running sync_server from master, and all our
mailstores have a single data partition) and hasn't been tested anywhere
else. Ken rewrote large chunks of the code to turn it into proper Cyrus
service, and Bron threw it all away to make something better.
--
David Carter
On Fri, 29 Jun 2012, Bron Gondwana wrote:
My real plan is to make the Cyrus code able to read old-version Cyrus
files directly without upgrading them...
I would certainly welcome this.
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service
entry, I'm just
debating what to put in that report. I am running Cyrus 2.4.13.
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK
://bugzilla.cyrusimap.org/show_bug.cgi?id=2772
which has been open since December 2005.
Should I open a new bug in bugzilla? Stripping the spurious whitespace
from the middle of msgids in the References header (as John Capo suggests
in Bugzilla #2772) feels like the most elegant solution.
--
David
don't have any
conflict with the mailbox namespace. For example a mailbox named 42 or
cyrus.index. I really wouldn't recommend this.
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service,Phone: (01223) 334502
New Museums Site
other functions in imapd.c: catenate_url() and cmd_urlfetch().
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.diff -udNr
The quota parameter to verify_user() is quota_t, not long.
Solaris 10 with bundled GCC 3.4.3 wasn't doing type promotion correctly.
(long)(-1) becomes (long long int)(4294967295), which is = 0.
Consequently IGNOREQUOTA had no effect.
--
David Carter Email
.
The exclusive POP lock on cyrus.cache is an additional complication, but
that is off to one side and seems to work reasonably well. (I'm not
entirely convinced that a server which implements delayed expunge actually
needs the exclusive lock, but it is mandated by RFC 1725, section 4).
--
David
happy about
mboxlist_findall which can run for several days in any case.
sync_combine_commit.patch
=
Update quota correctly in sync_combine_commit(). Only normally triggered
when replicating unexpunged messages in normal use.
--
David Carter Email
On Fri, 27 Mar 2009, Bron Gondwana wrote:
Why did this touch sieve.h? It's generated by bison. Did you not
work on a clean tree?
2.3.14 ships with a sieve/sieve.h I did remove a number of the .c files
which are generated from Bison or Flex, but missed this one.
--
David Carter
)
But if you are happier with two or three lines of code, then go for it.
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
amount of processor for
something that really was /[0-9]/ in meaning.
So - where would you put the macro?
lib/util.h? That already has:
#define TOUPPER(c) (convert_to_uppercase[(unsigned char)(c)])
#define TOLOWER(c) (convert_to_lowercase[(unsigned char)(c)])
--
David Carter
the fact both are just using message_parse_file().
I infer that isspace(-10) is normally false but occasionally true, at
least on Solaris x86. It appears to always be false on Linux x86.
--
David Carter Email: david.car...@ucs.cam.ac.uk
University Computing Service
may not want to add these to
GOODCHARS.
I had to remove '!' from the list above when we moved from 2.1.x.
and ? might conflict with modified UTF-7 encoding.
shouldn't be allowed without checking for a valid modified UTF-7
sequence. What is the problem with '?'.
--
David Carter
mboxlist_findall() quietly starts a transaction.
We use this patch as well. The only curiosity that I haven't tracked down
is the following from cyr_expire:
Apr 9 03:43:52 cyrus-35 cyr_expire[15410]: skiplist: unlock while not locked
(I haven't touched cyr_expire in 2.3.11).
--
David Carter
There is a small memory leak in the status cache code. Or at least I don't
believe that DB-store() automatically free()s its input.
Probably faster to just have a char data[250] on the stack.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service
...
I run quota on master and replica every once in a while and compare.
That's an easy way to find and fix the problem cases, regardless of
whether the master or replica was at fault. There are typically only about
half a dozen cases each time, so I don't think that it's a huge problem.
--
David
expunged in the last 24 hours for that mailbox. That's
the way that my old two phase expunge stuff used to work.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223
:
cyrusdb_skiplist.c: 913: 2 + 2 == 5
The error text is also sent to syslog() before a proper shut_down().
Here is a simple patch to change all of the '#include assert.h' into
'#include assert.h':
http://www-uxsup.csx.cam.ac.uk/~dpc22/cyrus/patches/2.3cvs/assert.patch
--
David Carter
:
. SCAN test/ 1 dpc22
which could just as easily be done using EXAMINE and SEARCH.
Things would probably go much faster searching one mailbox at a time
rather that adding random access disk I/O just for the hell of it.
--
David Carter Email: [EMAIL PROTECTED]
University
of the original code.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
condstore.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
packages to try how they work. Do you
recommend both for general consumption?
It is certainly very easy to break mailboxes.db using cyr_dbtool.
Kudos to Bron for tracking down the problems.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service
On Tue, 6 Nov 2007, Rudy Gevaert wrote:
Added -u and -U options to reconstruct -- courtesy of David Carter.
I don't find any information in the man of reconstruct or in reconstruct.c :)
UUIDs are now GUIDs, hence -g/-G
--
David Carter Email: [EMAIL PROTECTED
is possible.
This requirement is likely to be a pain for:
1. Leverage delayed expunge which already stores state for expunged
messages in cyrus.expunge (up until the records are purged by
cyr_expire).
Given that unexpunge assigns a new UIDvalidity to the mailbox.
--
David Carter
On Tue, 9 Oct 2007, David Carter wrote:
I've never faced a spilt brain situation which involved more than two or
three messages (the outstanding log on an old master system).
I suppose that it was predicable that a week after writing this I faced my
first serious split brain (3000 messages
.
And indeed it uses an explicit SA_RESTART flag, so it looks like I've
been worrying about nothing all along. By all means signal away!
