Autocompletion relevance

[Nikolai Weibull <epiphany-list pcppopper org>]

Hi.
I have some complaints about the autocompletion item relevance
calculation.  It seems that it weighs items rather poorly, at least for
my browsing patterns.  What I would like it to do is:

If I've typed a keyword, I expect bookmarks with that keyword to show up
first, and then, perhaps bookmarks and history items matching the text
typed so far sorted after, preferably visually separated in the list.
The bookmarks that belong to the given keyword should then be sorted
after visiting frequency and the same for the other bookmarks and
history items.  And finally, if two items in the same group (keyword,
bookmark, history) have the same visiting frequency, then they should be
sorted alphabetically.

As it is modelled now, I couldn't find a trivial solution for this.
Perhaps someone with a bit more knowledge of the code could?
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}