I had been looking for a better way to help me keep on top of software releases for the projects I’m interested in, be that either personally or for things we use at work.

Some projects have Atom feeds, some have mailing lists just for release updates, some post updates on sites like freshmeat and some have no useful update watching mechanism at all. Tracking all these resources is annoying and a simple unified solution would be much more workable.

cupage is that solution, at least for my purposes. Maybe it could be for you too!


With a local, unified tool we would instantly gain easy access to the updates database for use from other tools and applications.

JSON was chosen as it is simple to read and write, especially so from Python using the json module [1].

The database is a simple serialisation of the cupage.Sites object. The cupage.Sites object is a container for cupage.Site objects. Only persistent data from cupage.Site objects that can not be regenerated from the configuration file is stored in the database, namely last check time stamp and the current matches.

matches is an array, and contains the string matches of previous cupage runs.

checked is the offset in seconds from the Unix epoch that the site was last checked. It is normally a float, but may be null.

An example database file could be:

    "geany-plugins": {
        "matches": [
        "checked": 1256677592.0
    "interlude": {
        "matches": [
        "checked": null
[1]Pickle was used in versions prior to 0.3.0. The switch was made as Pickle provided no benefits over JSON, and some significant drawbacks including the lack of support for reading it from other languages.