[Kde-pim] [kcalcore] DST transitions determined from VTIMEZONE DAYLIGHT RRULE

[Chris Adams]

Hi Allen,

Just an update on this one: a community member took a deeper look into this one
and I believe has found (and provided a fix for) the true root cause of the issue.
I've abandoned my original PR in favour of this new one:

https://github.com/mer-packages/kcalcore/pull/9

Please let me know if you disagree with something it does or any of the assumptions
noted in the discussion :-)

Kind regards,
Chris.

________________________________________
From: Chris Adams
Sent: Tuesday, June 02, 2015 5:29 PM
To: Allen Winter; kde-pim at kde.org
Cc: Pekka Vuorela
Subject: RE: [Kde-pim] [kcalcore] DST transitions determined from VTIMEZONE DAYLIGHT RRULE

Hi Allen,

>> My fix for this issue is at: https://github.com/mer-packages/kcalcore/pull/7/files
>>
>> Firstly, do you think that this is the correct way to resolve this issue?
>> Or does this problem imply a deeper problem with the way the icaltimetype struct data is being
>> interpreted, which should require other changes to KCalCore?  I had a look at the toQTime()
>> function, and it seems sane, so I'm not sure what is going wrong...
>>
>> Secondly, if it is correct, please feel free to cherry pick that change upstream.
>
>Do the regression tests still pass?

I don't have a development environment for upstream set up, so I haven't run those tests.
I have built and run the tests in our (very old) version of kcalcore.  Unfortunately that wasn't
too helpful as there are test failures already in master branch.  At best, I can say my patch
didn't cause any more unit test failures in ours than existed already.  But, as mentioned,
our version of kcalcore is several years old.

>Your change looks ok to me but I'd be happy for David Jarvie and others to review as well.

The more eyes on it, the better.  I did some more investigation into the bug and the results
of that further investigation can be seen in discussion on the PR.  There are at least two
bugs there.  I think my fixes are sane but I'd prefer someone who is more familiar with
the codebase to check them also.

If someone can run the unit tests from mainline with a "normal" environment, that would
also be great.

>
>BTW: I am the libical maintainer.  you can submit merge requests and bugs at
>https://github.com/libical/libical

I think one of the bugs is in KCalCore and the other is in KDateTime.  I don't think either
of the bugs is caused by libical itself, although I may be wrong.

Cheers,
Chris.


_______________________________________________
KDE PIM mailing list kde-pim at kde.org
https://mail.kde.org/mailman/listinfo/kde-pim
KDE PIM home page at http://pim.kde.org/