2 * fail if item title is not called at first
3 * test for buffer overflows when reading from global
4 buffer into special ones (item, etc).
5 * test whether nomtime can handle complicated filenames
6 * test driven refactoring
8 * replace assertions by proper error handling
9 * feed name is cutted if too long (replace using full stops)
10 * feed/title/desc injection not possible
11 * check whether new items are retrieved, but old ones are
13 * check whether dates are handled properly (timezones!)
14 * fix remaining bugs (see README)
16 * item overwrite only if date >= olddate
17 else notice, skip offending item, and do not change
18 feed date, continue but exit with error
19 (invariant: date of every unfetched item is > feed date)
20 * Make directories/items read-only.
21 * support lastBuildDate and pubDate for channels
22 * do not stop on first old item, consume whole feed
23 channel pubDate is not old.
24 * maybe use http last-modified? it is not reliable on
25 some servers, though, and probably needs ugly hacks
27 * we're currently not just aggregating, we're aggregating
29 * Drop HTML formatting!
32 |-- news item title, probably truncated or hashed?
34 | |-- description (opt)
38 * now *every* interface can be built on top.