Age | Commit message (Collapse) | Author |
|
|
|
It's faster, includes a proper triple store with querying, and
generally should be more useful.
|
|
Sometimes pandoc takes unreasonably long time to parse a
document (though it warns about that in the documentation), so a
safeguard is needed.
|
|
With two types of caching: file-based for retrieved documents, and
memory-based for predicate labels extracted from those.
|
|
|
|
|
|
|
|
Finding the last HTTP 200 response via the availability API, so a
dummy "web-archive" scheme is added into the default configuration in
order to use a custom command.
The URI_ESCAPED environment variable is also introduced, though not
used here -- but may be handy for similar custom commands.
|
|
|
|
User commands are adjusted, and the code is slightly refactored in
order to fit this better.
|
|
Mode names mostly match file extensions, so try to use those when
other methods fail.
|
|
Identifiers get handled better with those. And since they also add
divs, rendering of lists of blocks is slightly improved to avoid
unnecessary spacing.
|
|
|
|
Reading effective URI and content type (if those are available) now,
aiming `curl -w` or similar commands.
Not a particularly nice way, and complicates both the program and the
configuration, but sometimes file name extensions are deceptive.
|
|
|
|
|