Re: [Evolution-hackers] Local calendars with custom files

[Matthew Barnes <mbarnes redhat com>]

On Mon, 2011-02-07 at 09:50 +0100, Milan Crha wrote:
> And that's what is wrong with it. I do not know the polling interval,
> chosen by gio developers, but I prefer to be able to set the polling
> time myself, for fine-tuning. Not talking about unnecessary network
> traffic for those (rather rare?) cases.

What network traffic?  Evolution only lets you choose a file on the
local file system.


> All three options are used for different purposes and use cases, to be
> able to configure one writer and multiple readers, for example, because
> there is done no file-locking and such.

I'm not suggesting dropping the "force read-only" option.

Users of these custom files expect the calendar data as shown through
Evolution to be kept up-to-date with the file contents and my guess is
they don't care how.  They should never be looking at stale data.

A file monitor is clearly the best mechanism for knowing when to refresh
a local calendar, and if the file never changes then the only resource
lost in monitoring the file is one file descriptor.  Big deal.  I see no
need to burden the user with this choice.