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke
and SVR4: BSD tries to
retry while SVR4 throws EINTR. Linux of course can work either way:
http://www.gnu.org/software/libc/manual/html_node/Interrupted-Primitives.html
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
arguments presumably?
It would be safe to convert -u and -m command line options based on
unixhiersep and I agree that this would make sense.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke
-uxsup.csx.cam.ac.uk/~dpc22/cyrus/patches/2.3.8/sha1_make_sha1.patch
which should probably be merged if we are working towards SHA1.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street
.
make_md5/make_sha1 generate two checksums for each message: one for the
message body and also one for the cyrus.cache entry. The latter has been
occasionally useful in showing up problems with other parts of Cyrus (for
example the ENVELOPE lcase() bug from a few weeks back).
--
David Carter
be useful, although probably more work.
The kind of random sampling which Fastmail do probably wouldn't hurt as an
extra sanity check.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street
systems to a tape spooling array each night. Replication is
transaction safe rsync for Cyrus, tailored around the cyrus.index files.
More later.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site
.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
. Based on the errors, I
think that you might be missing flex, which is bison's compatriot in arms.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
if -fstrict-aliasing or -O2 is set.
The GCC man page explains -fstrict-aliasing. I also found the following:
http://en.wikipedia.org/wiki/Type_punning
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums
On Sun, 30 Sep 2007, Rudy Gevaert wrote:
If people are going to upgrade from a configuration that uses
sync_machineid= X and provide_uuid=1, and don't configure guid_mode=sha,
what will happen?
They will get NULL GUIDs.
--
David Carter Email: [EMAIL PROTECTED
, it would have asked for that
copy in that mailbox, so the ids array would have been longer.
ids[count] is always allocated.
There is a one in hundred chance that ids[count+1] is not, which would
lead to a segmentation fault if ids[count] matched a GUID.
--
David Carter
. sync_server needs to know about the expunge index. It did in my
original code base, but replication got merged before delayed expunge.
More later. In theory I'm on holiday this week.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service
appears to be braindead.
You need an update_uidlast() call if the last batch is empty, as Bron
suggests. I'm also not convinced that:
} while (!r);
ever exits. I think that it should be:
} while (!r msgno = mailbox-exists);
Updated patch against current 2.3cvs attached.
--
David
On Tue, 25 Sep 2007, David Carter wrote:
I imagine that a single GUID appearing in cyrus.index and cyrus.expunge
is bad news. sync_server needs to know about the expunge index. It did
in my original code base, but replication got merged before delayed
expunge.
Single _UID_ appearing
this is probably the simplest solution.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
collecting (rewrite
on expunge and/or cyr_expire) when the file reaches some threshold.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
don't match... I really should test-case this!)
I don't think so. There is a separate test for UIDLAST if
do_mailbox_work() finds that it has no messages to UPLOAD.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
) {
sync_index_list_free(index_list);
return(r);
}
should just be:
if (index_list-count == 0) {
return(r);
}
As I get an exciting:
*** glibc detected ***
sync_client: double free or corruption (fasttop): 0x08134fd0 ***
Otherwise looks good. Nice work.
--
David Carter
be a NOOP in that case. Could
always add the (index_list-count == 0) test to upload_messages_list()
and upload_messages_from() as well.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke
values to
users, but the wrong case isn't the end of the world.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
-uxsup.csx.cam.ac.uk/~dpc22/cyrus/patches/2.3cvs/sync_do_mailbox_work.patch
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
reconstruct on a mailbox
which is missing cyrus.header.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
, not CMUs. I'll prepare a proper patch next week.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
On Sat, 1 Sep 2007, Ken Murchison wrote:
FYI, the code in CVS has sync_message_sync() declared as having no
return value. Does your code has it declared as returning an int, and
checking for it?
Yes.
--
David Carter Email: [EMAIL PROTECTED]
University Computing
seek using mailbox-exists is definitely more robust, although
it probably doesn't help if power fails halfway through the fsync() on the
cyrus.index file after both updates have been made (data=journal maybe?)
--
David Carter Email: [EMAIL PROTECTED]
University
wasn't expecting the replication code to ever get merged
when I first wrote it. All of the really awful stuff should be in
sync_commit.c. Those are the routines which create messages with
predetermined UIDs and mailboxes with predetermined UniqueIDs.
--
David Carter
embarrassment about broken code in CVS, so long as
we are able to generate a separate branch for emergency bug fix releases.
At the end of the day CMU's CVS repository is the canonical reference.
Trying to track third party patches (and patches to patches) is hard work.
--
David Carter
On Wed, 29 Aug 2007, David Carter wrote:
mboxname_isusermailbox() works on internal mailbox names. I don't think that
it needs to worry about IMAPOPT_UNIXHIERARCHYSEP.
Here's a trivial patch to remove the IMAPOPT_UNIXHIERARCHYSEP stuff.
Otherwise mboxname_isusermailbox() is broken when
.
Small bug fix (the p++ line) attached.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
Index: imap/cyr_expire.c
.
mboxname_isusermailbox() works on internal mailbox names. I don't think
that it needs to worry about IMAPOPT_UNIXHIERARCHYSEP.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street
lrswipkxtea
279 87 1
suggests that you have a missing eatline() somewhere.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223) 334679
Cambridge UK. CB2 3QH.
postponed messages mailboxes from our
Webmail application and PINE, which I imagine is rather unusual.
Deleted mailboxes are removed after 7 days.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site
off for a couple of days, but may have a go at
implementing this next week unless someone has a better idea.
--
David Carter Email: [EMAIL PROTECTED]
University Computing Service,Phone: (01223) 334502
New Museums Site, Pembroke Street, Fax: (01223
60 matches
Mail list logo
