[mythtv-commits] Ticket #9676: Obtaining program times across daylight savings time change gives inaccurate results
MythTV
noreply at mythtv.org
Sun Mar 27 02:27:40 UTC 2011
#9676: Obtaining program times across daylight savings time change gives
inaccurate results
---------------------------------------+----------------------------
Reporter: mythtv@… | Owner: mdean
Type: Patch - Bug Fix | Status: assigned
Priority: minor | Milestone: unknown
Component: MythTV - Mythfilldatabase | Version: 0.24-fixes
Severity: medium | Resolution:
Keywords: | Ticket locked: 0
---------------------------------------+----------------------------
Comment (by mythtv@…):
I believe the final patch I attached solves the problem. I did some unit
level testing, but it's hard to say 100% just because I can't force the
guide data to span a daylight savings time change. I could only rig up
some unit tests.
To summarize
* there were some calculations in the code that may have been written to
an older, less complete QtDateTime API. They didn't handle daylight
savings correctly.
* EITHelper is written to calculate an offset from UTC in its
constructor. It used that in its calculations rather than using
QtDateTime's features. When EITHelper lives across a daylight savings
change (which changes the offset from UTC) it starts calculating start
times incorrectly, so it updates the schedule to be an hour off over time
periods covered by the EIT.
I think this patch solves the problems I've seen discussed in several
places, like this thread: http://www.gossamer-
threads.com/lists/mythtv/users/427478
For myself, I happened to leave on vacation right before the time changed.
Myth recorded shows off by an hour for more than a week. I cleared out the
guide data and redownloaded it (without restarting myth) and it continued
to be off by an hour. Restarting myth AND redownloading the data worked
around the issue. This has happened to me on multiple previous time
changes, and I know from those times that restarting myth only doesn't
fully address the issue. In the thread above, the user's data being off
for the next two days only would be consistent with about how long EIT
data might exist for.
Thanks, mdean, for looking at this patch. The time change thing has gotten
me twice a year for years, so I finally decided the best thing to do was
try and fix it myself. I hope I got it.
--
Ticket URL: <http://code.mythtv.org/trac/ticket/9676#comment:5>
MythTV <http://code.mythtv.org/trac>
MythTV Media Center
More information about the mythtv-commits
mailing list