Please try the latest source or next nightly to see if it's
sufficiently improved.
Christiaan
On 27 Feb 2009, at 7:28 PM, Stuart Andrews wrote:
>
> Thanks Adam,
>
> As I am just connecting to sourceforge, I have not looked at the
> code. But once I get that far, I will be asking myself why
> we need to cycle through all publications after edited just one?
>
> Also, I liked your suggestion for disabling the updating of counts ...
> with my BibDesk usage, I would be happy if the smart groups were
> computed just once, at startup. Can you recommend a starting point to
> tweak this code?
>
> - Stu
>
>
> On Feb 26, 2009, at 9:28 PM, Maxwell, Adam R wrote:
>
>> On 02/26/09 16:59, "Stuart Andrews" <stu.andr...@gmail.com> wrote:
>>
>>> So, my question is whether there is a way to make smart groups less
>>> diligent in terms of updating their contents after every edit. I
>>> would be happy if I could trigger this update myself ... "Ok smart
>>> groups, show us your beach-ball magic".
>>
>> The problem is that the count is updated after each change, so it
>> can be
>> displayed next to the group name. You could hack the code so smart
>> groups
>> were updated when selected or something, if you didn't mind an
>> incorrect
>> count (or disable the count for smart groups).
>>
>> The best way to fix it would probably be a data model that would
>> maintain
>> state and relationships in groups (this is greatly complicated by
>> things
>> like crossrefs and macros). At present, smart groups don't maintain
>> any
>> state; they're essentially a persistent search. IIRC this works as
>> follows:
>>
>> for (group in smartGroups) {
>> for (publication in [document publications]) {
>> for (condition in [group conditions]) {
>> if ([condition isSatisfiedByPublication:publication])
>> // add to display list
>> }
>> }
>> }
>>
>> So any change to a publication triggers a linear search through the
>> entire
>> publications list, evaluating each condition...and that search is
>> repeated
>> for each group. If it could be changed to iterate the publications
>> at the
>> outer loop, that would probably help. Threading to evaluate groups
>> in
>> parallel would be a brute-force win if you have multiple cores, but
>> likely
>> not be worth the effort.
>>
>> --
>> adam
>>
>>
>> ------------------------------------------------------------------------------
>> Open Source Business Conference (OSBC), March 24-25, 2009, San
>> Francisco, CA
>> -OSBC tackles the biggest issue in open source: Open Sourcing the
>> Enterprise
>> -Strategies to boost innovation and cut costs with open source
>> participation
>> -Receive a $600 discount off the registration fee with the source
>> code: SFAD
>> http://p.sf.net/sfu/XcvMzF8H
>> _______________________________________________
>> Bibdesk-users mailing list
>> Bibdesk-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/bibdesk-users
>
> Stuart Andrews
> stu.andr...@gmail.com
>
>
>
>
> ------------------------------------------------------------------------------
> Open Source Business Conference (OSBC), March 24-25, 2009, San
> Francisco, CA
> -OSBC tackles the biggest issue in open source: Open Sourcing the
> Enterprise
> -Strategies to boost innovation and cut costs with open source
> participation
> -Receive a $600 discount off the registration fee with the source
> code: SFAD
> http://p.sf.net/sfu/XcvMzF8H
> _______________________________________________
> Bibdesk-users mailing list
> Bibdesk-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bibdesk-users
------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
Bibdesk-users mailing list
Bibdesk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bibdesk-users
