|well, here's what i figured out regarding DateList...
jens wille wrote:
>> In addition, the list should be sorted descending, with the
>> most recent date on top. I guess a look at the manual would
>> tell me how to control the number of items shown on the list,
>> but if you can answer that faster, I'd be much obliged :-)
> afaik, there is no option for reverse sorting.
i added an "newestfirst" option to DateList - and this works (at
least when used in combination with -nogroup).
> apart from that, i don't think AZCompactList is what you're
> looking for: even if it works (without the "empty" records ;-)
> you get single entries for _every_ date (dd-mm-yyyy); while you
> more likely want an hierarchical access (year -> month -> day),
> don't you?
> i tried to modify the DateList classifier to do exactly this, but
> i'm afraid there's more to it than just giving $date an
> appropriate value ("yyyymmdd" instead of seconds since the
> epoch). this just doesn't work when displaying the resulting
> DateList - probably there are yet more components involved (which
as seems to me there are some "weird" things going on with DateList:
1. if you have more than one DateList they get mixed up, because
cgiargdm isn't properly transferred between these classifiers.
2. i think this has to do with another point: the dm argument is
included only in the button href. so in case you use -buttonname
whithout a proper button (which i did, after modifying DateList to
accept the -buttonname option as other classifiers do) there will be
no cgiargdm, which makes greenstone use the default value (or the
last one used). i didn't try it with a custom image button though.
3. when displaying the DateList classifier the receptionist grabs
the date metadata (denoted by dm) on the fly, expecting it to be in
yyyymmdd format. so lastmodified will be displayed as something like
"1106 _textmonth53_" which is impossible to control by format
and because i can't control the dm arg either (there isn't even a
macro for this - it's all in the receptionist code) i don't see any
"easy" solution to this problem :-(
there's more competence in demand i'm afraid...