Community
Participate
Working Groups
If you invoke the Ctrl+H (Search dialog) when your focus is on task list, it should bring the "Task Search" tab on focus, like when you type Ctrl+H on a Java editor and it brings you the "Java Search" tab.
Agreed. Although I often wonder whether this workbench behavior is wrong more of the time than just bring up the previous search page would be, due to the frequency of re-running searches. Willian: do you know the workbench hook needed to support this?
No, unfortunately I don't know what's the "magic"... but I can investigate...
Created attachment 56118 [details] Patch I had to turn ITaskListElement into a IAdaptable, because this mechanism relies on getting a adapter for ISearchPageScoreComputer for current selection. Also, this only works if you have something selected on task list, i.e. only having the task list focused with nothing select won't bring the task search page.
Created attachment 56119 [details] mylar/context/zip
Also, with this patch you will need to include a dependency on org.eclipse.equinox.common in org.eclipse.mylar.doc plugin. I forgot to include this on previous patch.
Great. Since this changes API it will need to wait until we branch for 2.0, probably within the next couple of weeks. Also need to consider along with bug 107169.
Mik, are planning to apply this?
Will review this shortly.
Willian: my apologies, I let this go stale due to all the refactorings I did this week. Could you cut the patch again?
Created attachment 59441 [details] Updated patch
Created attachment 59442 [details] mylar/context/zip
Patch applied but I couldn't make the search page come up with the Task Search. Do you need some kind of plugin.xml entry to make this work?
I think I forgot to include the adapter extension from plugin.xml or for some unknown reason the context does not consider it "interesting" anymore, because it was present on the first patch. BTW Mik, do you have some article, paper, etc, describing how the increasing/decreasing of interest works internally on Mylar? (I don't mean source code ;) I have some contexts that have a strange behavior, like folders not touched for a long time does not getting uninterested, or editors not being restored after some days not touching the context. I thought about opening bugs for these, but I'd like to understand better how the score, rules work before to have a reproducible test case. Here is the snippet: Index: plugin.xml =================================================================== RCS file: /cvsroot/technology/org.eclipse.mylar/org.eclipse.mylar.tasks.ui/plugin.xml,v retrieving revision 1.185 diff -u -r1.185 plugin.xml --- plugin.xml 19 Dec 2006 23:06:07 -0000 1.185 +++ plugin.xml 23 Dec 2006 02:17:46 -0000 @@ -692,6 +692,16 @@ </commandParameter> </command> </extension> + <extension + point="org.eclipse.core.runtime.adapters"> + <factory + adaptableType="org.eclipse.mylar.tasks.core.ITaskListElement" + class="org.eclipse.mylar.tasks.ui.search.SearchScoreComputerAdapterFactory"> + <adapter + type="org.eclipse.search.ui.ISearchPageScoreComputer"> + </adapter> + </factory> + </extension> </plugin>
Patch applied and verified, great stuff Willian. Regarding packages getting 'stuck' as interesting, that's bug 118542 Regarding documentation on the model, I'll post a link to my PhD thesis soon, but for now the best description is the Foundations of Software Engineering paper available from the Publications page: http://www.eclipse.org/mylar/publications/2006-11-mylar-fse.pdf
(In reply to comment #14) > Regarding packages getting 'stuck' as interesting, that's bug 118542 > I think what I'm experiencing is somewhat related to this. I'll post there. > Regarding documentation on the model, I'll post a link to my PhD thesis soon, > but for now the best description is the Foundations of Software Engineering > paper available from the Publications page: > http://www.eclipse.org/mylar/publications/2006-11-mylar-fse.pdf > Good. I'll take a look.
