Present: ellie, Bron, Robert 
 
Robert: 
* CalConnect was good - everything was captured in Bron's notes 
* have pushed a couple of charset API changes 
 - JMAP mail handles wrongly encoded utf32 bodies (covers 2 of 3 raised issues) 
 - now will evaluate the two charset libraries for detecting the charset of 
arbitrary binary data 
 - have worked a bit with libicu internal detector, but might not be good 
enough, may need to add external dependency 
* planning to look at generating mboxevents for keyword changes 
* will update jscalendar drafts 
* for jscontact, follow up with Peter and join VCARD mailing list 
 
Partha: 
* Ran a bunch of YCSB benchmarks, but looked at LevelDB's benchmarks instead. 
 - compare to what they run against leveldb, kyoto, redis, etc. 
 - interfaces on leveldb benchmark are a closer match to Cyrus 
* Zeroskip - waiting for JMAP changes to settle before merging. 
 - Two big patches: one changes cyrusdb interfaces to unsigned char, then 
zeroskip driver. 
 - unsigned changes are large and touch many places! 
 - other possibility is to first pass just cast the other way for zeroskip and 
merge that first. 
 
ellie: 
* hoping to get 2.5.12 out this week 
* Cyrus is no longer being packaged directly by Debian 
 - we should look at how we want to distribute Cyrus for Debian users - nightly 
builds? Releases? 
* brought back annotatemore support behind a config option 
 
Bron: 
* have been looking at JMAP ACL uploads 
* in order to set sortOrder on intermediate folders, they need to be able to 
set annotations, did some work on that, still more to do there. 
 - Robert will have a look at it. 
* unlikely to have time for Cyrus work this week. 
* *
*Daylight savings will be hitting Australia soon, moving the Melbourne time to 
10pm.* 
 
 
Bron's Notes from CalConnect: 
 
The conference is hosted by 1&1. We started by talking through what we want to 
focus on during this time. Here's my list: 
 * See where we’re at on JSCalendar and push to completion 
 * Start JSContact work 
 * CalDAV sharing work maybe? (federation) 
 * CalDAVTester Cassandane integration 
 * Topic for discussion: event in multiple calendars with same UID (labels for 
calendars?) 
 
The conference has a new structure now - combined Technical and Conference 
sessions mixed each day rather than first half Technical/Testing and second 
half talking. 
 
Interesting topics covered: 
 * Server-side-subscriptions (similar to a FastMail feature, but standard and 
can be managed over CalDAV by clients) 
 * CalDAVTester and the framework for regression testing 
 * Autodiscovery & Push (will work through at IETF) 
 * Push 
 * RRULE edge cases 
 * CalConnect is working with ISO and M3AAWG and working on other partnerships. 
We'll start publishing or co-publishing standards. 
 * ISO8601-2 standard for much more flexible dates is going to be done soon 
 * VCARD work is underway to give more globally compatible address format, name 
format, etc. 
 * JSCalendar - Robert presented. I'll copy the full notes from that. 
 * Demoed some JMAP using the FM interface. 
 * Discussed ETags and how weak ETags are needed unless resources are byte 
identical due to caching and partial ranges. 
 * Next conference will be in Zurich in February 2019. 
 
*JSCalendar Notes:** *
 * current version -07, in WGLC. 
 * Lots of decisions made slightly smiling face  
 * Open points: 
 * Durations 
 * JSCalendar → requires it to be in UTC 
 * iCalendar → done in event timezone 
 * *DECISION OF ROOM: align with iCalendar* 
 * Participant email: 
 * key can be any URI 
 * if there’s an EMAIL= parameter, that becomes the “email” key 
 * if there’s no EMAIL= parameter, extract “email” key from the mailto: 
 * if neither, “email” key is null (default) 
 * Discussion of timeZone on location objects → change from MUST to MAY such 
that  
 * rsvpWanted → *replyWanted* 
 * rsvpResponse → *participationStatus* 
 * participation →* **attendance { none / optional / required }* 
 * roles → *object* mapping to true 
 * Splitting locations 
 * real locations vs virtualLocations 
 * participant can define which physical location, but not which virtual 
location (if any) 
 * physical locations can have timezone, coordinates, address, etc. 
 
-- 
 Bron Gondwana, CEO, FastMail Pty Ltd 
 br...@fastmailteam.com 
 
 

Reply via email to