True. And a little better. But it still like pushing a square into a
round hole: Tapestry does not support the "natural" way to do it. I
presume we agree that from a user point of view, start presenting
the
entries we are looking for is the "expected" thing.
Also, there is actually more in this. I can think of other,
reasonable
paging strategies e. g., some entries of overlap for each page. My
gut
feeling is that Grid should be more generic, and that a GridPager
should
be free to define whatever strategy it wants. I think the only thing
which needs to be changed is the Grid's idea of the "actual page"
being
a row nr instead of a page nr. It shouldn't be hard to make this
change
without affecting current code. If required, one could even let
setPage
remain with current semantics?!
Davor Hrg wrote:
> hm,
> you can relatively easily mark the first one,
> so it is noticed instantly
> and that way pager and indexer are not in conflict..
>
> definitely an user friendly feature you're creating there :)
>
> Davor Hrg
>
> On Tue, Mar 11, 2008 at 2:43 PM, Alec Leamas <[EMAIL PROTECTED]
> wrote:
>
>> It's an option, but not a good one. Looking for 'l' might
present 24
>> users beginning with 'k', and a last line of 'Larsson'. This
is just not
>> intuitive.
>>
>> The expected behaveour of an index link is to start presenting
entries
>> according to the link. (like javadoc :-) )
>>
>> --alec
>>
>>
>>
>> Davor Hrg wrote:
>> > why is calculating page not an option ?
>> >
>> > does selected row have to be first or you just
>> > wan to navigate to the fist page that contains the row ?
>> >
>> > Davor Hrg
>> >
>> >
>> > On Tue, Mar 11, 2008 at 2:15 PM, Alec Leamas <[EMAIL PROTECTED]
> wrote:
>> >
>> >> I'm about to convert some T4 code to T5. In this code, I
have a large ,
>> >> paged table of persons with index links with the letters
'a' ..'z' at
>> >> the bottom of each page . Clicking the 'c' link starts
presenting the
>> >> first persons with a name beginning with 'c', There are
also links to go
>> >> on page forward/backward. Like this:
>> >>
>> >> < a b c d e f g h i j k l m n o p q r s t u v x y z >
>> >>
>> >> This is actually useful, it's much easier to press 'l'
looking for
>> >> leamas the to try to guess which page nr he is at.
>> >>
>> >> In T4, I had to recode large part of the table stuff ,
including sort
>> >> etc, to implement this. I hoped T5 would be better, but
the problem seem
>> >> to be still here: The Grid has an internal model of a
fixed number of
>> >> pages, and a current page nr. This doesn't work with the
sliding page
>> >> window required to present "the first page of users
beginning with x'.
>> >> For this to work, the actual page must be defined by the
first row nr.
>> >>
>> >> In other words: A flexible Grid should be able to
position to any row
>> >> in the data, not just to even page boundaries (according
to the
>> >> default's definition of a page).
>> >>
>> >> So, two questions:
>> >> - Am I right thinking that implementing a GridIndex analog
to GridPager
>> >> isn't straightforward in current design?
>> >> - Would i be possible to change the code in Grid to open
up for other
>> >> paging strategies than a fixed nr of (numbered) pages? In
particular,
>> >> would it be possible to store currentRow instead of
currentPage in Grid?
>> >>
>> >>
>> >> Cheers,
>> >>
>> >> --alec
>> >>
>> >>
---------------------------------------------------------------------
>> >> To unsubscribe, e-mail: users-
[EMAIL PROTECTED]
>> >> For additional commands, e-mail: [EMAIL PROTECTED]
>> >>
>> >>
>> >>
>> >
>> >
---------------------------------------------------------------------
>> > To unsubscribe, e-mail: [EMAIL PROTECTED]
>> > For additional commands, e-mail: users-
[EMAIL PROTECTED]
>> >
>> >
>>
>>
>>
---------------------------------------------------------------------
>> To unsubscribe, e-mail: [EMAIL PROTECTED]
>> For additional commands, e-mail: [EMAIL PROTECTED]
>>
>>
>>
>
>
---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]