TextSearch: trim trailing whitespace.

Node monitoring changes:
* In the ChangesIterator, just remove removed files from the HashMap, regardless
if they could be considered "temporary" or not.
* If a file is removed, we can directly remove it from the results list. This
makes removing files from the result list more robust and quicker if this
was the only thing that happened with regards to node monitoring (the grep
process does not need to be run again).
* Refactored removing result items from the list on result notifications.
* Beginnings of supporting moving files within the watched folder hierarchy.
If they were just moved, the new location should update in the list.
(not well tested)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27377 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added convenience methods for debugging.
* Added better tracing support.
* GetNextName() was not incrementing the index when iterating to the next entry
and was therefor broken if the object managed more than one entry.
* Made a small simplification in EntryRemoved().


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26934 a95241bf-73f2-0310-859d-f6bbb57e9c96

Finished implementing and enabled node monitoring on the folders and files
of the current search. If new files match the pattern, the appear in the
results, or are removed if they don't match anymore. The results also
adapt to changes in the files.
Basically, I added another iterator that is also used to track changes when
node monitor events arrive. Only those changed files are grepped again after
a timeout of .5 seconds when no new node monitor events pour in.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26809 a95241bf-73f2-0310-859d-f6bbb57e9c96

TextSearch: trim trailing whitespace.

Node monitoring changes:
* In the ChangesIterator, just remove removed files from the HashMap, regardless
if they could be considered "temporary" or not.
* If a file is removed, we can directly remove it from the results list. This
makes removing files from the result list more robust and quicker if this
was the only thing that happened with regards to node monitoring (the grep
process does not need to be run again).
* Refactored removing result items from the list on result notifications.
* Beginnings of supporting moving files within the watched folder hierarchy.
If they were just moved, the new location should update in the list.
(not well tested)


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27377 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added convenience methods for debugging.
* Added better tracing support.
* GetNextName() was not incrementing the index when iterating to the next entry
and was therefor broken if the object managed more than one entry.
* Made a small simplification in EntryRemoved().


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26934 a95241bf-73f2-0310-859d-f6bbb57e9c96

Finished implementing and enabled node monitoring on the folders and files
of the current search. If new files match the pattern, the appear in the
results, or are removed if they don't match anymore. The results also
adapt to changes in the files.
Basically, I added another iterator that is also used to track changes when
node monitor events arrive. Only those changed files are grepped again after
a timeout of .5 seconds when no new node monitor events pour in.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26809 a95241bf-73f2-0310-859d-f6bbb57e9c96