I wanted to update my pipe by finally upgrading from the deprecated fetch page module to the [not so new anymore] XPath fetch page module.
I did, as exemplified in this pipe, but I encountered a bug.
The module tries to interpret every mark up as a tree branch, as it should, but it's not good for bold mark ups <b> or any text formatter mark-up for that matter that are placed mid-text. Look in my pipe, just below the url builder module there are 2 XPath fetch page modules: the one on the right in there to show the bug, fetches recent torrents from TPB. All TPB torrents have details formatted as a text that I want to use. The format of those details is such as this example:
<font class="detDesc">Uploaded <b>34 mins ago</b>, Size 586.74 KiB, Uploaded by <a class="detDesc" href="/user/BoscoBosco/" title="Browse BoscoBosco">BoscoBosco</a></font>
And here is the trouble. With the bold <b> mark-up mid-text, XPath FPM puts "Uploaded" in font.content field, "xx minutes ago" in a font.strong field (instead of a font.content.strong field btw) and forget the rest of the text. Weirder, in the debugger, the rest of the text is shown instead of "Uploaded", whereas when I try to use the content of the field, it prints "Uploaded".
Any thoughts, workarounds, ways to bring this bug to the attention of Yahoo devs?
yes, it changes. Actually the bold 'n minutes ago' is only there for less than a day after the publication of the torrent, afterwards it just prints the date, no boldface. However, as I fetch the feed once or twice per day, I get this for new versions when they're out.
That's why I put the fetching of most recent torrents (any) in the example pipe to show an example of this bold tag inside the font tag.
I'm guessing those cases (text mark ups inside text) are handled in a <p> mark up, but were overlooked in the case of a <font> mark up? (although using a font for a p is really not a best practice...)