in the server configuration would have that effect.
arguably, either wouldn't be so bad as such, as it would destroy only
the replica. however, it would be important to verify that the replica
does not contain any unpropagated mails (as opposed to any mails at all,
as is done currently).
> Drivers cfg section (imap only):
>
> * Suppress Keyword not supported warnings
> IgnoreKeywordWarnings yes|no
>
i wonder why a server would bleat about not supporting an optional
feature when it can (and probably does) announce that in a "civilized"
way, too. did these responses appear to be correlated with specific
messages, or did they always come when opening any mailbox?
> diff --git a/src/drv_imap.c b/src/drv_imap.c
> index e24c7d8..10da0cb 100644
> --- a/src/drv_imap.c
> +++ b/src/drv_imap.c
> @@ -1416,6 +1419,16 @@ imap_socket_read( void *aux )
> resp = RESP_NO;
> if (cmdp->param.failok)
> goto doresp;
> + } else if (!strcmp( "BAD", arg )) {
> + resp = RESP_NO;
> + warn( "Warning: IMAP command '%s' returned an error: %s %s\n",
> + starts_with( cmdp->cmd, -1, "LOGIN", 5 ) ?
> + "LOGIN <user> <pass>" :
> + starts_with( cmdp->cmd, -1, "AUTHENTICATE PLAIN", 18 ) ?
> + "AUTHENTICATE PLAIN <authdata>" :
> + cmdp->cmd,
> + arg, cmd ? cmd : "" );
> + goto doresp;
> } else /*if (!strcmp( "BAD", arg ))*/
> resp = RESP_CANCEL;
>
this hunk downgrades tagged BAD responses to warnings and suppresses the
subsequent client-side connection drop.
this doesn't seem like a terribly good idea to me - this server response
indicates that the client (allegedly) did something wrong. that may mean
that the subsequent command stream will be interpreted as garbage, which
may have unpredictable effects. it just isn't safe to continue at this
point.
i suppose you implemented this as a workaround before you identified the
line length issue?
============
and a last retour:
>> Common cfg section:
>>
>> * Either skip or fix messages with lines more than xxx bytes
>> (typically no more than 9900 bytes with exchange)
>> MaxLineLength xxx (in bytes)
>> CutLongLines yes|no (fix or skip message)
> as mentioned before, i'm concerned about the "sledge hammer" approach of
> hard-cutting the lines, because that falsifies the messages' content,
> which may very well render them unreadable (if it's not plain text).
Well you have the choice of just skipping them to allow the sync to
complete if you're concerned about the messages integrity
> meanwhile i found that this should at least not invalidate possibly
> present signatures, simply because the respective standards require
> complete normalization of the contents before signing - specifically to
> avoid the problem.
>
> still, a cleaner approach would be encapsulating the message in a MIME
> structure. i found in the imapsync FAQ that "reformime -r7" would do
> that (i'm not suggesting to use that, but it should serve as a good
> example).
I had a look at that, and found that completely overkill for my usage
(see below)
> i'd be interested in samples of such messages with excessively long
> lines to assess what the "target audience" actually is. i would expect
> that messages which already are MIME-encoded would not have this
> problem. but then, a sloppily encoded multipart text+html mail could
> very well be broken as well.
100% of those messages where having bad html code without line breaks
Non binary attachments where always correctly line wrapped.
It was either poorly done html signatures or even javascript (yeah,
inside an email !)
So I wasn't worried about the integrity of those messages, which where
already breaking the rules, but I needed the contents (messages from
customers we needed to keep)
>> * Allow to rescan all mails from a folder, ignoring the last sync
>> latest message pulled (usefull when playing with my new settings)
>> IgnoreMaxPulledUid yes|no
> that seems to be overkill to me given that it's a workaround and can be
> easily achieved by hacking the sync state files, for example by sed'ing
> them.
> i suppose you implemented this to resume syncing after implementing the
> line length workaround?
Yes it was mainly a flag I used for debugging (editing hundreds of sync
state files wasn't an option)
>> * Skip messages with raw binary content (bytes < 0x20 except CR/LF/TAB)
>> SkipBinaryContent yes|no
> i know that i suggested that this might be a problem, but i don't
> remember whether you reported actual instances of that.
> anyway, the treatment should be the same as for messages with excesively
> long lines - MIME-encoding (presumably as quoted-printable).
Those where bogus messages with the raw attachment in binary but with
base 64 headers correctly set.
Near 100% (if not 100%) of those where in the sent folder and are
probably the result of gmail + buggy email client (but you can still
open the attachment with gmail !)
>> * Allow to delete non empty folders on slave (when you are sure about
>> what you're doing)
>> DeleteNonEmpty yes|no
> i'll consider this.
