Minutes March 14th

[Ken Murchison]

Sorry I missed the meeting.

I'm still working on sharing, amid other distractions.  I am working 
through the idiosyncrasies between the iOS and OSX clients.  Some of the 
XML elements that have been pseudo-documented by Apple as being optional 
are really mandatory, so I'm altering my code appropriately and doing 
some code cleanup along the way.


On 03/14/2016 07:26 AM, Bron Gondwana via Cyrus-devel wrote:
> Present: Bron, Robert, Nicola, Raymond, Jeroen
>
> Bron:
> - have been working on a bunch of things at the last minute (just the few hours before the meeting)!
> - IMIP support back in master (if you don't set a notifier, it uses sendmail like before, but with the per-recipient logic that stops use sending multiple messages if there are multiple attendees, and also the update detection)
> - Merged Raymond's object storage work to master
> - Merged the cyrus-docs repository (with rewritten history to be under /doc) into cyrus-imapd repository master
> - Merged Robert's additional charset work onto master.
> - Worked with Robert's Xapian patches to build a local test build for testing the Xapian SearchFuzzy tests (not finished yet)
> - Other than that mostly working on FastMail support issues.
>
> Raymond:
> - Jean-Francois has promised he would work on OpenIO docs, haven't heard back yet.
> - Jeroen - does anyone have OpenIO completely packaged?  - Jean-Francois will provide packages.  Github wiki refers to packages. (Centos/Redhat/Fedora)
> - Raymond is running it on Ubuntu.  Not from packages, did installation a while ago.   They've been cutting down on dependencies, so it's probably easier now.
>
> Robert:
> - added some charsets to master repository (Japanese, Chinese, Baltic) - simple table based
> - now working on integrating libicu into charset library - support for nearly every charset!
> - should improve search (normalise all the charsets to UTF8).  Getting libicu working with pipeline.
> - working on character based now, but if performance is an issue we can pass blocks at a time to libicu.
>
> Jeroen:
> - last week - discovered that discrete murder forcefully uses getannotation/annotatemore against backends.
> - merged a patch that Ken has done based on backend capabilities
> - gets metadata responses to getannotation reports.  Got fix for selves.  D86.  Have asked Ellie to take a look.
> - also specialuse fetch from backend.  Ken already has some patches for that.  Will look into them.
> - some discussion of speed implications of doing annotation lookup for every line of LIST response vs fetching entire list of specialuse from the backend in one roundtrip and then using a hash or list lookup locally.
>
> Nicola:
> - nothing to add (but has made some Docs layout fixes during the week per the repo commits)
>

-- 
Kenneth Murchison
Principal Systems Software Engineeer
Carnegie Mellon University