The built-in components can not be all things to all people. You can often reuse some of the sub-components of Grid to build a customized Grid.
I think I would use a filter on the Grid; so you have a set of controls for selecting the search letter, that applies a filter to the Grid contents, then the Grid renders just the values within that letter. On Tue, Mar 11, 2008 at 6:58 AM, Alec Leamas <[EMAIL PROTECTED]> wrote: > 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: [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] > >> > >> > >> > > > > --------------------------------------------------------------------- > > 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] > > -- Howard M. Lewis Ship Creator Apache Tapestry and Apache HiveMind --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]