Quote:
Originally Posted by FaceDeer
...
I guess the way to fix this would be to have FFDL parse the chapter title out of the page's text rather than trusting the API. Something like that could also fix the rare accented-character-in-title-gives-null-title-in-API bug too. But that seems like a bit of a hassle and a kludge to fix something that's properly FiMFiction's responsibility to correct. My only real concern is that if one of my epubs gets an incorrect chapter title inside it there's no error message letting me know it needs to be fixed, but I guess that's a pretty minor flaw.
Edit: it just occurred to me that this could potentially cause a bigger problem under the right circumstances, if there's a 50-chapter epic and the first chapter is "hidden" every chapter afterward would have its name shifted down to the next chapter. That's kinda annoying. *makes a new years' resolution to keep bugging Knighty at regular intervals*
...
|
My first thought was to just stop using the API altogether and go back to screen scraping. But that would be a pain.
Then I made a version that does as you suggest. For fimfic (and only fimfic), it would update the chapter titles with the chapters that are downloaded.
But that's still broken in the face of an update to the 50 chapter epic--because it wouldn't be downloading all those chapters again, it would use the titles from the API for old chapters.
So really, the only compromise left is to scrape the chapter list. Turns out FFDL already has to scrape to get the characters, so it wasn't a huge hardship. But the fimf API does become less and less useful.
Try the attached version and see how it works for you. If you have an example for the 'accented-character-in-title-gives-null-title-in-API bug', I'd be interested to hear about it, too.
UPDATE Jan 04, 2012 -- Remove obsolete beta versions