Hi Timo (and list :)
No bug reports since rc4, so here it is.
[snip]
One small question: How far is nfs safety implemented now? You did a lot of work on nfs-safety for the index files, how about the other files? And does it need specific configure options?
Thank you for this release though, I will have a look at it tonight ;-)
Kind regards,
Maikel Verheijen.
On Mon, 2003-06-30 at 17:03, Maikel Verheijen wrote:
One small question: How far is nfs safety implemented now? You did a lot of work on nfs-safety for the index files, how about the other files? And does it need specific configure options?
No, I've done nothing for NFS-safe index files and likely will not do for a long time. NFS-safety is still pretty much the same as it was two months ago. From TODO:
- NFS safety:
- use link()s instead of relying on O_EXCL
- .subscriptions: use rename() like dovecot-uidlist
- .customflags: use rename(), but there's a problem when we have to remove unused flags to make room for new ones. to fix that add new field in the file, it would be set for flags which are currently unused. if that field is set, .customflags must be locked before the flag is set to any messages. but make sure there's no race conditions, we probably have to wait a few seconds just to make sure no-one set a flag we want to remove
On 30 Jun 2003, Timo Sirainen wrote:
On Mon, 2003-06-30 at 17:03, Maikel Verheijen wrote:
One small question: How far is nfs safety implemented now? You did a lot of work on nfs-safety for the index files, how about the other files? And does it need specific configure options?
No, I've done nothing for NFS-safe index files and likely will not do for a long time. NFS-safety is still pretty much the same as it was two months ago. From TODO:
- NFS safety:
- use link()s instead of relying on O_EXCL [etc]
It seems from a brief test that messages aren't sorted when new UIDs are assigned to existing messages in ./cur/. As a result, UIDs may be nothing like in message arrival order.
-- Charlie
On Tue, 2003-07-01 at 04:05, Charlie Brady wrote:
It seems from a brief test that messages aren't sorted when new UIDs are assigned to existing messages in ./cur/. As a result, UIDs may be nothing like in message arrival order.
I thought I fixed it just before 0.99.10 release. And I just tested with 4000 messages, they were ordered just fine.
On 1 Jul 2003, Timo Sirainen wrote:
On Tue, 2003-07-01 at 04:05, Charlie Brady wrote:
It seems from a brief test that messages aren't sorted when new UIDs are assigned to existing messages in ./cur/. As a result, UIDs may be nothing like in message arrival order.
I thought I fixed it just before 0.99.10 release.
Just before? Do you mean "after 0.99.10-rc3" (which is what I tested)? I don't recognise the fix in the ChangeLog.
And I just tested with 4000 messages, they were ordered just fine.
Here's some of my dovecot-uidlist:
1 1056728482 7490 1 1029035495.23819.allspice,U=3055,W=3232 2 1048787491.30439.allspice,U=14979,W=1922 3 1030760219.16598.allspice,U=4022,W=3563 4 1043671263.23068.allspice,U=11963,W=1892 5 1030739827.15369.allspice,U=4011,W=1981 6 1035300042.15815.allspice,U=7062,W=2257 7 1020252495.10690.allspice,U=188,W=3681 8 1040245470.4889.allspice,U=10279,W=2281 9 1043276959.18707.allspice,U=11811,W=2308 10 1046307299.12933.allspice,U=13793,W=4375 ...
-- Charlie Brady charlie_brady@mitel.com Lead Product Developer Network Server Solutions Group Mitel Networks Corporation http://www.mitel.com/smallbusiness Phone: +1 (613) 592 5660 or 592 2122 Fax: +1 (613) 592 1175
On Wed, 2003-07-02 at 18:40, Charlie Brady wrote:
It seems from a brief test that messages aren't sorted when new UIDs are assigned to existing messages in ./cur/. As a result, UIDs may be nothing like in message arrival order.
I thought I fixed it just before 0.99.10 release.
Just before? Do you mean "after 0.99.10-rc3" (which is what I tested)? I don't recognise the fix in the ChangeLog.
Yes, 11 minutes before release :)
2003-06-26 16:01 Timo Sirainen tss@iki.fi
* src/lib-index/maildir/maildir-sync.c: When assigning UIDs to
mails, we tried to sort them with strcmp(), but that wasn't
actually working. Also we now sort properly the mails that had unix
timestamp older than 1 billion (Sep 9 2001).
participants (3)
-
Charlie Brady
-
Maikel Verheijen
-
Timo Sirainen