tigefa4u projects Trac

Search API: matches in wiki page names also count for 5 normal hits, as for matches in ticket summaries, plus a bug fix in `get_matching_lines` when the given text might occasionally be `None`.

Search API: initial adaptation of the existing search sources to the new API

Search API: forgot to add style specification for the number of hits

TracSearch: refactor the Search API in a similar way than the recent Timeline API refactoring (see r5523).

Quite similar to the `TimelineEvent` object used to collect timeline entries, there's now a `SearchResult` object for collecting the matches provided by the search sources, instead of the tuple used before.

The advantage of such an approach over the use of a tuple is an increased clarity and greater extensibility of the API.

In addition to the fields and methods which are similar to the ones of the `TimelineEvent` object, there's also a '''weight''' property which will be used to sort the matches by order of relevance.

Each occurrence of the terms in the matched resource would normally count for 1. A search source can increase the weight of some hits if the search terms appear in some more sensitive place (in the title of the resource, in a ''keywords'' field, etc.). As this weight is not an exact measure of the hit count, it can also be referred to as the ''bogohits'' count.

The legacy ISearchSource providers are still supported.

Setup merge tracking on that branch.

Creating a sandbox for some improvements to the search subsystem.

We don't need to render the template when debugging the template data using ''hdfdump=1''

Improve rendering of page listing for search results. In particular, this makes more obvious what is the currently selected page.

Disposing of timeline-api-tmp branch merged in r5523

FieldRefactoring: merging changes from [5470:5523/trunk]