> my biggest concern is that some transient error would falsify the
> mailbox list and thus cause the folders to be nuked. similary, a
> permanent change in the server configuration would have that effect.
> arguably, either wouldn't be so bad as such, as it would destroy only
> the replica. however, it would be important to verify that the replica
> does not contain any unpropagated mails (as opposed to any mails at all,
> as is done currently).
Well, when you are sure about your settings, this can be usefull, as my
users where renaming folders while I was working on the sync
At start I was logging to the mailbox, deleted the folder, and syncing
again.
>> Drivers cfg section (imap only):
>>
>> * Suppress Keyword not supported warnings
>> IgnoreKeywordWarnings yes|no
>>
> i wonder why a server would bleat about not supporting an optional
> feature when it can (and probably does) announce that in a "civilized"
> way, too. did these responses appear to be correlated with specific
> messages, or did they always come when opening any mailbox?
Well, "exchange online", that sums it all ...
Tied to specific messages, I guess it happened when there was a word
between bracket in the message subject (no debug log of that)
Happends only one time, when the message is synced.
A rather ugly hack, but I needed clean logs to spot errors.
> i suppose you implemented this as a workaround before you identified the
> line length issue?
I implemented that before the binary content issue
It's exchange which is breaking all the rules that "forced" me to do
that to sync most of the messages
Cutting the connexion instead of reporting the right error is not the
right thing to do, but that's what exchange does (with Error 10 or 11,
but with BAD reponse)
when syncing flags but not re-newing non-fetched messages, there is no
need to query the message size for all messages, as the old ones are
queried only for their flags.
instead of a single hard-coded branch, use a generic method to split
ranges as needed.
this is of course entirely over-engineered as of now, but subsequent
commits will make good use of it.
turns out the comment advising against it was bogus - unlike for
memcmp(), the standard does indeed prescribe that the memchr()
implementation may not read past the first occurrence of the searched
char.
if AuthMechs includes more than just LOGIN and the server announces any
AUTH= mechanism, we try SASL. but that can still fail to find any
suitable authentication mechanism, and we must not error out in that
case if we are supposed to fall back to LOGIN.
specifically, if AuthMechs included more than just LOGIN (which would be
the case for '*') and the server announced any AUTH= mechanism, we'd
immediately error out upon seeing it, thus failing to actually try
LOGIN.
the number was chosen to make queries more comprehensible when the
server sends no UIDNEXT, but it appears that such insanely large UIDs
actually show up in the wild. so send 32-bit INT_MAX instead.
note that this is again making an assumption: that no server uses
unsigned ints for UIDs. but we can't sent UINT_MAX, as that would break
with servers which use signed ints. also, *we* use signed ints (which is
actually a clear violation of the spec).
it would be possible to special-case the range [1,inf] to 1:*, thus
entirely removing arbitrary limits. however, when the range doesn't
start at 1, we may actually get a single message instead of none due to
the imap uid range limits being unordered. this gets really nasty when
we need to issue multiple queries, as we may list the same message
twice.
a reliable way around this would be issuing a separate query to find the
actual value of UID '*', to make up for the server not sending UIDNEXT
in the first place. this would obviously imply an additional round-trip
per mailbox ...
trashing many messages at once inevitably overtaxes m$ exchange, and the
connection breaks. without any progress tracking, it would restart from
scratch each time, which would lead to a) it never finishing and b) many
copies of the messages in the trash.
full transactions as we do for "proper" syncing would be over the top,
as it's not *that* bad if some messages get duplicated in the trash. so
we record only the messages for which trashing completed, thus allowing
some overlap between the attempts.
turns out i misread the spec in a subtle way: while all other folders
are physically nested under INBOX, the IMAP view puts them at the same
(root) level. to get them shown as subfolders of INBOX, they need to
have _two_ leading dots.
this also implies that the Maildir++ mode has no use for a Path, so
reject attempts to specify one.
the mbsync manual says explicitly that the system's default certificate
store should *not* be specified.
however, the isync manual talked about CA certificates, which is (and
always was) exactly wrong.
also adjust both .sample rc files.
flock() may be implemented via fcntl(), which may cause the process to
deadlock itself when trying to apply both types of locks. this is the
case even on linux when the file lives on NFS.
it's unlikely that anything except mbsync would try to access the
.uidvalidity files anyway, so there is no point in trying to be
compatible with anything else ...
REFMAIL: uddy4g589ym.fsf@eismej-u14.spgear.lab.emc.com
it is legal for an email system to simply change the case of rfc2822
headers, and at least one imap server apparently does just that.
this would lead to us not finding our own header, which is obviously not
helpful.
REFMAIL: CA+fD2U3hJEszmvwBsXEpTsaWgJ2Dh373mCESM3M0kg3ZwAYjaw@mail.gmail.com
Oswald Buddenhagen
Anton Khirnov