In Horde_Timezone we would have to allow to set a date range that we 
need timezone data for. This way we could limit the distributed 
VTIMEZONE elements to the ones that are actually important to the 
event(s).

Hello,

For the same event exported from the calendar:
- Horde produce a 6.3 kB file
- Sogo produce a less than 1 kB file

After checking, the main difference is because Horde produce more Olson
data.

Sogo include:
TZNAME:CEST
TZNAME:CET

Horde include:
TZNAME:PMT
TZNAME:WEST
TZNAME:WE-T
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WE-T
TZNAME:WEST
TZNAME:WE-T
TZNAME:WEST
TZNAME:WE-T
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WEST
TZNAME:WE-T
TZNAME:WEST
TZNAME:CEST
TZNAME:CE-T
TZNAME:CEST
TZNAME:CE-T
TZNAME:CEST
TZNAME:CE-T
TZNAME:WEMT
TZNAME:WEST
TZNAME:WEMT
TZNAME:WE-T
TZNAME:CEMT
TZNAME:CE-T
TZNAME:CEST
TZNAME:CE-T
TZNAME:CEST
TZNAME:CE-T
TZNAME:CE-T
TZNAME:CE-T
TZNAME:CEST
TZNAME:CE-T

Except if I'm wrong, for TZID:Europe/Paris, only TZNAME:CET and
TZNAME:CEST are usefull. Sogo use only this two TZNAME.

I suggest to include only really useful TZNAME in the export done by
Horde.

Reducing the file size by 6 is very important as most of CalDAV traffic
is done with mobile device (roaming traffic is very expensive or the
bandwidth can be very limited and saving every Kbps is important for be
able to have fast sync).

I don't known how fix that as a loop is used in Horde/Timezone for
generate all TZNAME.

For info, Sogo have precompiled data for each timezone and don't use
tzdata database from IANA. I agree that the method used by Horde is more
reliable as it use tzdata directly from IANA.


Thanks

Best regards,

--
Nicolas DEFFAYET