[Vobject] timezone bugs prevent round-tripping an ical object

[Phil Mayers]

Jeffrey Harris wrote:
> Hi Phil,
> 
>> In the process of writing a Google Calendar sync backend for opensync
>> (using the python plugin) I ran into this.
> 
> Cool!

The good news is it ought to be working short - I need to check the 
mappings are all ok and clean it up, and I'll probably drop it in the 
opensync.org Trac tracker.

> 
>> I confess I find the vobject code impenetrable, so I can't really debug
>> it, but I guess what happens is this:
> 
> A detailed explanation of the code and more useful docstrings are both
> long overdue, so I'm not surprised you found it impenetrable, though of
> course I'm sorry to hear it.  Were there particular bits that you found
> especially opaque?

The core code was ok. The thing that confused me was the "behaviour" 
stuff and how it plugs together behind the scenes. It's powerful and a 
nice way to do it, but it took me a while to trace where the patch was 
needed as a result. Certainly no criticism intended - it's been 
awesomely useful code.

> 
>> -                comp.add('tzname').value  = rule['name']
>> +                if rule['name']:
>> +                    comp.add('tzname').value  = rule['name']
> 
> Yup, right you are, tzname isn't required and we were treating it as if
> it was.  Thanks for the patch, it's in Subversion r145.

Excellent, cheers
Phil