Parsing the LugRadio "episodes.rss" feed on a N900 (600 MHz Cortex A8) with an experimental parser I've written in C (yes, that's 0.8%, not 80%).
no plus ones
Shared publicly•View activity
- Finally giving up on Python then? ;-)Feb 8, 2013
- Not giving up on Python, but trying to write critical parts in C++ and load them as extension modules. Glueing components together and creating prototype implementations is still nicer in Python for now (also, there's a big existing codebase that I'm just not going to rewrite in C++ from one day to the next).
Also, the comparison is a bit unfair here - feedparser does more than what cpodder does, but cpodder only takes what it needs for parsing podcast feeds (and throws the rest away). Using the same algorithm and comparing C implementation, CPython (no JIT) and PyPy (with JIT) might give more realistic results - I still assume C++ will be faster (and will take less memory than PyPy w/ JIT).Feb 8, 2013
- ha! why our feed in particular? :-)Feb 8, 2013
- http://lugradio.org/episodes.rss) that I know by heart, and the first one that comes to mind when testing something ;)It's the shortest feed URL (Feb 8, 2013
- Yay! Our malign influence continues :)Feb 8, 2013
- Pretty cool!Feb 8, 2013
- Very impressive! After reading https://github.com/antono/vala-object , my interest in moving computationally intensive logic to a compiled library was definitely piqued. Now I need to make a test.Feb 8, 2013