Week 23

This will be a short update as I’ve a few things to still sort out before going to a funeral. I’ve only worked on my replacement ml_wire plug-in (as covered at the end of last week) and it’s now at the stage where it can do most of what the native plug-in could do i.e. check for and download new podcasts.

Currently it only reads RSS feeds (as that’s what I’m already using since that’s all the original plug-in supports) but ATOM feeds will be supported in the coming week. It also has a lower memory usage by ~11MB when using the same feeds data which is a decent saving (ymmv). More can be saved (~8MB for my test data) if dynamically loading most of the feeds information instead of just loading and caching all of the information but that involves more work to achieve than the simpler option I’ve gone with to get something that works.

-dro


Comments:

  1. Hello DrO,
    sorry to just send this to you, I was on holiday. I am also I big podcast user in winamp and I have noticed that sometimes it will download the metadata for a podcast and sometimes not. And these could be episodes in the same feed.I am not sure if the changes you are making will help with that or not,. but I thought that I would mention it. My condolences to the family of the dearly departed.

  2. I’ve a feeling my loss of feeds is similar to your inconsistent behaviour. As the native plug-in doesn’t appear to correctly handle certain character sequences which for me was leading to corruption of the settings files but I can see how that same issue could occur not as badly (depending on the feed) and appear to just skip over things instead of clearing everything out.

    Hopefully the new replacement plug-in will not suffer from such issues as I’ve implemented it using a completely different XML library (for reading and writing the feeds) compared to what the native plug-in uses and which under testing is working reliably (e.g. the feed from http://atp.fm/ is what ended up being the trigger for destroying my settings files and it now work correctly with my new plug-in).

    And thanks for the condolences, is appreciated.

  3. If you are still working on the new ml_wire plug-in, I guess that you could consider this a bug report for the old podcast feed plug-in.
    I have noticed that the old plug-in had a problem with special characters in filenames and even in the notes. A good example is in the following feed.
    http://feeds.feedburner.com/TheAdamCarollaPodcast
    If you download the file from 7/12/2016 with guest Dinesh D’Souza, you will see that the filename gets cut off at the apostrophe and also the notes that get downloaded in the metadata get cut off also at the same point. I think that I have seen the same issue with quotation marks, but I do not have anything to test that with at this time.

  4. I’ve just given that feed a go and the notes show up correctly in my replacement (the expat library doesn’t seem to handle such things nicely which is why i’m using something different for the podcast plug-in which seems to work correctly) though the attempt to download that entry is failing for some reason which I’ll now look into now (the others do download ok so it’s likely I’ve not filtered out the filepath correctly).

  5. Don’t spend too much time trying to download it. I think it is something with the feed because I have tried to download it countless times and it keeps saying failed in my status column. However, it was downloaded successfully at one point because I have the file. I have included a link to what got downloaded for me so you can compare.
    https://drive.google.com/open?id=0B4bA4MA-8e2sWG5PVFh6S19GLUU
    The metadata title says ACS: Dinesh D
    and the notes just say Dinesh D.
    Just stopping at the apostrophe.


Add comment:

Fill out the form below to add your own comments