Review Request: fix xkcd comic strip

Jonas Pollok jonas.p at gmail.com
Mon Feb 4 00:37:27 CET 2008


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://matt.rogers.name/r/54/#review55
-----------------------------------------------------------


Midnight passed and there is a bug of course. The cache depends on the date.

Before midnight:
03022008 => id = 378 (today with latestid)
02022008 => id = 377
etc.

After midnight:
04022008 => id = 378
03022008 => id = 378 (well in fact the picture and not the id is cached for a certain date. The calculations are OK - id = 377 would be calculated if the cache have not been used)

Well since the patch is not really fixing xkcd I'll remove it from the review-board. Please share your opinion on how to fix this on the mailing list.

- Jonas


On 2008-02-03 13:56:27, Jonas Pollok wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://matt.rogers.name/r/54/
> -----------------------------------------------------------
> 
> (Updated 2008-02-03 13:56:27)
> 
> 
> Review request for Plasma.
> 
> 
> Summary
> -------
> 
> I tried to fix the xkcd part of the comic dataengine.
> xkcd uses incremental ids and not dates to identify a comic strip.
> The old code assumes that xkcd releases one comic strip per day which is not the case.
> I'll write down what i _wanted_ to do and you can either find a better way or improve the code :)
> 
> On creation the latest comic strip id is parsed from 'http://xkcd.com' and associated with currentDate().
> When requesting a comic strip by providing a certain date the delta between the given date and currentDate
> should be the delta between the wanted id and the latest id.
> 
> 
> Diffs
> -----
> 
>   trunk/extragear/plasma/dataengines/comic/xkcdprovider.h
>   trunk/extragear/plasma/dataengines/comic/xkcdprovider.cpp
> 
> Diff: http://matt.rogers.name/r/54/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Jonas
> 
>



More information about the Panel-devel mailing list