For most of plain text messages we have two FETCH commands now:
C: FH12 UID FETCH 4790 (UID RFC822.SIZE FLAGS INTERNALDATE BODYSTRUCTURE
BODY.PEEK[HEADER.FIELDS (DATE FROM TO SUBJECT REPLY-TO IN-REPLY-TO CC
BCC CONTENT-TRANSFER-ENCODING CONTENT-TYPE MESSAGE-ID REFERENCES
DISPOSITION-NOTIFICATION-TO X-PRIORITY)])
[08-Jan-2009 15:21:29 +0100]: C: ftch FETCH 640 (BODY.PEEK[1])
We can prefetch a small part of message body (first part) using
BODY.PEEK[1]<0.4096> in first FETCH. Then we should check if message is
realy a plain text and if body size (from BODYSTRUCTURE) is not bigger
than 4096 bytes. If so we can skip the second FETCH.
--
Aleksander 'A.L.E.C' Machniak http://alec.pl gg:2275252
LAN Management System Developer http://lms.org.pl
Roundcube Webmail Project Developer http://roundcube.net
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hi! I have another great patch from my performance fixes set ;) I've
observed that when message is displayed (with disabled cache) we have:
1. FETCH for headers
2. FETCH for message (body) structure
3. one FETCH for each message/attachment part (it really sucks when
message contains several attachments)
4. FETCH for body
So, here's the patch which joins 1. and 2. into one command call and
joins all fetches from 3. into one command call. Speed up is significant
for all messages, specially for those having many attachments. Imagine a
message with ten attachments. Previously there was ca.10 + 3 fetches *
0.2 sec. (in my mailbox with ca.800 messages) = 2.6 sec. Now, only 3
fetches * 0.2 = 0.6 sec.
I'm posting this here, because it's too late to commit such changes (not
tested well) before 0.2-stable release. So, enjoy and test in your
environments.
ps. Thomas, maybe you can wait a few days and review this patch to
include in 0.2-stable? It would be nice.
--
Aleksander 'A.L.E.C' Machniak http://alec.pl gg:2275252
LAN Management System Developer http://lms.org.pl
Roundcube Webmail Developer http://roundcube.net
--- 8< --- detachments --- 8< ---
The following attachments have been detached and are available for viewing.
http://detached.gigo.com/rc/Wr/HvaAQAVp/perf.patch
Only click these links if you trust the sender, as well as this message.
--- 8< --- detachments --- 8< ---
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hello everybody !
So ! 0.2 stable is here, I see 0.2.1 is planned for bug fixes, so 0.3 is
almost here, and in 0.3 plugin api is planned so.. when we will able to
begin to code a plugin (maybe with a different branch ? )? :]
Regards,
--
Maximilien Cuony [The_Glu]
http://theglu.org
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hi,
Does any one know what was used to minimise the JS files in the 0.2stable
tarball? Also why the tarball is different from the 0.2stable branch/tag? I
generated some patches against 0.2stable from the SVN and then they don’t
work with the other version :/.
Thanks
Phil
_______________________________________________
List info: http://lists.roundcube.net/dev/
Thunderbird has a very cool feature called "Edit as new". What it does
is simple. It takes the message "as is" without quoting it like replying
or forwarding, so I can reuse my messages as templates and don't have to
painfully copy/paste the message every time I want to duplicate it. For
example: one of my customers has to send a password mail "by hand" to
his newly registered customers and uses an old send message as template
(changing the private data) using the "Edit as new" feature in
Thunderbird. When he is not at his computer he wants to send this
message using the webmailer (my RoundCube installation of course).
I already implemented this in my installation with minimal effort using
the forward command as template. I don't have a patch for the newest svn
trunk but can provide it if this is going to be implemented. I can
provide the patch and translation files for German. What I need is a
picture for the button that appears next to the forward button maybe on
of you has an idea?
Could you please add your thoughts about this?
Regards
Christian
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hi Devs,
I was hoping someone would have the time to look at
http://trac.roundcube.net/ticket/1485634
It's a bug patch I posted a few weeks ago which details how turning on
skip deleted and caching at the same time can cause serious
misbehavior. It's been a real pain for us in our environment. I think
it's serious enough that, until its fixed, it at least warrants
something like a warning in the config file that those options are
incompatible.
Thanks.
Ziba
Webmaster Team
University of Michigan
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hi,
Trying to fix #1485659, which hurts me too, I noticed a number of
raise_error() calls missing required parameters.
Attached patch fixes three of those. Not sure though if type in
utf8.class.php and spell_pspell.inc should be 'php', so I'm not
committing this yet.
There are two other occurances in rcube_ldap.php, but I have no clue
what error code to put there.
Robin
--- 8< --- detachments --- 8< ---
The following attachments have been detached and are available for viewing.
http://detached.gigo.com/rc/Vp/NN6EXh1a/raise_error.patch
Only click these links if you trust the sender, as well as this message.
--- 8< --- detachments --- 8< ---
_______________________________________________
List info: http://lists.roundcube.net/dev/
Hi,
Here is the updated Korean Localization files.
Hanjo Kim
--- 8< --- detachments --- 8< ---
The following attachments have been detached and are available for viewing.
http://detached.gigo.com/rc/Kw/WvhG8XA3/ko_KR.zip
Only click these links if you trust the sender, as well as this message.
--- 8< --- detachments --- 8< ---
_______________________________________________
List info: http://lists.roundcube.net/dev/
Could I just poke you devs to please include STARTTLS support sooner
rather than later? I just lost a few hours writing a patch to add this
(I don't know PHP) and later discovered that someone had already done
the same:
http://trac.roundcube.net/ticket/1485284
The implementation isn't that complex and I've found more than a few
messages online involving people trying to get roundcube to use TLS. I
believe STARTTLS is considered the standard way to do both encrypted
IMAP and SMTP (STARTTLS in SMTP is already supported) and it's quite
irritating to have to open up an extra port for IMAP over SSL.
Regards,
Francis
_______________________________________________
List info: http://lists.roundcube.net/dev/