The latest commit on github reflects a solution that is 99.9% done. I have used the reader over the last few days since the previous update. A time boundary now exist. A given feed is never requested more than once in an hour. POCO C++ libraries streamlined that a bit. I changed the rss feed parse to not only focus on the main data values for each headline but do so with a more compact recursive function. The overall solution looks great.
Other than that, I also studied request headers using Mozilla Firefox Developer Tools to tune the request headers emitted through the http interface in the program. I made great progress there for websites accessible on port 80. I am almost at 100%. The POCO C++ library has an issue with the basic http client abstraction when accessing some websites on port 443 aka SSL and increasingly TLS. I am sure it works from Microsoft Windows, but on a mainstream Linux desktop, the standard abstraction for HTTPSClientSession in the POCO C++ library does not work well. Only www.phoronix.com is inaccessible from the current reader implementation. Content comes back but decodes garbled. I suspect an issue with certificate authority access and client selection. No matter, I prefer an abstraction that handles it well with minimal initialization and API calls. I plan to change the program to use a different network communication library. Present candidates are curl and cpr. Fortunately, most of the feeds I use are accessible on port 80 which means all the feeds I am testing with except one pulls into the reader for viewing and access.
Gautier Talks About Technology 